ENGLISH HISTORICAL SYNTAX AND MORPHOLOGY

AMSTERDAM STUDIES IN THE THEORY AND

HISTORY OF LINGUISTIC SCIENCE
General Editor
E. F. KONRAD KOERNER
(University of Ottawa)

Series IV – CURRENT ISSUES IN LINGUISTIC THEORY

Advisory Editorial Board

Raimo Anttila (Los Angeles); Lyle Campbell (Christchurch, N.Z.)

Sheila Embleton (Toronto); John E. Joseph (Edinburgh)
Manfred Krifka (Berlin); Hans-Heinrich Lieb (Berlin)
E. Wyn Roberts (Vancouver, B.C.); Hans-Jürgen Sasse (Köln)

Volume 223

Teresa Fanego, María José López-Couso and Javier Pérez-Guerra (eds)

English Historical Syntax and Morphology

ENGLISH HISTORICAL
SYNTAX AND
MORPHOLOGY
SELECTED PAPERS FROM 11 ICEHL,
SANTIAGO DE COMPOSTELA,
7–11 SEPTEMBER 2000

TERESA FANEGO
MARÍA JOSÉ LÓPEZ-COUSO
University of Santiago de Compostela

JAVIER PÉREZ-GUERRA
University of Vigo

JOHN BENJAMINS PUBLISHING COMPANY

AMSTERDAM/PHILADELPHIA
 TM The paper used in this publication meets the minimum requirements of American
8

National Standard for Information Sciences — Permanence of Paper for Printed

Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data

International Conference on English Historical Linguistics (11th : 2000 : Santiago de Compostela,
Spain)
English historical syntax and morphology : selected papers from 11 ICEHL, Santiago de Compostela,
7–11 September 2000 / [edited by] Teresa Fanego, Maria José López-Couso, Javier Pérez-Guerra.
p. cm. -- (Amsterdam studies in the theory and history of linguistic science. Series IV, Current
issues in linguistic theory, ISSN 0304-0763 ; v. 223)
Includes bibliographical references and indexes.
1. English--language--History--Congresses. 2. English language--Grammar, Historical--Congresses.
3. English language--Morphology--Congresses. 4. English language--Syntax--Congresses. I. Fanego,
Teresa. II. López-Couso, Maria José. III. Pérez-Guerra, Javier. IV. Title. V. Series.
PE1075.I57 2002
420’.9--dc21 2002025421
ISBN 90 272 4731 5 (Eur.) / 1 58811 192 X (US) (Hb; alk. paper)

© 2002 – John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other
means, without written permission from the publisher.
John Benjamins Publishing Co. • [Link] 36224 • 1020 ME Amsterdam • The Netherlands
John Benjamins North America • [Link] 27519 • Philadelphia PA 19118-0519 • USA
<TARGET
< R/ RE E F"toc"
F DOCINFO

"AUTHOR
"ack">
"intro">
"aki">
"all">
"bib">
"bri">
"kas">
"kor">
"kru">
a d dress"
""

TITLE "Table of contents"

>

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Table of contents

Addresses vii
Acknowledgements ix
Introduction 1
Teresa Fanego
Two types of passivization of ‘V+NP+P’ constructions in relation to
idiomatization 9
Minoji Akimoto
On the development of a friend of mine 23
Cynthia L. Allen
Historical shifts in modification patterns with complex noun phrase
structures: How long can you go without a verb? 43
Douglas Biber and Victoria Clark
Grammaticalization versus lexicalization reconsidered: On the late
use of temporal adverbs 67
Laurel J. Brinton
The derivation of ornative, locative, ablative, privative and
reversative verbs in English: A historical sketch 99
Dieter Kastovsky
From gold-gifa to chimney sweep? Morphological (un)markedness of
Modern English agent nouns in a diachronic perspective 111
Lucia Kornexl
A path to volitional modality 131
Manfred Krug
< /R/TREARGET
E FF

"len">
"los">
"meu">
"moh">
"sch">
"ni">
"si">
"toc">

vi Table of contents

Is it, stylewise or otherwise, wise to use -wise? Domain adverbials and

the history of English -wise 157
Ursula Lenker
The loss of the indefinite pronoun man: Syntactic change and
information structure 181
Bettelou Los
The progressive in Older Scots 203
Anneli Meurman-Solin
Detransitivization in the history of English from a semantic
perspective 231
Ruth Möhlig and Monika Klages
Morphology recycled: The Principle of Rhythmic Alternation at
work in Early and Late Modern English grammatical variation 255
Julia Schlüter

Name index 283

Subject index 289
<TARGET "address" DOCINFO

AUTHOR ""

TITLE "List of Contributors"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Addresses

Minoji Akimoto Teresa Fanego

Aoyama Gakuin University Department of English
3–12–35 Azamino Facultad de Filología
Aoba-ku Universidad de Santiago de Compostela
Yokohama-shi E-15782 Santiago de Compostela
Kanagawa-ken 225–0011 Spain
Japan iafanego@[Link]
miha-ru@[Link]
Dieter Kastovsky
Cynthia L. Allen Institut für Anglistik und Amerikanistik
School of Language Studies Universität Wien
Australian National University Universitätszentrum Altes AKH
Canberra ACT 0200 Spitalgasse 2–4, Hof 8
Australia A-1090 Wien
[Link]@[Link] Austria
[Link]@[Link]
Douglas Biber
Department of English Monika Klages
Northern Arizona University Englisches Seminar
Flagstaff, AZ 86011–6032 Universität zu Köln
USA Albertus-Magnus-Platz
[Link]@[Link] D-50923 Köln
Laurel J. Brinton Germany
[Link]@[Link]
Department of English
University of British Columbia Lucia Kornexl
#397–1873 East Mall Institut für Anglistik/Amerikanistik
Vancouver, B. C. V6T 1Z1 Ernst-Moritz-Arndt-Universität
Canada Greifswald
brinton@[Link] Steinbecker Str. 15
Victoria Clark D-17487 Greifswald
Department of English Germany
kornexl@[Link]
Northern Arizona University
Flagstaff, AZ 86011–6032
USA
vec@[Link]
</TARGET "address">

viii Addresses

Manfred Krug Anneli Meurman-Solin

Englisches Seminar Department of English
Albert-Ludwigs-Universität University of Helsinki
Postfach P. O. BOX 4
D-79085 Freiburg Yliopistonkatu 3
Germany Helsinki 00014
[Link]@[Link] Finland
[Link]-Solin@helsinki.fi
Ursula Lenker
Institut für Englische Philologie Ruth Möhlig
LMU München Englisches Seminar
Schellingstr. 3 RG Universität zu Köln
D-80799 München Albertus-Magnus-Platz
Germany D-50923 Köln
[Link]@[Link] Germany
[Link]@[Link]
María José López-Couso
Department of English Javier Pérez-Guerra
Facultad de Filología Department of English, French &
Universidad de Santiago de Compostela German
E-15782 Santiago de Compostela Facultad de Filología y Traducción
Spain Universidad de Vigo
iacouso@[Link] Campus As Lagoas Marcosende
Bettelou Los E-36200 Vigo
Vakgroep ATW Spain
jperez@[Link]
Vrije Universiteit
De Boelelaan 1105 Julia Schlüter
1081 HV Amsterdam Fachbereich 3
The Netherlands Anglistik — Amerikanistik
[Link]@[Link] Universität Paderborn
Warburger Str. 100
D-33098 Paderborn
Germany
schlueter@[Link]
<TARGET "ack"
</TARGET "ack">DOCINFO

AUTHOR ""

TITLE "Acknowledgements"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Acknowledgements

This volume and its companion one (CILT 224) were realized with the help of
a great many people. First of all, we would like to thank all authors for their
contributions and cooperation. We are also greatly indebted to the following
colleagues who acted as anonymous external reviewers: Claire Cowie, Dolores
González-Álvarez, Santiago González Fernández-Corugedo, Maurizio Gotti,
Dieter Kastovsky, Ans van Kemenade, Ursula Lenker, Angelika Lutz, Donka
Minkova, Arja Nurmi, Susan Pintzuk, Helena Raumolin-Brunberg, Herbert
Schendl, Jeremy J. Smith, Erik Smitterberg, Anthony Warner, Irené Wother-
spoon, Laura Wright and Wim van der Wurff.
Thanks are also due to the editorial staff of John Benjamins, in particular
Konrad Koerner for his cooperation, and Anke de Looper for her good counsel
in response to our many queries. We owe a special debt of gratitude to Belén
Méndez-Naya and Elena Seoane, our friends and colleagues at the Department
of English of the University of Santiago de Compostela, for their close collabo-
ration with us during the editing process. Last but not least, we are grateful to
the Spanish Ministry of Science and Technology (grant no. BFF2001–2914) and
the Xunta de Galicia (grant no. PGIDT01PXI20404PR) for financial support,
and to our colleagues Paloma Núñez-Pertejo and Carla Dechant for their help
in various ways.

Santiago de Compostela, October 2001

The Editors
<TARGET "intro" DOCINFO

AUTHOR "Teresa Fanego"

TITLE "Introduction"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Introduction

Teresa Fanego
University of Santiago de Compostela

The Eleventh International Conference on English Historical Linguistics (11

ICEHL) was held at the University of Santiago de Compostela between 7th and
11th September 2000. The number of participants exceeded 250, while the
papers delivered within the conference’s main programme came to 120. The
distinguished panel of plenary speakers featured Douglas Biber, Laurel J.
Brinton, Santiago González Fernández-Corugedo, Raymond Hickey, Chris
McCully, Frans Plank, Irma Taavitsainen, Ingrid Tieken-Boon van Ostade and
Anthony Warner. There were also several events running concurrently with the
main programme, notably a workshop on historical word-formation, a para-
session on electronic corpora and a poster session.
This volume is a companion to another one also containing papers from the
same conference: Sounds, Words, Texts and Change. Selected Papers from 11
ICEHL, Santiago de Compostela, 7–11 September 2000, edited by Teresa Fanego,
Belén Méndez-Naya & Elena Seoane (CILT 224). The two volumes together
offer a representative sample of the contributions presented at the conference,
including some of those delivered during the workshop on historical word-
formation. The papers that have survived the successive selection procedures
for presentation and publication1 quite accurately reflect the various concerns
of English historical linguistics at the turn of the millennium and the different
methodologies applied to address them.
Though grouped together under the convenient heading of ‘Syntax and
Morphology’, the articles contained herein testify to the often noted fact (see
e.g. van Kemenade 1999: 1002) that some of the current approaches to linguis-
tic research — most notably grammaticalization theory — are increasingly
calling into question a number of the basic axioms of structural linguistics,
such as the notion of the discreteness of categories or the autonomy of the
domains of grammar. In several of the contributions morphology, syntax,
2 Teresa Fanego

semantics and communicative strategies of various kinds are seen as impinging

on one another (see e.g. the papers by Allen, Biber & Clark, Kornexl, Krug,
Lenker, Los, or Möhlig & Klages). Largely for this reason, we have made no
attempt to organize the papers thematically and have simply presented them in
alphabetical order. However, so as to give the reader some preliminary idea of
what this volume has to offer, we will give a brief summary of the main issues
in each individual paper.
Grammaticalization processes are discussed in the papers by Akimoto,
Brinton and Krug. Akimoto (“Two Types of Passivization of ‘V+NP+P’
Constructions in relation to Idiomatization”) examines the passivization of
composite predicates like make allowance for or put an end to on the basis of
extensive data from the seventeenth to the twentieth centuries and relates the
type of passivization allowed by each predicate to its degree of idiomatization
(Akimoto 1995). He shows that deverbal nouns with overt suffixes, such as
allowance, are less easily fused with their verbs than deverbal nouns without
suffixes, such as end, which accounts for their different behaviour regarding
facts of passivization: inner passives (“allowance ought to be made for the
Society”) are preferred by the former class of nouns; outer passives (“the pause
was put an end to”), by contrast, are more characteristic of nouns without
suffixes. Brinton’s article (“Grammaticalization versus Lexicalization Reconsid-
ered: On the late Use of Temporal Adverbs”) discusses a problem in the history
of English which seems to present a challenge for concepts such as grammatical-
ization and lexicalization: the use of temporal adverbs as attributive adjectives,
as in the often remembrance. Such usage was common in Early Modern English,
but for most of the forms it was transient, and they are found only adverbially
in Modern English. Brinton looks at possible ways of explaining this categorial
shift and, although she tends to view the change as a case of grammaticalization,
notes that doing so is not unproblematic: the forms discussed do not exhibit a
downgrading of categorial status (i.e. from more major to more minor catego-
ry) and hence do not meet the principle of decategorialization (Hopper 1991)
which is usually considered central to grammaticalization processes. Finally, in
“A Path to Volitional Modality” Krug traces the syntactic and semantic changes
undergone by the verb want from its early use as a transitive verb meaning ‘lack’
(“the fool wants wit”), and later ‘need’ (“this room wants cleaning”) and
‘desire’ (“I want an apple”), to its use as a catenative encoding modal meanings
like volition (“I want to go to the cinema”) or, more recently, obligation (“you
want to be careful”). He argues that this semantic development was motivated
by the gradual conventionalization of pragmatic inferences (Hopper & Traugott
 Introduction 3

1993: 63ff.) in appropriate syntactic contexts: what is lacked by somebody will

often be needed, and hence desired — from which it is only a small step to the
volitional reading where what is desired is no longer a concrete nominal object
but an action or a state of affairs. Semantically, want has therefore become
increasingly abstract and more grammatical. Syntactically, the want + to-infin-
itive construction, which involves the concatenation of two verb phrases, is also
more grammaticalized than the original transitive verb governing a NP object.
Noun phrase structure is the topic of two papers. In “On the Development
of a friend of mine” Allen discusses the development of the double genitive
construction. She traces the first examples of expressions like “a friend of mine”
to the middle of the fourteenth century, and argues that the precursor of the
double genitive can be found in Early Middle English constructions without a
nominal head like Šif þu mare spenest of þine ‘if you spend more of yours’ (i.e.
‘of your money’); these constructions, which involve partitive of and the
absolute use of a possessive pronoun, make reference to ‘part of a set’. From this
the meaning shifted to the ‘membership in a set’ meaning that characterizes the
noun-headed double genitive. In turn Biber & Clark (“Historical Shifts in
Modification Patterns with Complex Noun Phrase Structures: How Long Can
You Go without a Verb?”) examine patterns of noun modification in English
from 1650 to the present. They show that over the past 100 years (especially the
past 50 years in the case of noun–noun sequences) a major historical shift has
taken place to favour the use of nouns as premodifiers (which are now nearly as
important as attributive adjectives) and of prepositional phrases (rather than
relative clauses) as postmodifiers. One major consequence of this historical shift
is the development of a much more compressed, less explicit style of presenta-
tion which, Biber & Clark suggest, may have been facilitated by functional
factors like recent advances in the technology of literacy, coupled with the
‘informational explosion’ resulting in pressure to communicate information as
economically as possible.
The paper by Los (“The Loss of the Indefinite Pronoun man: Syntactic
Change and Information Structure”) analyses how the indefinite pronoun man
‘one’ fell into disuse in Late Middle English and relates this to the loss of V2 and
to the spread of to-infinitival complementation. The loss of V2 affected the
information structure of English main clauses: with the generalization of SV,
subject NPs — unless preceded by another clause constituent fronted for special
discourse effects (e.g. “tennis I like”) — came to function as unmarked themes
(cf. Halliday 1994) encoding given information and maintaining textual cohesion.
This left little scope for the indefinite man, whose main role had been to
4 Teresa Fanego

provide a contentless subject, and promoted the use of impersonal (i.e. agent-
less) passives, which increasingly took over the function of man in main clauses.
In the transition from Old English to Middle English there was also a decline of
man in dependent clauses as a result of competition between subjunctive
that-clauses and to-infinitives after certain verbs, which led to man in such
clauses being largely ousted by generic PRO (e.g. ic æfre bebead [PROgen] þone
drihtelican dæg to healdenne ‘I have ever ordered to keep the Lord’s day’).
Schlüter’s paper (“Morphology Recycled: The Principle of Rhythmic
Alternation at Work in Early and Late Modern English Grammatical Varia-
tion”) is concerned with the so-called Principle of Rhythmic Alternation (cf.
Selkirk 1984: 37: “[t]here is arguably a universal rhythmic ideal, one that favors
a strict alternation of strong and weak beats”), whose effects she illustrates with
reference to a number of grammatical variation phenomena taking place chiefly
in the Early and Late Modern English periods, such as the distribution of
mono- and disyllabic past participle variants (drunk/drunken), or the variable
marking of infinitives dependent on the verb make in the passive (e.g. made to
glisten / made draw water). All her analyses are concerned with intermediate
phases of language change in which morphemes and markers are no longer or
not yet quite obligatory in terms of grammatical motivations. Schlüter shows
that it is precisely in these phases of indeterminacy that the Principle of Rhyth-
mic Alternation may assume the role of an influential determinant.
An area of research pursued with great interest at the 11th ICEHL was
word-formation processes. Three of the papers in this line — by Kornexl,
Lenker and Kastovsky — appear in this volume. Kornexl (“From gold-gifa to
chimney sweep? Morphological (Un)markedness of Modern English Agent
Nouns in a Diachronic Perspective”) examines nominal formations and
challenges assumptions about the diachronic continuity from Old English times
of agentive zero-derivation: according to such assumptions, nouns like cheat,
cook or chimney sweep would be the Modern English analogues of Old English
zero derivatives like bora ‘bearer’ or gold-gifa ‘gold-giver’. Kornexl argues,
however, that English has never developed a productive type of zero-derived
agentives; alleged zero-derived agentive nouns are either loans (cook, guide,
judge) or members of a special category of formally unmarked ‘attitudinal’
nouns (cheat, bore, tramp) reflecting the speaker’s attitude towards certain
people or things and thus crucially differing, both semantically and pragmati-
cally, from prototypical agentives. Lenker’s paper (“Is It, Stylewise or Otherwise,
Wise to Use -wise? Domain Adverbials and the History of English -wise”)
discusses the recent emergence of viewpoint adverbials in -wise (e.g. “there are
 Introduction 5

two types of hydrogen atoms positionwise”). After an examination of the

distribution and productivity of sentence adverbial -wise in Present-day English,
she connects its emergence to the history of viewpoint adverbials in German
and English. Viewpoint adverbials are a new category of sentence adverb not
attested until the nineteenth century and apparently arising in scientific
registers as a way of indicating the speaker’s perspective without having to name
the speaker directly (e.g. “botanically, this is the region of plants”). English
-(c)ally is recorded in this use from the second half of the nineteenth century
onwards, but is restricted to Latin or Greek adjectival bases. In the course of the
twentieth century, -wise, which unlike -cally can be combined with nominal
roots irrespective of their etymological origin, came to fill this obvious gap in
word-formation. Finally, in “The Derivation of Ornative, Locative, Ablative,
Privative and Reversative Verbs in English: A Historical Sketch”, Kastovsky
examines the historical development of five classes of verbs sharing a common
cognitive-semantic basis and using the same derivational means: the underlying
meaning of ornative (encrown), locative (encage), ablative (unsaddle), privative
(behead) and reversative (unlock) verbs is that some Theme is caused to be
located in some Location (e.g. encage) or is removed from this Location (e.g.
unseat). Kastovsky shows that, with the exception of ablative verbs, these
semantic categories were well established in Old English. An ablative pattern
with un-, as in unsaddle or unearth, emerged in Middle English, probably under
the influence of ablative dis-formations borrowed from French, such as dislodge.
In addition, the introduction in Middle and Early Modern English of other
Romance/Latin prefixes (de-, en-/em-) generally strengthened the whole
derivational set.
Verbs are also the focus of Möhlig & Klages’ “Detransitivization in the
History of English from a Semantic Perspective”. They adopt a functional
framework (Dik 1989, 1997) to look into the historical development from Old
English times of four different uses of selected transitive verbs which result in
non-transitive constructions. The four patterns of detransitivization in question
are co-referential intransitives (John washed), ergatives (the door opened),
generics (he is wise who reads) and middles (this book reads well). Starting from
the functionalist assumption that “the properties of clause structure are
predictable from the semantics of predicates” (Faber & Mairal Usón 1999: 37),
Möhlig & Klages examine the semantic properties of the verbs involved in these
non-transitive presentations of essentially transitive events.
With the paper by Meurman-Solin we move away from Standard English
into the fields of dialectology and sociolinguistics. In “The Progressive in Older
6 Teresa Fanego

Scots” she discusses the use of the progressive in Scots from 1450 onwards, in
particular the variants be doand, be doing and be in/a doing, and the linguistic,
idiolectal and genre-specific factors affecting their distribution. The frequency
of the progressive correlates with the type of discourse: narratives and speech-
based texts yield a generally greater number of occurrences of the construction.
Regarding the possible influence of Celtic languages on the development of the
progressive in English,2 Meurman-Solin does not find evidence supporting the
existence of a ‘Celtic connection’ and points out that, despite the size of her
corpus (over 850,000 words of running text), to say anything conclusive in this
respect an even larger database would be necessary. Among other interesting
findings, she suggests that the prepositional type be in/a doing may have
provided a structure with a passive sense (i.e. ‘be being done’) at a time when
the verbal types did not yet have a passive transform.
We would like to close this brief introduction by thanking the many people and
institutions that helped to make the 11th ICEHL a success. Among the former,
we are grateful to all those who delivered papers, as well as to the several
academics who helped us in the difficult task of selecting from the large number
of valuable abstracts submitted the contributions that were accepted for
presentation at the conference. Our thanks also to the students who collaborat-
ed with the Organizing Committee both before and during the conference.
Sponsorship was gratefully received from the Xunta de Galicia (Secretaría Xeral
de Investigación e Desenvolvemento and Dirección Xeral de Turismo), the
Spanish Ministry of Education and Culture, the University of Santiago de
Compostela, the Department of English, the City of Santiago de Compostela,
the British Council, the Spanish Association for Canadian Studies (AEEC), and
the Fundación Pedro Barrié de la Maza.

Notes

1. The selection process was not an easy task, for the number of papers submitted for
publication came to 55.
2. On this issue see Mittendorf & Poppe (2000) and Vennemann (2001).
</TARGET "intro">

Introduction 7

References

Akimoto, Minoji. 1995. “Grammaticalization and Idiomatization”. The Twenty-first LACUS

Forum 1994 ed. by Mava Jo Powell, 583–590. Chapel Hill, NC: LACUS.
Dik, Simon. 1989. The Theory of Functional Grammar. Part I: The Structure of the Clause.
Dordrecht: Foris.
Dik, Simon. 1997. The Theory of Functional Grammar. Part I: The Structure of the Clause /
Part II: Complex and Derived Constructions. 2nd rev. ed. by Kees Hengeveld. Berlin &
New York: Mouton de Gruyter.
Faber, Pamela & Ricardo Mairal Usón. 1999. Constructing a Lexicon of English Verbs. Berlin
& New York: Mouton de Gruyter.
Halliday, M. A. K. 1994 [2nd rev. ed.]. An Introduction to Functional Grammar. London:
Edward Arnold.
Hopper, Paul J. 1991. “On Some Principles of Grammaticization”. Approaches to Grammatic-
alization ed. by Elizabeth Closs Traugott & Bernd Heine, vol. I, 17–35. Amsterdam &
Philadelphia: John Benjamins.
Hopper, Paul J. & Elizabeth Closs Traugott. 1993. Grammaticalization. Cambridge: Cam-
bridge University Press.
Kemenade, Ans van. 1999. “Functional Categories, Morphosyntactic Change, Grammatical-
ization”. Linguistics 37.997–1010. Special issue on Functional Properties of Morpho-
syntactic Change ed. by Ans van Kemenade.
Mittendorf, Ingo & Erich Poppe. 2000. “Celtic Contacts of the English Progressive?” The
Celtic Englishes II ed. by Hildegard L. C. Tristram, 117–145. Heidelberg: Carl Winter.
Selkirk, Elizabeth O. 1984. Phonology and Syntax: The Relation between Sound and Structure.
Cambridge, Mass. & London: MIT Press.
Vennemann, Theo. 2001. “Atlantis Semitica: Structural Contact Features in Celtic and
English”. Historical Linguistics 1999. Selected Papers from the 14th International Confer-
ence on Historical Linguistics, Vancouver, 9–13 August 1999 ed. by Laurel J. Brinton,
351–369. Amsterdam & Philadelphia: John Benjamins.
<LINK "aki-n*">

<TARGET "aki" DOCINFO

AUTHOR "Minoji Akimoto"

TITLE "Two types of passivization of ‘V+NP+P’ constructions in relation to idiomatization"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Two types of passivization of ‘V+NP+P’

constructions in relation to idiomatization*

Minoji Akimoto
Aoyama Gakuin University

1. Introduction

This paper examines so-called composite predicates1 of the form ‘V+NP+P’

(e.g. lose sight of, take care of) in relation to passivization. The structures under
analysis involve a verb of general actional meaning (do, give, make, take), a
(usually) deverbal noun (charge, leave, advantage) and a following prepositional
phrase, as in (1):
(1) We took advantage of the students.

As is well known, English allows two possible types of passive for ‘V+NP+P’
constructions, as illustrated in (2)–(3):
(2) Advantage was taken of the students.
(3) The students were taken advantage of.

Following Nunberg et al. (1994), I will call the former passive the ‘inner
passive’, and the latter the ‘outer passive’. In a historical perspective, these two
types of passivization present interesting problems. First, which type appeared
first? Second, which type is more frequent with the ‘V+NP+P’ construction?
Third, how is the development of these types of passivization related to the
process of idiomatization (Akimoto 1995: 588)?
The aim of this paper is to discuss the issues raised above on the basis of
examples collected from a variety of written sources (see the list of texts at the
end of this paper), and also from the Oxford English Dictionary on CD-ROM
(OED) and from the COBUILD CD-ROM (for Present-day English). Previous
studies (see Section 2 below) suggest that the Late Modern English period
witnessed an increase in the passivization of ‘V+NP+P’ constructions, hence in
10 Minoji Akimoto

what follows I will concentrate on the seventeenth, eighteenth, nineteenth and

early twentieth centuries.2
The texts examined date from the seventeenth to the early twentieth
centuries, and comprise fiction, essays, letters and drama among others. In
terms of degree of formality, the texts can be considered representative of
various styles, ranging from more (e.g. essays) to less formal (e.g. letters). Since
the examples gleaned from my corpus were quite limited, I made use of the
OED and COBUILD CD-ROMs for additional data. The total number of words
of the texts examined in each century is as follows: (1) 17th century: around
873,600 words; (2) 18th century: around 609,700 words; (3) 19th century:
around 925,500 words; (4) 20th century: around 568,800 words.
The OED represents a fairly random collection of data collected for no
specific grammatical purpose, but even so, it can still be interpreted as showing
general trends. Furthermore, my own collection of data, as listed below, can
remedy that arbitrariness to a certain extent.

2. Previous studies

Poutsma (1926: 118–122) gives four possibilities of passivization for ‘V+NP+P’

constructions:
a. Passivization of the (pro)noun governed by the preposition as the only
possible passive (the outer passive). E.g. some things had been lost sight of.
b. Passivization of the noun phrase as the only possible passive (the inner
passive). E.g. every allowance is made for difficulties.
c. Admission of two passive constructions. E.g. new means must of necessity be
had recourse to / recourse was had to the present participle.
d. No passive conversion, as is the case with the composite predicates pay
court to, set foot on or take leave of.
Jespersen (1927: 315–317) also gives examples of ‘V+NP+P’ passivization. His
examples are mostly from the eighteenth and nineteenth centuries, except
Mandeville and North. The type of passivization illustrated by Jespersen is the
outer passive; he adduces no examples of the inner passive.
Denison (1993:153–154) gives examples of the outer passive, such as be taken
heed to and be made mind of, in the fifteenth century. He gives no examples of
the inner passive. In an earlier study (1985: 202), he points out that the first
“safe example” of the outer passive known so far dates back to the fifteenth century.
 Two types of passivization of ‘V+NP+P’ constructions 11

Matsumoto (1999: 88–89) and Tanabe (1999: 119–120) briefly discuss the
passivization of ‘V+NP+P’ constructions in the Middle English period. Matsu-
moto states that the inner passive is common and that the outer passive is rare, but
she does not adduce any examples of her own. Tanabe, like Matsumoto, does
not make a distinction between deverbal nouns with and without suffixes (e.g.
appointment vs. sight), a crucial issue to which I will return later in this section.
This latter observation is also true of Claridge (2000). She examines three
groups of verbo-nominal combinations in the Lampeter Corpus (1640–1740).
The second group consists of ‘verb+noun+preposition’ units, such as set fire to
and take care of. She refers to the two types of passivization allowed by such
structures in various places (80, 161, 276) and, without drawing a distinction
between deverbal nouns with and without suffixes, she suggests that there is a
predominance of the inner passive over the outer passive.
In his comprehensive analysis of John Dryden’s verb syntax, Söderlind
(1951: 26–27) notes that “[t]he construction verb + object + prepositional
adjunct turned into the passive is extremely rare (…) The type was apparently
only just beginning to expand”. He has found only three instances — all of
them of the outer passive — with the predicates do good to, take hold of and
take notice of.
Visser (1973: 2163–2176) discusses the passivizability of ‘V+NP+P’ con-
structions extensively, adducing numerous examples, mostly of the outer
passive. With regard to the different passivization possibilities of the predicates
listed, he points out that “[i]t is not easy to account for this divided usage”.
Though valuable, Visser’s examples cannot be taken uncritically, since not all of
them can be accepted as genuine instances of the pattern under analysis. In any
case, except for a few examples such as those involving find fault with (1400;
with the inner passive) or take heed at/about/to (1380; with the outer passive),
most date from the seventeenth century onwards.
In their study on idioms, Nunberg et al. (1994) refer to “double passives”,
that is, the possibility of inner and outer passives with the same construction.
They call our attention to the following points:
a. The outer passive is vastly more frequent than the inner passive: in an
extensive database of newspaper texts they found 1,200 examples of outer
passives with take advantage of, as against only 71 examples of inner passives
with this same predicate (1994: 521).
b. In the examples of inner passive, the head houn is most often preceded by
adjectives or quantifiers (47 out of the 71 examples involving advantage was
taken of).
12 Minoji Akimoto

c. Generally speaking, idioms with only outer passives are semantically more
opaque than idioms with only inner passives, since the former are treated as
phrases that have been lexicalized into verbs, whereas the latter are typically
conventionalized metaphors whose parts have identifiable meanings (1994:
524). For example, take hold of, which allows only the outer passive, means
roughly ‘grasp’ and it is difficult to see what part of this interpretation can be
assigned to hold.
d. Double passives are not very common, constituting less than 10% of the 97
idioms involving make and take.
With regard to these various studies on the passivization of composite predi-
cates, there are several aspects which are worth noting. First, none of the
scholars referred to above describes in any detail the characteristics of the
‘V+NP+P’ constructions they examine. Yet an aspect of prime importance is
whether the preposition is obligatory or not. In this respect, Nunberg et al.
(1994) seem to have mixed genuine ‘V+NP+P’ structures with those which are
not, because they include constructions in which the preposition is optional.
For example, make an appointment (with), make an impression (on), make
arrangements (for/with), take a chance (on), take a look (at) and take a stab (at).
Secondly, no distinction is made in previous studies between deverbal
nouns with suffixes and those without suffixes. Nouns participating in the
construction under analysis are mostly deverbal nouns; those with suffixes,
such as calculation or preference, usually allow only the inner passive. This is
partly because the overt nominal suffix increases the nouniness of the deverbal
noun, which consequently resists the outer passive. By contrast, deverbal
nouns without suffixes (or zero derivational nouns) often allow the outer
passive, although at times they allow the inner passive as well. In this study, I
will adopt a narrow definition of ‘V+NP+P’ constructions which includes only
nouns without suffixation, although for comparison I give examples with
suffixation in Table 1.
Finally, no less important is the fact that the presence of indefinite articles
is strongly related to the passivizability of these nouns. Generally speaking,
nouns without indefinite articles, such as catch sight of and take care of, are
indicative of the fixed unity of the phrases in question, and hence are more
susceptible to the outer passive than nouns with indefinite articles.
 Two types of passivization of ‘V+NP+P’ constructions 13

Table 1. Frequency of passives from the 17th to the 20th centuries

17th 17th 18th 18th 19th 19th 20th 20th
inner outer inner outer inner outer inner outer

catch hold of 1 (2)

do harm to 1 1 (2)
do justice to 1 2 (2) 1 (3) (1)
find fault with 3 1 (1) 1 (1) (2) (1)
give countenance to (1) 1
give preference to 1 (1) (8) (8)
have attention to 1 (1)
have recourse to (1) 2 (4) (1) (13) (11) (3)
have regard to (3) 1 (7) (6) (1)
lay hands on 1 (1)
lay hold of 1 (1) 2 (3)
lose sight of (1) 1 (10) (3)
make addresses to 1 (1)
make allowance for (2) 3 (2) (26) (8)
make allusion to 2 (7)
make a calculation of 1 (1)
make choice of 1 (1) (2) (1) (1)
make complaint of 2
make conjecture of 1
make estimate of (1) 1
make an example of 1 (1) (1) (1)
make a fool of (2) 1 (1) 2 (2) 2 (2)
make mention of 5 (5) 1 (6) (13) (5)
make preparations for 1 (5) (6)
make provision for 1 (5) 1 (4) (21) (30)
make use of 3 (11) 10 (23) 4 (18) 7 (100) 1 (21) 1 (65) (47) (12)
pay attention to 1 (7) (25) (30)
pay regard to 1 (1) (6)
put an end to (1) (2) (2) (1) 5 (5)
put a stop to 1 (7) (2) 1 (4) (3) (1) (1)
set bounds to (1) 1
set one’s mind on 1
take advantage of 1 (1) 1 (14) 7 (7) (6)+1 (6)+7
take care of 2 (3) 1 4 (8) 5 (5) +2 (6)+65
take a dislike against 1
take leave of 1
take measure of 1
take notice of 5 (4) 13 (14) 2 (5) 11 (13) (4) (1)+3 +1
take pity on 1
Totals 70 69 94 157 188 141 157 105
14 Minoji Akimoto

3. Analysis and discussion

3.1 Distribution of passivizations in the data

The passives in Table 1 have been obtained from the various sources examined;
figures for passives in the OED CD-ROM are indicated in parenthesis. For the
high frequency phrases take advantage of, take care of and take notice of, I have
added data from the COBUILD CD-ROM.
It seems that the general trend is for the inner passive to be more frequent
than the outer passive except in the eighteenth century. However, one has to
take into consideration the number of deverbal nouns with suffixes in the
‘V+NP+P’ constructions listed. The ‘V+NP+P’ patterns which allow both inner
and outer passives are usually limited to nouns without suffixation. Nouns with
suffixation usually allow only the inner passive. As I will discuss later, the
nominals forming the more idiomatic phrases are those without suffixes.
Deverbal nouns with suffixes, because of their strong nouniness, are not
suitable as constituents in idiomatic phrases, their fusion into the verbal phrase
as a whole being prevented. If we exclude from the list the predicates containing
deverbal nouns with suffixes, such as do justice to, give countenance to, give
preference to, have attention to, make allowance for, make allusion to, make a
calculation of, make complaint of, make preparations for, make provision for and
pay attention to, then the totals for the inner and outer passives in each century
are as shown in Table 2.3

Table 2. Revised frequency of passives in each century

17th 17th 18th 18th 19th 19th 20th 20th
inner outer inner outer inner outer inner outer

57 69 66 157 89 141 74 105

The revised Table 2 shows that, while the frequency of inner passives
remains constant through time, outer passives have been predominant particu-
larly since the eighteenth century onwards, a trend which corresponds to the
process of idiomatization to be discussed later.
Because of space limitations, I can only give a few examples of the passive
structures occurring in my data:
(4) great justice was done to the collation by the guests in general (Smollett 243)
 Two types of passivization of ‘V+NP+P’ constructions 15

(5) a. and so pleasing is the snare as, till it hath ruined one, no fault is
found with it (Anthology of Seventeenth-Century Fiction 176)
b. Lucan is found fault with for not writing according to the lawes of a
Poeme (Conway Letters 31)
(6) Be that as it may, recourse will always be had to this place (Smollett 181)
(7) There is also another consideration not to be lost sight of (Mill 212)
(8) much allowance ought to be made for the Society (Burke 72)
(9) though it was possible that some passing allusion might still be made to
the hospital (Trollope 100)
(10) Complaint being made yesterday of the great abuse in selling of offices &
places of trust (Marvell 35)
(11) a. some severe example might be made of such who (…) (Marvell 187)
b. They must be made an example of. (1803 Pic Nic No. 4 (1806) I.140;
OED s.v. example n. 3)
(12) a. No mention hath yet been made of Sr John Coventryes misfortune
(Marvell 124)
b. whereas in mine to Him was made mention of a cap and a paire of
pistole barrells (Conway Letters 284)
(13) a. no advantage may be taken of your proceedings hitherto (Marvell 84)
b. As the variety of each species (…) will be taken advantage of by natu-
ral selection (Darwin 348)

By comparison with Present-day English, the following aspects deserve mention:

a. It appears that the phrases make use of and take notice of were originally
more frequently used in the outer passive, whereas in Present-day English the
inner passive is more common. On the other hand, take care of and take
advantage of show a shift in frequency from the inner to the outer passive (see
also fn. 4 below).
b. The phrase have recourse to allowed both types of passivization, but in
Present-day English neither passive is possible (see Cowie & Mackin 1975: 154).
c. In Visser’s examples, lose sight of and make love to permitted both types of
passive, but in Present-day English only the outer passive is possible. On the
other hand, take account of, which is not recorded in my material but appears in
Visser (1973: 2174), permitted only the inner passive, but now allows two types
of passive (see Cowie & Mackin 1975: 324).
16 Minoji Akimoto

d. The phrase make mention of is used mostly in the inner passive. As a phrase
containing a noun with a zero suffix, this phrase is slightly exceptional, because
nouns without a suffix such as mention usually allow the outer passive. Inciden-
tally, make mention of was modelled on the French expression ‘faire mention de’
(Prins 1952: 206), and French influence may have had some connection with
the use of the inner passive.
e. Phrases containing deverbal nouns with suffixation, such as attention,
preference, allowance and all those listed in Table 1, permit only the inner
passive. That is, the deverbal noun becomes the subject.

4. Idiomatization

Idiomatization can be described as the linguistic process, both synchronic and

diachronic, of reorganizing certain phrases into fixed / fossilized expressions,
whose meanings cannot be deduced from their constituents (Akimoto 1999:
225; see also Akimoto 1995: 588). Idiomatization usually involves semantic
change, lexical fixing and syntactic ossification. I consider the development of
‘V+NP+P’ constructions as a four-stage process of idiomatization, as follows
(cf. Akimoto 1989: 354–358):
a. in the first stage, all of the constituents are unrestricted;
b. in the second stage, the relation between the verb and noun becomes
stabilized, the verb and preposition become fixed, and the noun loses some
of its nominal features (number, definite/indefinite articles), i.e. becomes
decategorialized;
c. in the third stage, reanalysis of the constituent structure occurs; and
d. in the fourth stage, all of the constituents are idiomaticized into a single
lexical item.
Let us take lose sight of as an example. In the first stage, there is variability in the
verbs that can collocate with the noun sight (OED s.v. sight n.1):
have/lose/take + the + (modifier) + sight + of
The phrase occurred with a definite article and was used for anything that can
be seen:
(14) They loste ye syght of ye castell, it was clene vanysshyd a way (a1533 Ld
Berners Huon xxiii 68; OED s.v. sight n.1 II.4.a)
 Two types of passivization of ‘V+NP+P’ constructions 17

In the second stage, the verb lose becomes more strongly collocated with sight
than the other verbs, the definite article the being sometimes deleted. In the
third stage, reanalysis takes place, as follows:
[lose] [sight of X] > [lose sight of] [X]
In the fourth stage, lose sight of is idiomaticized into a single lexical item. The
phrase can now be used for something unseen:
(15) There is also another consideration not to be lost sight of. (Mill 212)

A similar process can be observed in the case of take advantage of. The noun
advantage, unlike its Present-day English counterpart, was formerly a full noun
that could take modifiers, definite/indefinite articles, and plural forms (see
Akimoto 1999: 212–213, 226), and could also be followed by a variety of
different prepositions (e.g. to, on, of). Example (16) below illustrates stage 3
(reanalysis of the constituent structure), as described above:
(16) they may have means on any emergency occasion, or sudden need, to
resist, or take advantage on their Enemies (Hobbes, Leviathan 239)

Thus, semantically idiomatization is a process whereby meaning becomes more

opaque, and, syntactically, the phrasal unity becomes more fixed than before. In
this respect, take care of and take advantage of, with both of which the outer
passive is now predominant, have become more idiomatized than make use of
and take notice of.4 Other phrases, such as lose sight of and make love to (the
latter not recorded in my material), show a higher degree of idiomaticity in
Present-day English; these phrases used to permit both types of passivization
(see Visser 1973: 2168, 2170), but in Present-day English they permit only the
outer passive.5 Have recourse to, which formerly allowed both the inner and
outer passives, does not allow either type of passivization in Present-day
English.6 This could be interpreted as indicating a high degree of idiomaticity,
since idioms usually resist grammatical transformations such as passivization
and topicalization (see Chafe 1968). But a more general principle may also be
at work. Because of its stative nature in Present-day English, the verb have
cannot be passivized. Formerly, however, in composite strings with have this
verb seems to have had a relatively dynamic meaning, as discussed by Traugott
(1999:255). Thus the non-passivizability of have-phrases (in this respect see also
have attention to and have regard to in Table 1) may be a consequence of the
change of this verb from dynamic to stative.
18 Minoji Akimoto

As pointed out above, of the ‘V+NP+P’ patterns take care of, take advantage
of, take notice of and make use of, the first two have gone in different directions
from the second two in terms of idiomatization. Take care of and take advantage
of have shifted their frequency from the inner to the outer passive, and take
notice of and make use of from the outer to the inner passive. This difference is
partly attributable to the strength of the fusion of the noun into the verb in the
phrase. The nouns care and advantage are more strongly fused into their verb
phrases than notice and use in theirs. The degree of fusion can be largely
determined by whether the noun can be modified or not (cf. Nunberg et al.
1994: 500–503). The nouns in take care of and take advantage of are normally
used with no or a limited number of modifiers,7 whereas the nouns in take
notice of and make use of are usually accompanied by modifiers. In this way, the
idiomatic bond is weakened, so that, as pointed out by Quirk et al. (1985: 1160),
in idiomatic phrases where the object contains a modifier or determiner, it is
easier “to separate the object from the rest of the construction by the regular
passive transformation” (e.g. some notice was taken of …).
Also characteristic of idiomatization is the gradual abstractness of nouns, as
illustrated above with the noun sight in lose sight of. This direction from
concreteness to abstractness is very common, and the development of outer
passives matches this direction: an abstracted noun that has lost such nominal
features as the ability to take an indefinite article can easily be fused into the
verb phrase, which promotes the possibility of the outer passive.
It should be borne in mind, however, that some composite predicates have
failed to move in the expected direction. Thus, as pointed out above, in the case
of make use of inner passives have grown more frequent in the course of time,
rather than the other way round. This is an issue that clearly deserves further
investigation.

5. Conclusion

I have discussed the two types of passivization allowed by ‘V+NP+P’ constructions

— the inner passive and the outer passive — from the seventeenth to the twentieth
centuries. Prior research on this topic has failed to make a distinction between
phrases involving nouns with suffixes and those without suffixes, yet it has been
shown here that establishing such a distinction is crucial, as nouns like e.g.
calculation or sight behave differently with regard to facts of passivization.
<DEST "aki-n*">

Two types of passivization of ‘V+NP+P’ constructions 19

Another finding of the present study is that since at least the seventeenth
century the outer passive has been more frequent than the inner passive, a trend
that matches the increased process of idiomatization of the structures under
analysis. While this is true of most of the idioms examined, phrases like make
use of and take notice of seem to resist the direction of idiomatization.

Notes

* I am grateful to Laurel Brinton, Teresa Fanego and an anonymous referee for helpful
comments and suggestions on earlier versions of this paper. Thanks are also due to Elena
Seoane-Posse and Teresa Moralejo-Gárate for sending me copies of Seoane-Posse’s research
on the passive.
1. For the label cf. Cattell (1984).
2. I tried to find examples of these types of passivization in sixteenth-century authors, such
as Deloney, Nashe and Sidney, but did not come across any. This result may be natural in
view of the development of the passive construction. The passive voice only became frequent
from Early Modern English onwards (Görlach 1991: 116; Seoane-Posse 1996: 222–223, 407;
1999: 134–136), and it seems reasonable that the more ‘marked’ passive construction of
‘V+NP+P’ should have appeared even later.
3. I excluded choice and sight from the category of deverbal nouns with suffixation.
4. Although in Table 1 there seems to be no clear difference in the frequency of the inner and
outer passives with take advantage of, Nunberg et al. (1994: 521) found 1,200 examples of
outer passives, as against 71 examples of inner passives.
5. As regards the passivization of make love to, I am relying on Nunberg et al.’s (1994: 533)
judgement.
6. See Cowie & Mackin (1975: 154), although Poutsma (1926:121) gives two types of passive.
7. As regards the modifiers of advantage in the phrase take advantage of, see Nunberg et al.
(1994: 522). With respect to the modifiers of care in take care of, in COBUILD I recorded a
total of 67 passive examples, of which only one — with the modifier great — was an example
of the inner passive.

References

Akimoto, Minoji. 1989. A Study of Verbo-Nominal Structures in English. Tokyo: Shinozaki

Shorin.
Akimoto, Minoji. 1995. “Grammaticalization and Idiomatization”. The Twenty-first LACUS
Forum 1994 ed. by Mava Jo Powell, 583–590. Chapel Hill, NC: LACUS.
20 Minoji Akimoto

Akimoto, Minoji. 1999. “Collocations and Idioms in Late Modern English”. Brinton &
Akimoto 1999. 207–238.
Brinton, Laurel & Minoji Akimoto, eds. 1999. Collocational and Idiomatic Aspects of Composite
Predicates in the History of English. Amsterdam & Philadelphia: John Benjamins.
Cattell, Ray. 1984. Syntax and Semantics. Vol. 17: Composite Predicates in English. Sidney,
Australia: Academic Press.
Chafe, Wallace L. 1968. “Idiomaticity as an Anomaly in the Chomskyan Paradigm”.
Foundations of Language 4.102–127.
Claridge, Claudia. 2000. Multi-word Verbs in Early Modern English. Amsterdam & Atlanta:
Rodopi.
COBUILD Written and Spoken Corpus. Birmingham.
Cowie, A. P. & R. M. Mackin. 1975. Oxford Dictionary of Current Idiomatic English. Vol. I:
Verbs with Prepositions and Particles. Oxford: Oxford University Press.
Denison, David. 1985. “Why Old English Had No Prepositional Passive”. English Studies
66.189–204.
Denison, David. 1993. English Historical Syntax: Verbal Constructions. London & New York:
Longman.
Görlach, Manfred. 1991. Introduction to Early Modern English. Cambridge: Cambridge
University Press.
Jespersen, Otto. 1927 [repr. London, 1961]. A Modern English Grammar on Historical
Principles. Part III: Syntax (Second Volume). Copenhagen: Ejnar Munksgaard / London:
George Allen & Unwin.
Matsumoto, Meiko. 1999. “Composite Predicates in Middle English”. Brinton & Akimoto
1999.59–95.
Nunberg, Geoffrey, Ivan A. Sag & Thomas Wasow. 1994. “Idioms”. Language 70.491–538.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S. C. Weiner. Oxford: Oxford University Press.
Poutsma, Hendrik. 1926. A Grammar of Late Modern English. Part 2: The Parts of Speech,
Section 2: The Verb and the Particles. Groningen: P. Noordhoff.
Prins, Anton Adriaan. 1952. French Influence in English Phrasing. Leiden: Universitaire Pers
Leiden.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive
Grammar of the English Language. London & New York: Longman.
Seoane-Posse, Elena. 1996. The Passive Voice in Early Modern English. A Corpus-Based Study.
Santiago de Compostela: Universidade de Santiago de Compostela (Servicio de
Publicacións e Intercambio Científico).
Seoane-Posse, Elena. 1999. “The Consolidation of the Indirect and Prepositional Passive in
Early Modern English: Evidence from the Helsinki Corpus”. Estudios Ingleses de la
Universidad Complutense 7.119–139.
Söderlind, Johannes. 1951. Verb Syntax in John Dryden’s Prose. Vol. I. Uppsala: A.-B. Lundequist.
Tanabe, Harumi. 1999. “Composite Predicates and Phrasal Verbs in The Paston Letters”.
Brinton & Akimoto 1999.97–132.
Traugott, Elizabeth C. 1999. “A Historical Overview of Complex Predicate Types”. Brinton
& Akimoto 1999.239–260.
 Two types of passivization of ‘V+NP+P’ constructions 21

Visser, Frederikus Theodorus. 1973. An Historical Syntax of the English Language. Part III,
Second Half: Syntactical Units with Two and with More Verbs. Leiden: E. J. Brill.

Sources

An Anthology of Elizabethan Prose Fiction. 1573–1588 (1987). Ed. by Paul Salzman. Oxford:
Oxford University Press.
An Anthology of Seventeenth-Century Fiction. 1621–1698 (1991). Ed. by Paul Salzman.
Oxford: Oxford University Press. [c. 199,000 words]
Austen, Jane. 1818 (1998). Northanger Abbey ed. by John Davie. Oxford: Oxford University
Press. [c. 153,400 words]
Bennet, Arnold. 1908 (1995). The Old Wives’ Tale ed. by Margaret Harris. Oxford: Oxford
University Press. [c. 94,000 words]
Bunyan, John. 1678 (1965). The Pilgrim’s Progress ed. by R. Sharrock. Harmondsworth:
Penguin. [c. 116,900 words]
Burke, Edmund. 1790 (1993). Reflections on the Revolution in France ed. by L. G. Mitchell.
Oxford: Oxford University Press. [c. 104,400 words]
Butler, Samuel. 1660s (1979). Prose Observations ed. by Hugh De Quehen. Oxford: Claren-
don. [c. 75,000 words]
Butler, Samuel. 1903 (1966). The Way of All Flesh ed. by James Cochrane. Harmondsworth:
Penguin [c. 48,900 words]
Carlyle, Thomas. 1837 (1989). The French Revolution ed. by K. J. Fielding and David
Sorensen. Oxford: Oxford University Press. [c. 81,600 words]
Caxton’s Blanchardyn and Eglantine. c. 1489 (1962). Ed. by Leon Kellner. London, New York
& Toronto: Oxford University Press.
Chesterfield, Philip Dormer Stanhope, Earl of. 1774 (1992). Lord Chesterfield’s Letters ed. by
David Roberts. Oxford: Oxford University Press. [c. 103,400 words]
The Conway Letters. 1642–1684 (1992). Ed. by Marjorie Hope Nicolson. Oxford: Clarendon.
[c. 123,800 words]
The Correspondence of Robert Dodsley 1733–1764. 1988. Ed. by James E. Tierney. Cambridge:
Cambridge University Press. [c. 90,000 words]
Darwin, Charles. 1859 (1982). The Origin of Species ed. by J. W. Burrow. Harmondsworth:
Penguin. [c. 185,000 words]
Deloney, Thomas. 1583–1600 (1912). The Works of Thomas Deloney ed. by Francis Oscar
Mann. Oxford: Clarendon. [c. 117,000 words]
Doyle, Sir Arthur Conan. 1927 (1966). The Case-Book of Sherlock Holmes. London: John
Murray. [c. 74,000 words]
Dryden, John. 1668–1691 (1971). The Works of John Dryden ed. by S. H. Monk & A. E. W.
Mauer. Berkeley & Los Angeles: University of California Press. [c. 127,900 words]
Farquhar, George. 1706 (1995). The Recruiting Officer and Other Plays ed. by William Myers.
Oxford: Oxford University Press. [c. 89,000 words]
Forster, E.M. 1947 (1954). Collected Short Stories. Harmondsworth: Penguin. [c. 64,200 words]
</TARGET "aki">

22 Minoji Akimoto

Hobbes, Thomas. 1651 (1982). Leviathan ed. by Crawford B. MacPherson. Harmonds-

worth: Penguin.
Lamb, Charles. 1811–1848 (1985). Selected Prose ed. by Adam Phillips. Harmondsworth:
Penguin. [c. 165,500 words]
Marvell, Andrew. 1661–1678 (1971). The Poems and Letters of Andrew Marvell. Vol. II: Letters
ed. by H. M. Margoliouth & Pierre Legouis. Oxford: Clarendon. [c. 107,000 words]
Maugham, William Somerset. 1919 (1944). The Moon and Sixpence. Harmondsworth:
Penguin. [c. 74,200 words]
Mill, John Stuart. 1859 (1991). On Liberty and Other Essays ed. by John Gray. Oxford:
Oxford University Press. [c. 76,400 words]
Nashe, Thomas. 1594 (1972). The Unfortunate Traveller and Other Works ed. by J. B. Steane.
Harmondsworth: Penguin [c. 153,600 words]
The Oxford Book of Letters. 1535–1985 (1995). Ed. by Frank Kermode & Anita Kermode.
Oxford: Oxford University Press. [c. 294,500 words]
Richardson, Samuel. 1740–1741 (1980). Pamela, or Virtue Rewarded ed. by Peter Sabor.
Harmondsworth: Penguin. [c. 100,000 words]
Sidney, Sir Philip. 1580 (1977). The Countess of Pembroke’s Arcadia ed. by Maurice Evans.
Harmondsworth: Penguin. [c. 84,800 words]
Smith, Adam. 1776 (1986). The Wealth of Nations Books I-III ed. by Andrew Skinner.
Harmondsworth: Penguin. [c. 48,900 words]
Smollett, Tobias. 1771 (1992). The Expedition of Humphry Clinker ed. by Lewis M. Knapp,
rev. by Paul-Gabriel Boucé. Oxford: Oxford University Press. [c. 116,000 words]
Trollope, Anthony. 1855 (1984). The Warden ed. by Robin Gilmour. Harmondsworth:
Penguin. [c. 67,000 words]
Wentworth Papers. 1597–1628 (1973). Ed. by John Phillips Cooper. London: Royal Historical
Society. [c. 99,000 words]
Wilde, Oscar. 1888–1898 (1992). The Complete Works of Oscar Wilde ed. by Vyvyan Holland.
London: Collins. [c. 138,600 words]
Woolf, Virginia. 1927 (1992). To the Lighthouse ed. by Stella McNichol. Harmondsworth:
Penguin. [c. 60,000 words]
<LINK "all-n*">

<TARGET "all" DOCINFO

AUTHOR "Cynthia L. Allen"

TITLE "On the development of A Friend of Mine"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

On the development of a friend of mine*

Cynthia L. Allen
Australian National University

1. Introduction

In this paper, I will present some findings of an investigation into the develop-
ment of the construction illustrated in (1):
(1) He’s a friend of mine/yours/her neighbor’s.

I will refer to this construction as the ‘double genitive’, which is one of the
traditional terms for it; the construction has also been called the ‘post-genitive’
or simply the ‘friend of mine’ construction. Various investigations have been
made into the origins and early development of this construction; important
references include Einenkel (1905), Jespersen (1927), van der Gaaf (1927),
Hatcher (1950) and Mustanoja (1960). Altenberg (1982) includes a valuable
discussion of the construction in Early Modern English. However, these
investigations have by no means exhausted the subject.
Van der Gaaf ’s (1927) article on this construction purports to identify the
earliest examples, so one might expect that we must at least know when the
construction first appeared in English. But it turns out that there are problems with
van der Gaaf’s dating of the advent of the construction. There are two problems
here. First, the dates of the examples themselves are often faulty. For instance,
he dates examples from The North English Legends edited by Horstmann (1881)
c.1300. But the Middle English Dictionary (MED) dates the manuscript which
Horstmann used (Harley 4196) as “a1425(?c1375)”, which means that the
manuscript itself was probably written before 1425 but not before 1400, while
the text is thought to have been composed earlier, perhaps (but not certainly)
around 1375. Thus van der Gaaf ’s date is too early by at least 75 years.
In some instances, van der Gaaf’s dates are probably off because scholarship
since his time has revised earlier opinions. But a more pervasive problem, still
24 Cynthia L. Allen

found in much more recent works, is a conflation of the date of a manuscript

with the probable date of composition of a text which the manuscript contains.
It is important to keep these two dates apart, because it is always possible that
a later scribe may have changed the original text. As discussed in Allen (1992),
we have clear evidence that scribes sometimes did substitute newer syntax for
the older syntax of an original. We cannot simply use the supposed date of
composition of a text as the date of an example, and we cannot assume that a
given construction existed at the time of the date of composition when it is
found only in a manuscript of a significantly later date. While van der Gaaf ’s
pioneering work remains a wonderful source of examples, the examples must
be scrutinized carefully.
Another problem with dating the advent of this construction involves
defining the ‘friend of mine’ construction. Van der Gaaf included examples like
(2) as examples of this construction:
(2) Haue we nought of þe kinges.
“We have nothing of the king’s.” (CM Vesp. MS 4908)

While it can be called a ‘double genitive’, the construction of (2), which is

headed by what is traditionally called an ‘indefinite pronoun’, is arguably a
different construction from what I will refer to as the ‘noun-headed’ construc-
tion found in a friend of mine (see Section 3.1 for more discussion). Further-
more, examples of double genitives without nominal heads (e.g. (12)) appear in
the texts before noun-headed ones do, and this fact alone is reason to keep the
two types separate. I believe that if we do not differentiate examples of the
noun-headed construction from ones with no nominal head, we run the risk of
obscuring some facts about how the construction with a nominal head entered
the language. Although I believe that a full explanation of the origins of the
double genitive must consider the relationship of the construction with a
nominal head to the one without one, I have focused in this investigation on the
appearance and development of the construction which has an explicitly
mentioned head noun, whether it is preceded by an overt quantifier or deter-
miner, as in (1), or the quantifier/determiner is unexpressed because the noun
is plural, as in men of the king’s. Previous investigations have not made it clear
when the noun-headed construction first appeared and have sometimes given
a mistaken impression of the relative timing of this construction with the noun
modified by an indefinite article or quantifier, a demonstrative and a definite
article. Accordingly, in this paper the term ‘double genitive’ should be taken to
be restricted to the noun-headed type unless otherwise indicated.
 On the development of a friend of mine 25

I have also excluded examples like the king had no children of his own from
my investigation. While van der Gaaf included such examples in his survey, it
seems to me that this should be treated as a distinct construction, which
appeared earlier.

2. Earliest examples

Since van der Gaaf ’s early examples cannot all be treated as genuine examples
of the (noun-headed) double genitive construction, we need to look again at
when the earliest examples are to be found. The Oxford English Dictionary
(OED) discusses the double genitive construction in more than one place. There
is a discussion of the construction in general in the entry for of (OED s.v. of
prep., 44). There is also a brief mention of the construction in the entry for
hers (possessive pronoun):
(3) OED s.v. hers poss. pron.1
[In form, a double possessive, f. poss. pron. hire, HER, thus hires, her’s,
hers (cf. ours, yours, theirs), app. by association with the possessive case
in such phrases as ‘a friend of John’s’, whence ‘a friend of her’s’, formerly
‘a friend of her (hire)’. Of northern origin; the midland and southern
equivalent being HERN1.]

This brief mention contains two comments which I think merit revision. First, it
suggests that a friend of hire is found before a friend of hers is. Second, it suggests
that the ‘double possessive’ form hers apparently comes from association with
the possessive case found in double genitives with nominal possessives.
Regarding the first of these claims, note that the OED does not offer any
evidence that of hire antedated of hers, either in its entry for her (s.v. her poss.
pron.) or in any other entry as far as I can determine. Nor did a full-text search
of the on-line OED for the string of hire provide any evidence. Of 260 hits for
this string and the 602 hits for of hir, in none of them was her genitive. An
extension of my search to all the forms1 which the OED listed for all the
personal pronouns also failed to yield any examples earlier than c.1350 of any
double genitive constructions where the pronoun is unstrengthened.
Why then does the OED claim that the construction appeared first with
unstrengthened possessives? It seems likely that the claim is based on the
assumption that the double genitive had already made its appearance by the late
26 Cynthia L. Allen

OE stage. I know of two OE examples which look rather like double genitives.
Both are from the same manuscript, edited by Skeat (1871):
(4) eac sume wif of urum us bregdon.
also some women of our-dat us amazed
“Also some women of our people amazed us.” (WScp Luke 24.22)
Latin: et mulieres quedam ex nostris terruerunt nos
(5) þa ferdun sume of urum to þære byrgune.
then went some of our-dat to the tomb
“Then some of those who were with us went to the tomb.”
(WScp Luke 24.24)
Latin: quidam ex nostris

I agree with van der Gaaf (1927: 19) in rejecting these examples, both in a fairly
close translation from Latin, as early examples of the double genitive construc-
tion. It is true that the fact that these examples exist at all suggests that they were
not completely alien to the grammar. But the fact that no convincing examples
of the double genitive construction are found until the fourteenth century
suggests that there was no unbroken transmission of the construction. A
translation often stretches the grammar beyond its normal limits, and that
seems to me what has happened here. Van der Gaaf suggests that such transla-
tions are parallel to the use of an of þissum by some OE writers to translate unus
ex istis.
If we accepted these two examples as forerunners of the double genitive
construction, we would indeed expect an intermediate stage, which would
develop when inflection for case agreement dropped off the possessives, in
which we found a friend of hire but not a friend of hers. But there does not seem
to have been any such intermediate stage. It does in fact turn out that the very
first example I have found of a double genitive which I would consider genuine
has an unstrengthened, rather than a strengthened, genitive. This example is
from the Vespasian version of the Cursor Mundi of c.1350:2
(6) a man o þair
a man of their
“a man of their(s)” (CM (Vsp, c1350) 7465)

However, the fact that I found no examples of the double genitive construction
with a strengthened genitive in this text cannot be given much significance. (6)
is in fact the only example3 of the double genitive construction I found in the
portion I read. Since both strengthened and unstrengthened forms4 are used in
 On the development of a friend of mine 27

this manuscript in other constructions, including the double genitive construction

which has no overt noun (e.g. If yee me aught of yurs giue at line 15109), we
cannot make much of the fact that it is the unstrengthened form that is used in
this example. We do find a few more examples of the double genitive construc-
tion with an unstrengthened possessive pronoun, but nearly all come from a
much later date than the earliest examples with a strengthened pronoun in this
construction, in writings from dialects which did not adopt the strengthened
pronoun in any construction until late. I conclude that there is no reason to
assume that the double genitive with an unstrengthened pronoun antedated the
same construction with a strengthened pronoun in all dialects.
The second questionable suggestion that the OED makes in the hers entry
is that the ‘double possessive’ form hers apparently comes from association with
the possessive case found in double genitives with noun possessors, in other
words, that the earliest examples of our construction should have a noun as the
genitive. The OED gives no evidence for such a development, and I found none
in my own investigation, as the Appendix below demonstrates. Indeed, Kellner
(1892: 114) suggested that it was the pronouns which came first as the possessor
in this construction. Both van der Gaaf and Jespersen mention that an example
of the construction with a noun can be found in Chaucer’s Canterbury Tales:
(7) Oon Maximus, that was an officer /Of the prefectes and his corniculere
“One Maximus, who was an officer of the prefect’s and his aide”
(Chaucer CT G 369)

Restricting our definition of the construction as suggested above and correcting

van der Gaaf’s dates, this is the earliest example of the construction with a noun
that I have found.
The earliest example I have found with a pronominal possessor has already
been given as (6). This example is probably a bit earlier than Chaucer’s writings.
Given the uncertainty of the Cursor Mundi date and the small number of
examples at the time, it would probably be going too far to say that pronouns
were possible in the construction before nouns were. However, it is definitely
true that examples with pronouns were more numerous than examples with
nouns were at the earlier stage. For example, there are at least three examples of
the construction with a pronominal possessor in Chaucer’s writings, but (7) is
the only example which I know of with a nominal possessor. Nominal possess-
ors are hard to find until the early fifteenth century, when they become reason-
ably common but are still outnumbered by the examples with a pronoun:
28 Cynthia L. Allen

(8) Julian Lampit, that was fre tenant of the kynges there
“Julian Lampit, that was (a) free tenant of the king’s there”
(Lon. Eng. 226.47 (1413–1419))

So while we cannot say for certain that the construction was ever impossible
with a nominal possessor when it was possible with a pronominal one, it is clear
that there is no reason to believe that the use with a pronoun was modelled on
an earlier use with nouns.
I conclude that the first examples of expressions like ‘a friend of mine’ and ‘a
friend of hers’ come from the middle of the fourteenth century and that they were
not based on expressions like ‘a friend of the king’s’. It appears that the double
genitive construction was used in any given dialect with whatever the ‘indepen-
dent’ form of the pronouns was, whether strengthened or unstrengthened (4).

3. Expansion of the construction

3.1 Original meaning of the construction

Now let us turn to the question of whether the double genitive construction has
expanded its limits. In its entry for of, the OED states that all of the early
examples of the double genitive construction are “capable of explanation as a
partitive”. There have been objections to the use of the term ‘partitive’ for the
double genitive construction. For example, Jespersen (1927: §1.5) points out
that (9a) and (9b) are not synonymous:
(9) a. The general and some of his friends left the house.
b. The general and some friends of his left the house.

The double genitive construction of (9b) does not imply that the general has
more friends than this group, while the construction of (9a) does.
It seems to me that although we might not want to use the term ‘partitive’,
the construction does have a meaning of membership in a set when it is used
with an indefinite determiner or a quantifier. It is just that the set may have
only one member. What is required is that there be a potential for a member-
ship greater than one. Consider:
(10) It is not the sort of pleasantry which I like to hear from a daughter of mine.
(Tempest-Tost, by Robertson Davies 1951; p. 108 of The Salterton Trilogy,
London, Penguin, 1986)
 On the development of a friend of mine 29

Professor Vambrace, who utters this sentence, is able to do so even though he

has only one daughter. This is because the construction merely says that his
daughter Pearl is a member of a particular set, the set of people in a particular
relationship: daughter to me. No implication is made that there are necessarily
other members in this set, although the use of the indefinite article indicates
that the set is one which could in principle contain more than one member.
Using a simple possessive my daughter here would focus on the individual,
rather than on the membership of the individual in question in a particular set.
Let us consider briefly how this construction with a meaning of member-
ship in a set might have developed in the first place. Space does not permit a full
discussion of this question. However, it is notable that in Old and Early Middle
English, the possessive pronouns could be used to mean something like ‘those/
that connected with Pro’:
(11) and macode hine wel bliðe mid his
“and made himself merry with his (people)”
(Vesp. [Link] 233.7)

It seems reasonable to suggest that examples like (12) resulted when the
partitive of combined with this sort of meaning:5
(12) Šif þu mare spenest of þine
“if you spend more of yours” (= ‘your money’)
(Lambeth Hom. 79.13 1185–1225, comp. OE?)

I agree with Mustanoja’s (1960: 165–166) comment that the construction of

(12) was “probably the prototype of the a friend of mine construction”. (12)
talks about a part of a set (what belongs to me); it is a small step to picking out
an entity which is a member of a set which denotes a relationship.

3.2 Examples with a demonstrative

While the double genitive construction has an implication that the designated
set could have a membership of more than one when it is used with an indefi-
nite determiner, the construction with a demonstrative has no such implication
in examples like that nose of his. However, the earliest examples with a demon-
strative are open to an interpretation that there could be more than one
member in the set:
30 Cynthia L. Allen

(13) Or thou gettis this stede of myne

“before you get this place of mine”
(Ipom.A6 7747 a1500 (a1400))

The construction with a demonstrative seems to be an extension of the original

construction, and it appears that the extension was first to adding deixis
(pointing to a specific member or members of the set being referred to) and
then to allowing the demonstrative to be used when the set could only logically
contain one member.
The examples in the OED’s of entry (s.v. of prep., 44) are consistent with
the idea, first put forward by Kellner (1892: 115), that the use of the construc-
tion with a demonstrative was a later development, since the OED’s earliest
examples with a demonstrative are later than its earliest examples with an
indefinite article. Jespersen (1927: §1.5) does not dispute Kellner’s claim, but
only notes (p.16) that “there is nothing strange” in the later occurrence of the
demonstrative. But van der Gaaf (1927) includes examples with a demonstrative
in the constructions which he claims were “more or less usual about the middle
of the 14th century” (1927: 25). So it’s worth looking again at this question.7
In fact, (13) is the only example which van der Gaaf offers which he dates
as belonging to the fourteenth century; he dates it at c. 1350. But it appears that
this is much too early a date for this manuscript. The MED puts it at a1500
(a1400), meaning that expert opinion does not even date the translation of this
text from French as being as early as van der Gaaf ’s date.
So when are the earliest examples from? The OED does not offer any
examples from before the seventeenth century in its discussion of the construc-
tion in its entry for of (OED s.v. of prep., 44), although it does give examples
from the sixteenth century in its entries for ours and yours. But van der Gaaf
offers this genuine-looking example from 1449:8
(14) These godis of myne (…) mowen be seid the goddis of poor men
“These goods of mine (…) may be called the goods of poor men”
(Peacock Repressor p. 409)

Van der Gaaf lists several unassailable examples from later in the fifteenth
century, and so we can conclude that the construction was established by the
middle of the fifteenth century. Neither Mustanoja nor Jespersen mention any
earlier examples, nor have I found any. We remain indebted to van der Gaaf for
locating so many early examples.
 On the development of a friend of mine 31

3.3 Examples with a definite determiner

Let us conclude by looking at a type which appears, to some extent at least, in
earlier English but is not possible in Present-day English. This type uses a
definite determiner:
(15) vnder the fellys of Tomas Bettsons
“under Thomas Bettson’s sheepskins”
(Cely 131.46 (1481, William Cely))

As has often been noted, in Present-day English a definite determiner is only

possible in this construction if the possessed thing has a restrictive modifica-
tion, typically a relative clause:
(16) I (…) took off the old, comfortable navy blue pullover of Elliot’s that I
liked to wear for walking (…)
(Lucille Kallen, C. B. Greenfield: No Lady in the House, p. 85. Ballantine
Books, 1983)

Examples like this are possible because a relative clause has the effect of turning
its head into a member of a set.
Both Jespersen and van der Gaaf comment on the disappearance of the
construction with no such modification, although their comments are rather
different. Jespersen comments only that “[e]xamples with the definite article are
comparatively frequent in Shakespeare, but rare in PE” (1927: 22). He makes no
claim about Middle English. Van der Gaaf, however, comments that the
construction with the definite article is “often met with in M. E. and early
Modern English” (1927: 26) but seems to be obsolete now. Van der Gaaf thus
rather gives the impression that the construction with the definite article is as
old as other types of the construction, although he does not explicitly say this
anywhere in his paper.
But examination of van der Gaaf ’s examples must cause us to dismiss
several of them. Two examples from the Paston Letters turn out to be plurals,
not genitives:
(17) set or cause my lord to do þynges otherwise than accordith to the pleasir of my
lordes
“set or cause my lord to do things otherwise than accords with my Lords’
[not Lord’s] pleasure” (Paston Letter 908. 9, 1469)
32 Cynthia L. Allen

(18) I have wretyn by the meanes of my Lordes heere

“I have written through the agency of my lords here”
(Paston Letter 785.8, 1469)

His only genuine example from the Paston Letters involves a relative clause:
(19) the horse that he hathe of myne
“the horse of mine that he has”
(Paston Letter 383.13, 1471)

Of the six examples before 1450 which van der Gaaf adduces, four examples
(three from Ipomadon) suffer from the dating problems mentioned earlier. One
from 1399 has a possessive pronoun which should arguably be interpreted as
meaning ‘of her possessions’ and at any rate involves a relative clause:
(20) I bequethe (…) for þe gode I haue had of heres by any way, fourty pounde of
gold.
“I bequeath (…) because of the goods I have had of hers in any way,
forty pounds of gold.”
(Will, Robert Folkingham, Yorksh. Wr. II 449 (1399))

And the very earliest example, from the AŠenbite of Inwit, must be rejected:
(21) And þe mesure of þe purse of his þet is zorŠuol and scarse
“And the measure of the purse of him (?) that is miserable and parsimo-
nious” (AŠen. 54.3)

In her edition of this text, Gradon (1979: 132) has a note concerning this
sentence indicating that we should read ine his meaning ‘in his own house’. She
considers this to be a calque on the French ou sien ‘in his own (house)’. Van der
Gaaf admits that Dan Michel made a mistake with his French here, but suggests
that Michel simply substituted an idiom which was familiar to him for some
French which he didn’t understand. But given that there are no similar exam-
ples for more than a century and a half, it does not seem likely that this was a
familiar idiom at the time. Gradon’s explanation for the unusual sentence
seems superior.
This leaves us with a1450 as the date of the first indisputable example not
involving a relative clause:
(22) to tente þe tree of his
“to tend his tree”
(York Plays, 25/86 a1450)
 On the development of a friend of mine 33

Van der Gaaf offers several other examples from the second half of the fifteenth
century which indicate that this was an established construction by this time,
although it is interesting to note that a disproportionate number of the exam-
ples come from North Midlands or Northern texts, suggesting that the con-
struction may have been limited to specific dialects. I have myself noted five
examples from the Cely Letters of the late fifteenth century.
It does not appear that the construction with a definite determiner ever
really flourished. It is true that there are some examples in Shakespeare’s
works, e.g. (23):
(23) unity in the proofs, The mantle of Queen Hermione’s, the jewel about
the neck of it
(Sh. WT [Link].33–4)

But even in the texts where the construction with the definite determiner is
used, it is never as common as the construction with the indefinite determiner
or a quantifier. I recorded all examples of the double genitive construction in
King Lear and The Winter’s Tale. (23) is the only example I found with a definite
determiner, compared with 11 examples with a demonstrative and two exam-
ples with an indefinite determiner. Altenberg (1982: 72) found only two
examples in which the head of the double genitive was preceded by a definite
article in his seventeenth-century corpus, and remarks: “it is difficult to say
whether these cases should be taken as mere lapses or as instances of uncertain
usage”. I think that enough examples are to be found of this construction with
the definite article to make us unable to dismiss all of them as mistakes, and it
seems likely that a careful study of the contexts of these examples would reveal
some discourse and/or stylistic factors that distinguish the double genitive with
a definite article.9
I conclude that although the construction with the definite determiner
existed from the second half of the fifteenth century to at least into the early
seventeenth century, it was a rather marginal construction which never really
took firm root.
Why should this construction be different from the one with a demonstra-
tive, which did take firm root? After all, both of these constructions seem to
appear in English texts around the same time and both allow an English speaker
to get around the restriction against having a determiner followed by a posses-
sive adjective/determiner.
34 Cynthia L. Allen

First, we saw above that the double genitive construction does not by itself
(in Modern English) imply that the set of which the head noun is a member
necessarily has other members in it (cf. example (10)). Many uses of the
demonstrative in the double genitive construction have no implication that the
head noun is the sole member of the set of things in a particular relationship; if
we say let’s talk about that problem of John’s, there is no implication that John
has only one problem, only an implication that you know which of John’s
problems I mean. Such uses of the demonstrative do not seem particularly
problematic, since they are not at a terribly far remove from the original
partitive meaning of the double genitive construction. This would give the
demonstrative an advantage over the definite article, which explicitly indicates
uniqueness. Nevertheless, we still need some explanation of why one would ever
want to use the double genitive construction when only one entity could
logically be involved in the ‘possessive’ relationship, e.g. that nose of yours. Why
not just say your nose?
I think the answer lies in the discourse functions of the demonstratives.
Consider the difference between (24a) and (24b):
(24) a. We’ve got to do something about the water heater.
b. We’ve got to do something about that water heater.

Either of these could be said upon emerging from a shower which ran out of hot
water. But I believe that I would be likely to use (24b) when I have already
mentioned that there is a problem with the water heater. The definite deter-
miner would be enough to signal ‘you know what I’m talking about’, since we
only have one water heater. But the demonstrative that is very useful for
signaling annoyance. I think this is because it is used to signal a further bit of
information: ‘we’ve talked about this before’. Not only is the noun assumed to
be accessible to the hearer, it is also assumed to have been the subject of
previous discussion, which is why we can use demonstratives with proper
nouns, as in I’d like to meet this Thomas Brown. So the that of annoyance follows
from the fact that if the water heater has been discussed, the hearer ought to
have done something about it already.
The definite determiner has no such discourse function. Using the definite
determiner does not usually give an advantage over using the possessive by
itself, since the construction with a simple possessive is already definite.
 On the development of a friend of mine 35

4. Conclusion

The double genitive construction seems to have gone through these stages:
1. Early Middle English: Poss. Pro. used to mean ‘people/things associated with
Pro’, e.g. (11), combined with partitive of to produce examples like (12), with
no nominal head and referring to part of a set.
2. Fourteenth century: partitive meaning shifts to ‘member of a set’ meaning;
noun-headed double genitive construction arises.
3. Fifteenth century: double genitive extended to demonstratives (and margin-
ally to definite determiner), e.g. (14) and (15).
While space has not allowed a thorough discussion of the origins of the double
genitive construction, I hope to have demonstrated conclusively that stage 2
preceded stage 3 and that the construction with a definite determiner never
really flourished in the language. Further research into the discourse functions
of the variants of the construction is likely to help explain its limitations and
distribution.

Appendix: Some findings of the investigation

Text Date (Q) N (Q) N Dem N The N

of N’s of GPro of Gen of Gen
PC(Cont.) c1131–1154 NO NO NO NO
[Link] a1150, OE NO NO NO NO
[Link] a1150, c1125 NO NO NO NO
Trinity a1225, OE NO NO NO NO
Lambeth c1185–1225, OE NO NO NO NO
V&V c1200 NO NO NO NO
Orm c1200 NO NO NO NO
Dialect AB c1220–1225 NO NO NO NO
Wohunge 1st half 13thC NO NO NO NO
Gen&Ex c1250 NO NO NO NO
BrutC post 1189, [Link] NO NO NO NO
O&N 1189–1216, [Link] NO NO NO NO
Kentse c1275 NO NO NO NO
Havelok c1300, c1295–1310 NO NO NO NO
SEL c1300 NO NO NO NO
[Link]. c1325, c1300 NO NO NO NO
36 Cynthia L. Allen

Text Date (Q) N (Q) N Dem N The N

of N’s of GPro of Gen of Gen
AŠenbite 1340 NO NO NO NO
Cursor M c1350 NO 1ex NO NO
PiersP c1400, c1377–1381 NO NO NO NO
Chaucer 1368–1400, c1400–10 1 ex YES NO NO
Lon. Eng. 1385–1425 1 ex YES NO NO
Vernon c1390 NO 1 ex NO NO
BBrut c1400 1 ex 1 ex NO NO
Thornton MS 1430–1440 NO 1 ex NO NO
Kempe a1450 YES NO NO NO
Shillingfd 1447–1450 NO 1ex NO NO
Capgrave b.1398, MS 1451 NO YES NO NO
Stonor 1424–1475 YES YES NO NO
Gregory MS c1475 YES YES NO 1 ex
Malory fl. 1475 YES YES NO NO
Caxton b. 1422 (?) NO YES NO NO
Cely 1472–1488 YES YES NO YES
J. Paston II b. 1442, 1461–1479 YES YES 1 ex NO
Ricart 1479–1506 NO NO NO NO
Elyot b. ?1490 NO NO NO NO
Ascham b. 1515 (1552) NO YES NO NO
Lyly b. 1554 (?), (1579) NO 1 ex 1 ex NO
Shkspre b. 1564 1 ex 1ex YES 1 ex
Harley 1626–1643 NO 1 ex NO NO
Barrington 1629–1630 NO YES YES NO
Nicholas 1641–1660 YES YES YES NO

Notes:
When I have indicated more than one date for a text, the first is the presumed date of the
manuscript while the second is the presumed date of composition of the text. I have
sometimes given an author’s birth date instead of the date of the work. Columns 3 and 4
provide data bearing on the question of whether pronominal possessors antedated nominal
ones in the post-genitive construction. ‘Q’ stands both for quantifiers (e.g. none) and for the
indefinite determiner a; thus column 3 provides data on friends of John’s, some friends of
John’s, a friend of John’s, etc. Column 4, where ‘GPro’ stands for any genitive pronoun,
whether a strengthened or an unstrengthened form, includes (some) friends of his, friend of
our(s), etc. Columns 5 and 6 provide data concerning whether the construction is found with
a demonstrative or a definite determiner, respectively, and here I have not distinguished
between nominal and pronominal possessors, using ‘Gen’ to stand for any genitive form,
including any unstrengthened forms of genitive pronouns. ‘YES’ in any cell indicates that I
found more than one example.
<DEST "all-n*">

On the development of a friend of mine 37

The data come mostly from my own readings of the texts listed. In a few instances, the
data reflect additional examples found in the OED or MED or other sources. I have supple-
mented my reading with searches for genitive NPs in some of the selections found in the
Penn-Helsinki Parsed Corpus of Middle English (PPCME), an annotated and somewhat
extended version of the Helsinki Corpus of English Texts created under the direction of
Matti Rissanen at the University of Helsinki. The annotations were carried out under the
direction of Anthony Kroch at the University of Pennsylvania with the support of the
National Science Foundation. I gratefully acknowledge the contribution which this corpus
has made to my research.

Notes

* I am grateful to an anonymous referee for helpful comments on an earlier version of this

paper.
1. I looked at every example occurring in an entry before 1500 to ascertain that the string in
question was not to be construed as having an ‘independent’ genitive form, rather than a
possessive ‘adjective’ modifying a following noun.
2. The Göttingen MS of perhaps 50 years later has a man of ouris in this line.
3. I have not included if any barn of hir war þine (l. 2601) as a double genitive because I think
it is more likely that the pronoun here should be interpreted as an object pronoun (“any
child from her”) rather than a possessive pronoun. Note that none of the manuscripts of this
edition have an unambiguously possessive pronoun (i.e. some version of hers) here, although
the Göttingen MS typically has a strengthened genitive where the Vespasian MS has an
unstrengthened one.
4. The Cursor Mundi is not the first text to exhibit the ‘strengthened’ forms. The forms hires,
hores, ures and owres are found in the Titus MS of the Ancrene Riwle (as noted by Dieth 1919)
and hiren and heoren are found in the Ancrene Wisse. There seems to be no reason in
principle to expect that the double genitive construction with unstrengthened forms should
have antedated the same construction using strengthened forms in the dialects in which the
strengthened forms emerged early.
5. Hatcher (1950: 10) suggests the progression ‘to take of mine’ > ‘aught of mine’ > ‘anything
of mine’ > ‘a book of mine’.
6. Citations for Ipom.A refer to line numbers in Kölbing’s (1889) edition of the A MS.
7. One question which I will not address here is that of the possible relationship between the
double genitive construction with a demonstrative and what I will refer to as the ‘Dem Poss N’
construction, in which a demonstrative precedes a possessive, as in this my son. Hatcher
(1950: 12) comments “the type this my book, this thy face gave way (in time, exclusively) to
this book of mine, this face of thine”. Hatcher also comments in her footnote 37 “Old English
could combine so easily possessive and demonstrative (article): þis min sunu etc”. But it is
unsafe to assume a direct descent of an Early Modern English construction from an Old
English one that looks similar without evidence of the continued existence of the construction
38 Cynthia L. Allen

through the intervening period, and Mitchell (1985: §108) comments that the OE Dem Poss
N pattern is “perhaps less frequent than the last [the pattern in which the possessive is
followed by a demonstrative + N] but is common in eMnE, e.g. ‘this our old friend’ and
‘these my gifts’. Whether it has had a continuous existence since OE and whether it was in
origin a calque on Latin remains to be established”. My own impression is that the Dem Poss
N construction is uncommon in OE texts and is not to be found again until the early fifteenth
century, although I must emphasize that I have not carried out an investigation into this
question. A systematic examination of the relationship between the double genitive construction
with a demonstrative and the Dem Poss N construction, including an examination of the
limitations of the Dem Poss N construction and possible discourse differences between the
two constructions when both are found in a given text, is likely to yield some very interesting
results. However, although I have suggested a possible pathway of development for the
double genitive construction with a demonstrative, my main concern here has been not to
explain the appearance of this construction, but rather to establish the relative timing of the
appearance of the different variants of the double genitive construction in the texts.
8. This example contrasts the use of the genitive with a demonstrative with the use of a non-
genitive with the definite article. Van der Gaaf offers a putatively earlier example from the
Towneley Plays, but his date of a1400 is not in accordance with the date given by the MED,
viz. a1500(a1460). Cawley & Stevens (1976: xvii, fn. 19) comment that the manuscript is
“certainly not earlier than the 1480s”.
9. It can also be mentioned that a large proportion of the earlier examples with a definite
determiner and no relative clause are used in situations where the possessive is placed to
rhyme. This is true, for example, in all four examples from Ipomadon A. All of these examples
strike me as peculiar in comparison with other texts of the period, and it would not be
surprising if this rather second-rate poet had stretched a construction beyond its normal
limits to fit a rhyme scheme.

References

Allen, Cynthia L. 1992. “Old English and the Syntactician: Some Remarks and a Syntact-
ician’s Guide to the Works of Ælfric”. Evidence for Old English ed. by Fran Colman,
1–19. Edinburgh: John Donald.
Altenberg, Bengt. 1982. The Genitive v. the of-Construction: A Study of Syntactic Variation in
17th Century English. (=Lund Series in English, 62). Lund: Gleerup.
Cawley, A. C. & Martin Stevens, eds. 1976. The Towneley Cycle: A Facsimile of Huntington MS
HM 1. Leeds: University of Leeds, School of English.
Dieth, Eugen. 1919. Flexivisches und Syntaktisches über das Pronomen in der Ancren Riwle.
Zürich: Buchdruckerei Aschmann und Scheller.
Einenkel, Eugen. 1905. “A Friend of Mine”. Anglia 28.504–508.
Gaaf, Willem van der. 1927. “A Friend of Mine”. Neophilologus 12.18–31.
Gradon, Pamela, ed. 1979. Dan Michel’s AŠenbite of Inwyt. Vol. II: Introduction, Notes and
Glossary. (=Early English Text Society, 278). Oxford: Oxford University Press.
 On the development of a friend of mine 39

Hatcher, Anna Granville. 1950. “The English Construction A Friend Of Mine”. Word 6.1–25.
Horstmann, Carl, ed. 1881 [repr. 1969]. Altenglische Legenden: Neue Folge. Hildesheim &
New York: Georg Olms.
Jespersen, Otto. 1927 [repr. London, 1961]. A Modern English Grammar on Historical
Principles. Part III: Syntax (Second Volume). Copenhagen: Ejnar Munksgaard / London:
George Allen & Unwin.
Kellner, Leon. 1892. Historical Outlines of English Syntax. London & New York: Macmillan.
Kölbing, Eugen, ed. 1889. Ipomedon in drei englischen Bearbeitungen. Breslau: Wilhelm Koebner.
Middle English Dictionary. 1952–2001. Ed. by Hans Kurath, Sherman M. Kuhn & Robert
Lewis. Ann Arbor, Mich.: University of Michigan Press.
Mitchell, Bruce. 1985. Old English Syntax. Oxford: Clarendon.
Mustanoja, Tauno F. 1960. A Middle English Syntax. Part I: Parts of Speech. Helsinki: Société
Néophilologique.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S. C. Weiner. Oxford: Oxford University Press.
Skeat, Walter W., ed. 1871. The Holy Gospels in Anglo-Saxon, Northumbrian and Old Mercian
Versions, Synoptically Arranged, with Collations Exhibiting all the Readings of all the Mss.:
Together with the Early Latin Version as Contained in the Lindisfarne Mss., Collated with
the Latin Version in the Rushworth Ms. Cambridge: Cambridge University Press.

Sources

[PPCME = Penn-Helsinki Parsed Corpus of Middle English; see notes to Appendix, above.
Where the PPCME is indicated, I have looked at the PPCME selections in addition to
any other portions of a text mentioned.]
PC(Cont) = the First and Second Continuations of the Peterborough Chronicle. Clark, Cecily,
ed. 1970. The Peterborough Chronicle 1070–1154. Oxford: Clarendon.
Vsp. Axxii = the homilies in Cotton MS Vespasian A. xxii. Morris, Richard, ed. 1867–1868.
OE Homilies of the 12th and 13th Centuries. Vol. I. EETS 29, 34.
Vsp.D. xiv = two pieces found in MS Vespasian D. xiv which were composed in the twelfth
century. Förster, Max. 1925. “Die spätaltenglische Übersetzung der pseudo-anselmschen
Marien Predigt”. Palaestra 148.15–40. Warner, Rubie, ed. 1917. Early English Homilies
from the Twelfth Century MS Vesp. D. xiv. EETS 152.
Trinity B 14.52 = Morris, Richard, ed. 1873. OE Homilies of the 12th and 13th Centuries Vol.
II. EETS 53.
Lambeth = the homilies and the version of the Poema Morale in Lambeth MS 487. Morris,
Richard, ed. 1867–1868. OE Homilies of the 12th and 13th Centuries. Vol. I. EETS 29, 34.
V&V = Holthausen, Ferdinand, ed. 1889, 1921. Vices and Virtues. EETS 89.
Orm = Holt, Robert, ed. 1878. The Ormulum: With the Notes and Glossary of Dr. R.M. White.
Oxford: Clarendon. Dedication and lines 1–3,000 of the Homilies only.
Dialect AB = various works written in the ‘AB’ dialect of the early thirteenth century. I read
the first five parts of the Ancrene Wisse as well as all of Sawles Warde, along with the
portions found in the PPCME. Tolkien, J. R. R., ed. 1962. The English Text of the
40 Cynthia L. Allen

‘Ancrene Riwle’: Ancrene Wisse. EETS 249. Wilson, R. M., ed. 1938. Sawles Warde.
Kendal: Titus Wilson.
Wohunge = the ‘Wooing’ group. Thompson, W. Meredith, ed. 1958. Þe wohunge of ure
lauerd. EETS 241.
Gen&Ex = Arngart, Olaf, ed. 1968. The Middle English Genesis and Exodus. Lund: C. W. K.
Gleerup.
BrutC = the Cotton Caligula A ix version of LaŠamon’s Brut. Brook, G. L. & R. F. Leslie, eds.
1963. LaŠamon: Brut. EETS 250. First 3,000 lines.
O&N = Stanley, Eric G., ed. 1972. The Owl and the Nightingale. Manchester: Manchester
University Press.
Kentse = the Kentish Sermons. Hall, Joseph, ed. 1920. Selections from Early Middle English
1130–1250. Oxford: Clarendon. 214–222.
Havelok = Smithers, G. V., ed. 1987. Havelok. Oxford: Clarendon.
SEL = The South English Legendary (Laud 108 version). Horstmann, Carl., ed. 1887. The
Early South-English Legendary; or, Lives of Saints. EETS 87. Pages 1–150.
[Link]. = Wright, William Aldis, ed. 1965 [1887]. The Metrical Chronicle of Robert of
Gloucester. Rolls Series 86. London: Kraus Reprint. Volume I of the ‘A’ MS (Cotton MS
Caligula A. xi) version.
AŠenbite = Gradon, Pamela, ed. 1979. See references. PPCME portions and pp. 1–81.
Cursor M = the first two volumes of the Vespasian [Link] MS version of the Cursor Mundi.
Morris, Richard, ed. 1874, 1876. Cursor Mundi. EETS 57, 59.
PiersP = the ‘B’ version of The Vision of Piers Plowman. Schmidt, A. V. C., ed. 1978. The
Vision of Piers Plowman. London: Dent.
Chaucer = various works of Geoffrey Chaucer. Benson, Larry D., ed. 1987 [3rd ed.]. The
Riverside Chaucer. Boston: Houghton Mifflin. The pieces read were The Book of the
Duchess, House of Fame, and books I and II of Troilus and Cressida.
[Link]. = Daunt, Marjorie and R.W. Chambers, eds. 1967 [1931]. A Book of London English
1384–1425. Oxford: Clarendon.
Vernon = the Vernon MS version of the ‘Mirror of St. Edmund’. Edition used: Horstmann,
Carl, ed. 1895. Yorkshire Writers: Richard Rolle of Hampole. Pages 240–261.
BBrut = Brie, F., ed. 1906. The Brut, or the Chronicle of England. EETS 131. PPCME selec-
tions and pp. 1–16.17 & 28.9–217.5.
Thornton = Perry, G. G., ed. 1921 [1866]. English Prose Treatises of Richard Rolle de Hampole.
EETS 20. I excluded pp. 21–27 as they are not from the Thornton MS.
Kempe = Meech, Sanford & Hope E. Allen, eds. 1940. The Book of Margery Kempe. EETS
212. PPCME selections and pp. 100–160.
Shillingfd = Moore, S. A., ed. 1965 [1871]. Letters and Papers of John Shillingford, Mayor of
Exeter 1447–50. (Camden Society N.S. 2). PPCME selections and pp. 1–32, 35–40, 43–60
& 66–68.
Capgrave = Munro, John James, ed. 1910. Lives of St. Augustine and St. Gilbert of Sempring-
ham and a Sermon. EETS 140.
Stonor = Kingsford, C. L., ed. 1919. The Stonor Letters and Papers, 1290–1483. (Camden
Society Third Series 29, 30). PPCME selections and the letters of vol. I covering 1424–
1475 (excluding items which are not letters, e.g. inventories).
</TARGET "all">

On the development of a friend of mine 41

Gregory = Gairdner, James, ed. 1876. Gregory’s Chronicle: The Historical Collections of a
Citizen of London in the Fifteenth Century. (Camden Society N. S. 17). I searched the
PPCME selections and read the remaining part of the chronicle (i.e. pp. 197–end).
Malory = Field, P. J. C., ed. 1990 [3rd ed.]. The Works of Sir Thomas Malory. Oxford & New
York: Clarendon. Pages 7–152.
Caxton = pieces by William Caxton. PPCME selections and Caxton’s prologues and
epilogues not included in the PPCME. Crotch, J. B. 1956 [1928]. The Prologues and
Epilogues of William Caxton. EETS 176.
Cely = Hanham, A., ed. 1975. The Cely Letters, 1472–1488. EETS 273.
J Paston II = letters of John Paston II. Davis, Norman, ed. 1971. Paston Letters and Papers of
the Fifteenth Century. Oxford: Clarendon.
Ricart = Smith, Lucy Toulmin, ed. 1872. Ricart’s Kalendar: The Maire of Bristowe is Kalendar.
(Camden Society N. S. 5). I read only the parts which were by Ricart himself.
Elyot = Elyot, Thomas. 1970 [1531]. The Book Named the Governor, 1531. Menston, England:
Scolar Press. First 60 pages.
Ascham = William Wright, ed. 1904. Roger Ascham: English Works. Cambridge: Cambridge
University Press. I read the Scholemaster and A Discourse on the Affaires of the State of
Germanie.
Lyly = Bond, Richard Warwick, ed. 1902. The Complete Works of John Lyly. Vol. I. Oxford:
Clarendon.
Shkspre = two plays by William Shakespeare. Pafford, J. H. P. ed. 1963. The Winter’s Tale.
London: Methuen. / Fraser, Russell, ed. 1963. The Tragedy of King Lear. New York:
Signet Classics.
Harley = Lewis, Thomas Taylor, ed. 1854. The Letters of Lady Brilliana Harley. (Camden
Society N.S. 58). First 110 pages.
Barrington = Searle, Arthur, ed. 1983. Barrington Family Letters 1628–32. (Camden Society
Fourth Series 28). First 160 pages.
Nicholas = Warner, George F., ed. 1920. The Nicholas Papers: Correspondence of Sir Edward
Nicholas. Vol. IV: 1657–1660. (Camden Society Third Series 31). I did not include the
letters attributed to Nicholas himself, as the edition contains only extracts made nearly
a century later. My investigation was confined to letters to Nicholas or to people
concerned with him.
<TARGET "bib" DOCINFO

AUTHOR "Douglas Biber and Victoria Clark"

TITLE "Historical shifts in modification patterns with complex noun phrase structures"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Historical shifts in modification patterns

with complex noun phrase structures
How long can you go without a verb?

Douglas Biber and Victoria Clark

Northern Arizona University

1. Introduction

Formal written prose has long been regarded as more complex than spoken
discourse, but there has been surprising disagreement concerning the structural
locus of complexity. For example, several scholars have claimed that written
discourse has longer sentences and a greater use of subordination (e.g. O’Don-
nell et al. 1967; O’Donnell 1974; Kroll 1977; Akinnaso 1982). Others have
claimed that writing is more explicit and more decontextualized than speech
(e.g. DeVito 1966, 1967; Kay 1977; Olson 1977). And some have focused on the
way that writing tends to be more carefully planned and organized than speech
(Ochs 1979; Brown & Yule 1983; Gumperz, Kaltman & O’Connor 1984).
More recently, large-scale corpus-based investigations have been undertak-
en to provide more generalizable comparisons of spoken and written discourse.
Many of these studies have focused on the noun phrase as the central distin-
guishing structure: formal written prose frequently incorporates long, complex
noun phrases, while spoken registers rely much more heavily on pronouns and
other simple noun phrase types (see e.g. Biber et al. 1999: Ch. 8). At the same
time, these studies show that some earlier generalizations are not accurate. In
particular, it turns out that subordination is prevalent in both conversation and
academic writing, although spoken and written registers rely on different types
of embedded clauses (see e.g. Biber et al. 1999: Chs. 8–10).
These differences were noted as early as 1979 by Halliday (see also Halliday
1988). Chafe (1982) characterizes written prose as being tightly “integrated”,
referring in part to the ways in which information is packed into noun phrases.
44 Douglas Biber and Victoria Clark

This complex nature of noun phrases in written prose was documented much
more fully in studies such as Varantola (1984) and de Haan (1989).
The title of our paper illustrates the kind of complex noun phrase structure
typical of formal written prose. Note especially that the main title consists of a
single noun phrase (with the head noun shifts), modified by an attributive
adjective (historical) and two prepositional phrases as postmodifiers (in…
patterns; with… structures). Each of those prepositional phrases also contains
a complex noun phrase, with multiple examples of attributive adjectives and
nouns occurring as premodifiers (modification; complex noun phrase). Al-
though this is a long and complex noun phrase, it contains no clausal modifi-
ers. Such structures suggest that the answer to the question raised in the
subtitle is “a very long time”.
In the following section, we show that non-clausal modifiers are much
more common than clausal modifiers in Present-day English. In addition, these
structures are sharply stratified across registers: while readers regularly encoun-
ter structures of this type in academic writing, they are virtually unattested in
normal conversation. This situation differs from that found in earlier historical
periods: written registers in earlier historical periods showed a greater reliance
on clausal modifiers, and there was less difference between technical and
colloquial written registers in the use of these features.
After describing the present-day patterns of use, we attempt to track the
historical shift towards non-clausal modifiers as the preferred pattern of use in
informational written registers. In Section 2, we first summarize findings from
the Longman Grammar of Spoken and Written English (LGSWE; Biber et al.
1999) that document the present-day patterns of use. Section 3 provides
additional background, summarizing previous research on historical changes in
noun modification patterns over the past four centuries. Then, in Section 4, we
present a series of detailed linguistic investigations based on the ARCHER
Corpus, to more accurately pinpoint when these shifts occurred and how these
changes were realized across different registers. Finally, we summarize the
patterns of use in Section 5 and attempt to provide functional explanations for
this recent dramatic shift towards non-clausal modification; in particular, we
suggest that recent advances in the technology of literacy coupled with the
recent ‘informational explosion’ have facilitated these historical developments.
 Historical shifts in modification patterns with complex noun phrase structures 45

2. Present-day patterns of noun modification in English

The Longman Grammar of Spoken and Written English (Biber et al. 1999)
provides a detailed account of noun modification in English. That work
describes the major patterns of use associated with each grammatical variant,
employing a corpus-based analysis of texts from four registers: conversation,
fiction, newspaper language and academic prose. The analyses were carried out
on the Longman Spoken and Written English (LSWE) Corpus, which contains
approximately 40 million words of text overall, with approximately 4–5 million
words from each of these four registers. To provide a baseline for the historical
analyses reported in Sections 3–4, we describe here some of the major synch-
ronic patterns of noun modification.

300000
Noun Phrases per million words

250000

200000

150000

100000

50000

0
Conv Fict News Acad
Register
No modifier Premodifier Postmodifier Both pre- and post-

Figure 1. Distribution of Noun Phrases with Premodifiers and Postmodifiers (based on

Biber et al. 1999, Figure 8.4).

First, as Figure 1 shows, noun phrases occur with roughly the same frequen-
cy in all registers. However, the internal complexity of those noun phrases
differs dramatically across registers: in conversation, most noun phrases (85%)
have no modifiers at all. Over half of those noun phrases are realized by a
simple pronoun. Newspaper language and academic prose show strikingly
different patterns, with approximately 60% of all noun phrases having some
modifier, and many noun phrases having multiple modifiers.
46 Douglas Biber and Victoria Clark

There are a number of different structures that serve as premodifiers and

postmodifiers:
major premodifier types:
Attributive adjectives: “a special project”, “an internal memo”
Participial adjectives: “hidden variables”, “detecting devices”
Nouns as premodifiers: “the bus strike”, “the police report”
major postmodifier types:
Relative clause: “the penny-pinching circumstances that surrounded this
international event”
Ing-clause: “the imperious man standing under the lamp-post”
Ed-clause: “a stationary element held in position by the outer casting”
Prepositional phrase: “compensation for emotional damage”
Of-phrase: “this list of requirements”
We tend to think of attributive adjectives as the major device used for pre-
modification, and relative clauses as the major type of postmodification.
However, as Figures 2–4 show, these perceptions about use are not entirely
accurate. With premodifiers (Figure 2), nouns are nearly as important as
attributive adjectives, accounting for 30–40% of all premodifiers in news and
academic prose. The actual situation of use is even more surprising with
postmodifiers (Figure 3): prepositional phrases (rather than relative clauses) are
overwhelmingly the most common type of postmodification, accounting for
almost 80% of all postmodifiers in news and academic prose. Relative clauses
are the most common type of clausal postmodification (Figure 4), but they
account for only 10–15% of all postmodifiers.
Overall these findings show the dominance of non-clausal types of noun
modification in Present-day English. In the following sections, we briefly
describe noun modification patterns in earlier historical periods, and then go
on to trace the shifts to the current patterns of use.

3. Previous historical research on noun modification patterns

There have been few historical studies that specifically address the preferred
patterns of noun modification in earlier historical periods. Probably the most
useful for our purposes is Raumolin-Brunberg (1991), who described the
typical characteristics of noun phrases in sixteenth-century English, based on
analysis of the writings of Sir Thomas More.
 Historical shifts in modification patterns with complex noun phrase structures 47

80000

70000
Noun Phrases per million words

60000
50000
40000
30000
20000
10000
0
Conv Fict News Acad
Register
common adjective -ing adjective -ed adjective noun

Figure 2. Distribution of Noun Phrases with Single Premodifiers (based on Biber et al.
1999, Figure 8.7).

90000
80000
Postmodifiers per million words

70000
60000
50000
40000
30000
20000
10000
0
Conv Fict News Acad
Register
Prep phrases Other postmodifiers
Figure 3. Prepositional vs. other postmodification across registers (based on Biber et al.
1999, Figure 8.12).

The typical patterns of use in Sir Thomas More’s writings differ in impor-
tant ways from Present-day English. First, attributive adjectives are by far the
preferred type of premodifier, occurring ten times more often than noun
premodifiers. Further, when nouns are used as premodifiers, they are usually a
title (such as King, Master or Doctor). Turning to postmodifiers, Raumolin-
48 Douglas Biber and Victoria Clark

25000
Postmodifiers per million words

20000

15000

10000

5000

0
Conv Fict News Acad
Register
Relative clauses -ing clauses -ed clauses TO clauses Appositives

Figure 4. Non-prepositional postmodification across registers (based on Biber et al.

1999, Figure 8.13).

Brunberg shows that prepositional phrases are already slightly more common
than relative clauses in the sixteenth century. However, the difference is not
nearly as marked as in Present-day English. Further, Raumolin-Brunberg points
out that the range of prepositional phrases used for postmodification is restrict-
ed, with 70% of these being of-phrases. These patterns of use are strikingly
different from the present-day patterns documented in Biber et al. (1999),
raising the question of when and how the present-day patterns emerged.
Previous ‘multi-dimensional’ studies of register variation provide some
partial answers to this question. The Multi-Dimensional (MD) analytical
framework (Biber 1988) is useful here in that it allows comparison of texts and
text varieties with respect to a large number of co-occurring linguistic features,
representing underlying ‘dimensions’ of variation. In particular, two of the
dimensions identified in previous research have functional/linguistic correlates
that are directly relevant to issues of noun phrase complexity: Dimension 1
(‘Involved versus Informational Production’) and Dimension 3 (‘Situation-
Dependent versus Elaborated Reference’). The ‘informational’ pole of Dimension
1 includes co-occurring features like nouns, attributive adjectives and preposi-
tional phrases. The ‘elaborated reference’ pole of Dimension 3 includes several
different types of relative clause constructions. Thus both dimensions include
co-occurring features that reflect increased complexity in the noun phrase.
 Historical shifts in modification patterns with complex noun phrase structures 49

The MD framework has been used for a number of historical analyses of

register variation (e.g. Biber & Finegan 1989, 1997; see also several of the papers
in Conrad & Biber 2001). Most of these studies have used the ARCHER Corpus.
ARCHER — A Representative Corpus of Historical English Registers —
was designed for a specific major research agenda: to analyze historical change
in the range of written and speech-based registers of English from 1650 to the
present. The general design goal for the corpus has thus been to represent as
wide a range of register variation as possible, sampled systematically across texts
from the last three and a half centuries.
The overall structure of the corpus comprises ten major register categories,
sampled in 50 year periods from 1650 to the present, as summarized in Table 1.
Altogether, the complete corpus includes 1,037 texts and approximately 1.7
million words. Among the written registers, the corpus includes personal styles
of communication (journals/diaries and personal letters), prose fiction, popular
exposition represented by news reportage, and specialist expository registers,
represented by legal opinions, medical prose and scientific prose. The corpus
similarly includes several different kinds of speech-based registers: dialogue in
drama and dialogue in fiction as reflections of casual face-to-face conversation,
and sermons as a reflection of planned monologue styles.
Registers are represented by at least 10 texts per 50-year period, in most
cases chosen using random selection techniques (with available bibliographies

Table 1.
Overview of ARCHER

Total size: 1,037 texts; approximately 1.7 million words

Time-span covered: 1650–1990, divided into 50-year periods

Dialects covered: British (all periods) and American (1 period per century)
Genres/Registers:
7 Written Categories: journals/diaries, personal letters, fiction prose, news reportage,
legal opinions, medical prose, scientific prose
3 Speech-based Categories: drama, fiction dialogue, sermons
Target Sampling: 10 texts, at least 2,000 words, per genre (and dialect) in each period
A full sampling for a genre includes 100 texts:
1650–1699, British: 10 texts
1700–1749, British: 10 texts
1750–1799, British: 10 texts 1750–1799, American: 10 texts
1800–1849, British: 10 texts
1850–1899, British: 10 texts 1850–1899, American: 10 texts
1900–1949, British: 10 texts
1950–1990, British: 10 texts 1950–1990, American: 10 texts
50 Douglas Biber and Victoria Clark

serving as sampling frames). American English registers are sampled for only
one 50-year period per century. Table 2 summarizes the final composition of
the corpus by register category. Biber, Finegan & Atkinson (1994) and Biber,
Conrad & Reppen (1998: Methodology Box 2) provide more details about the
design, sampling and compilation of ARCHER.

Table 2.
Breakdown of texts in ARCHER by register category

Text category Number of texts

Journals 100
Letters (more than 10 texts per period; most texts shorter than 1,000 275
words)
Fiction prose 100
News 100
Legal (1750–1990; American only)
57
Medicine (no 18th c. American)
90
Science (from Philosophical Transactions of the Royal Society; no Ameri-
70
can texts)
Drama (only 5 texts from 18th c. American)
95
Fiction dialogue 100
Sermons (only 5 texts per period)
50

Biber & Finegan (1997) use ARCHER to track overall changes in the
patterns of register variation over the past four centuries. The developments
along Dimensions 1 and 3 are especially relevant for our purposes here. Along
Dimension 1, specialist written registers (e.g. medical prose and science prose)
steadily evolve towards an increased use of the ‘negative pole’ features, includ-
ing more frequent nouns, attributive adjectives and prepositional phrases.
Popular speech-based registers, like drama, shift in the opposite direction,
representing a less frequent use of those same features. The patterns of change
along Dimension 3 are similar: specialist written registers shift towards a greater
use of relative clause constructions over time, while speech-based registers shift
towards a less frequent use of those features. Biber & Finegan interpret these
historical developments as reflecting a major split between specialist written
registers and popular registers (both written and speech-based). In earlier periods,
all of these registers were relatively similar in their linguistic characteristics. Over
time, though, the specialist registers have shifted to become sharply distinguished
from the other registers, marked by a steady increase in the structural and
informational complexity of noun phrases.
 Historical shifts in modification patterns with complex noun phrase structures 51

These studies document the general trends towards greater complexity in

the noun phrase, especially in informational written registers such as academic
prose. However, they do not allow detailed consideration of the particular
structural resources employed at different stages of this shift. In the following
section we take up that question.

4. Tracking the shifts towards non-clausal modification

For our analysis of noun phrase modification, we decided to compare the

patterns of use in four registers from ARCHER: drama, fiction, newspaper
reportage and medical prose. We chose these four because they cover much of
the range of register variation available in ARCHER, and because they allow
easy comparisons to the findings from the LGSWE (summarized in Section 2
above). In Section 4.1 we discuss the patterns of change for premodification,
and then we focus on postmodifiers in Section 4.2.

4.1 Premodification patterns

For premodification, we focused on attributive adjectives (Figures 5 and 6) and
noun–noun sequences (Figure 7). Figure 5 shows the same major split between
specialist written registers (news and medical prose) and popular registers (fiction

60

50
Freq per 1,000 words

40

30

20

10

0
17th C. 18th C. 19th C. 20th C.
Period
Drama Fiction News Medical

Figure 5.
Attributive adjectives across periods.

52 Douglas Biber and Victoria Clark

14

12

10
Freq per 1,000 words

0
17th C. 18th C. 19th C. 20th C.
Period

Drama Fiction News Medical

Figure 6.
ADJ and/or ADJ sequences across periods.

60

50
Freq per 1,000 words

40

30

20

10

0
17th C. 18th C. 19th C. 20th C.
Period
Drama Fiction News Medical

Figure 7.
Noun–noun sequences across periods.

 Historical shifts in modification patterns with complex noun phrase structures 53

and drama) documented in earlier MD studies (see Section 3 above). That is,
the frequency of attributive adjectives remains constant in drama and fiction
across these periods, but it increases dramatically in news and medical prose.
Conjoined sequences of adjectives (e.g. male and female workers; racial or
religious cohesion) show a somewhat different distribution (Figure 6): relatively
rare across all periods in drama, fiction and news, but showing a marked
increase in frequency in medical prose over the last two centuries. Thus, while
news reportage is similar to academic prose in many respects, features like this
indicate that academic registers depart to an even greater extent from the norms
of earlier historical periods.
Interestingly, Figure 7 shows that news reportage has consistently used
noun–noun sequences (e.g. bus station) to a greater extent than medical prose.
In fact, medical prose, fiction and drama are nearly the same in their infrequent
use of noun–noun sequences over the seventeenth, eighteenth, and nineteenth
centuries. In the twentieth century, though, both medical prose and newspaper
reportage show marked increases in the use of noun–noun sequences, with
these features remaining more common in news.
In Section 3 above, we noted that premodifying nouns were used primarily
as titles in the sixteenth century. It turns out that this same distinction is
important in accounting for the greater use of these features in newspaper
reportage. Figure 8 shows that roughly half of all premodifying nouns in news
function as titles modifying a proper noun (e.g. forms like Duke, Archbishop,
President). It is only in the last 50 years that we see a marked increase in the use
of nouns premodifying common nouns. Medical prose represents strikingly
different patterns of use, as shown in Figure 9: premodifying nouns used as
titles (i.e. modifying a proper noun) are rare in all periods, and they actually
become less common in the last century. In contrast, the use of nouns modify-
ing a common noun increases dramatically, especially in the last 50 years. A
comparison of Figures 8 and 9 shows that medical prose and news use noun +
common noun to nearly the same extent; the higher frequency of noun–noun
sequences in news (shown in Figure 7) reflects the greater use of premodifying
nouns as titles in that register.
Text samples 1–3 illustrate these patterns of use. Text Sample 1 is from an
eighteenth-century news report; Text Sample 2 is from a twentieth-century
news report; and Text Sample 3 is from a twentieth-century medical article.
Noun–noun sequences are underscored in all three texts.
54 Douglas Biber and Victoria Clark

50

45

40

35
Freq per 1,000 words

30

25

20

15

10

0
1650 to 99 1700 to 49 1750 to 99 1800 to 49 1850 to 99 1900 to 49 1950 to 90
Period
N + Proper N N + Common N

Figure 8.
Breakdown of N+N sequences in News.

Key to text samples:

– blood pressure (underscored) marks noun–noun sequences
– OF (bold, CAPS) marks postnominal of-phrases
– in (bold, italics) marks other prepositional postnominal phrases
(1) NEWS. 1743. The London Gazette. #8294, 8295, Jan. 17–21, 21–24, Feb. 18–21.
AT the Court at St. James’s, the Nineteenth Day of January, 1743.
PRESENT,
The King’s most Excellent Majesty in Council.
His Majesty having been graciously pleased to deliver the
Custody OF the Seals OF the Dutchy and County Palatine OF
Lancazster, to the Right Honourable Lord Richard Edgeumbe, the
Oath OF Chancellor OF the said Dutchy was this Day, by his
Majesty’s Command, administred to his Lordship.
[…]
Naples, January 4. In the last Week the King performed the
Ceremony OF investing Cardinal Corsini with the Cardinal Cap.
The Chamber OF Financaes has fixed the Manner OF raising the
 Historical shifts in modification patterns with complex noun phrase structures 55

50

45

40

35
Freq per 1,000 words

30

25

20

15

10

0
1650 to 99 1700 to 49 1750 to 99 1800 to 49 1850 to 99 1900 to 49 1950 to 90
Period
N + Proper N N + Common N

Figure 9.
Breakdown of N+N sequences in Medical Prose.

extraordinary Contribution OF a Million OF Ducats, required by

the King, towards the Expences OF putting his Forces in the best
Posture against all Events; which is exacted with Rigour, and
occasions an extraordinary Uneasiness and Discontent. Frequent
Councils are held at Court on the Dispatches brought by the
Couriers from Spain, and upon the daily Arrival OF Officers from
M. de Gages’s Army. The Military Dispositions, and the Raising
OF the Militia, are continued with great Assiduity. From
Calabria we have received the melancholy Account OF an Earthquake
that happened there, by which the City OF Chieti has suffered
greatly. The Accounts OF the Sickness in that Province are this
Week rather more favourable, the Malady seeming to decrease
daily.
(2) NEWS. 1989. The Times. 1/2/1989. No. 63, 280. Pp. 1 & 18.
Thatcher says no to revenge for Lockerbie
Eye for an eye warning to the United States.
By Philip Webster, Harvey Elliott and Christopher Thomas.
56 Douglas Biber and Victoria Clark

The Prime Minister declared her opposition to avenging the

Lockerbie air disaster yesterday, as American intelligence chiefs
admitted failing to link the crash with any terrorist organization.
[…]
With the incoming Bush administration in America certain to
face pressure for retaliation, Mrs Thatcher’s outspoken rejection
OF reprisal raids could pose the first difficulties in her
relationship with the new President.
[…]
Meanwhile, a further tightening OF baggage inspection
procedures is likely to emerge from a review OF Britain’s airline
and airport security. It will be launched this week at a meeting
OF the national Aviation Security Committee, comprising government,
airline, union and safety officials.
The committee, under the chairmanship OF the senior Civil
Servant in the Department OF Transport’s Civil Aviation Policy
Directorate, will report urgently to Mr Paul Channon, Secretary OF
State for Transport, and is certain to recommend rearranging the
areas OF responsibility for baggage checks.
(3) Medical Prose. Irvine, N., et al. 1985. The results of coronary arteriography
in young men after myocardial infarction in north-east Scotland. Scottish
Medical Journal 30: 8–14.
The case records OF 50 consecutive male patients aged 40
years or under who were investigated by selective coronary
arteriography after myocardial infarction were reviewed. Fourteen
patients had normal coronary vessels and 36 patients had
significant occlusive disease. Eighteen were considered to be in
need OF surgical treatment. The features OF myocardial infarction
on the ECG were less marked in the group OF patients with normal
coronary arteriograms. Many OF these patients were asymptomatic and
had complete resolution OF the ECG changes. As well as having
normal coronary arteries, many also had normal left ventricular
angiograms.
Cigarette smoking was very common in the whole group, 86 per
cent OF patients being moderately heavy cigarette smokers. Five OF
the 14 patients in the ‘non-occlusive’ group were non-smokers and
only two OF the 36 patients in the ‘occlusive’ group were
 Historical shifts in modification patterns with complex noun phrase structures 57

non-smokers (P ·is less thanÒ 0.01). The fasting serum cholesterol

was significantly lower in the ‘non-occlusive’ group than in the
‘occlusive’ group. There was no significant difference between the
two groups regarding blood pressure, family history OF ischaemic
heart disease, obesity or alcohol consumption. There was, however,
a high incidence OF heavy alcohol consumption amongst patients who
subsequently required coronary artery surgery.
[…]
The present study was carried out to investigate the pattern
OF coronary artery disease in young men in North East Scotland
following myocardial infarction and to determine whether there is
any relationship between the clinical features OF infarction, risk
factors, post-infarction progress and the presence or absence OF
obstructive coronary artery disease.
Text Sample 1 illustrates the primary use of premodifying nouns as titles in
earlier newspaper reports (e.g. County, Lord, Cardinal). This use continues in
present-day news reportage, but there is an even greater increase in the use of
nouns modifying common nouns, as illustrated in Text Sample 2 (e.g. air
disaster; intelligence chiefs; terrorist organization). Text Sample 3 illustrates the
dense use of these same modification patterns in present-day medical prose
(e.g. case records; male patients). In fact, both Sample 2 and Sample 3 illustrate
how multiple nouns can be used in sequence modifying following head nouns
(e.g. baggage inspection procedures; fasting serum cholesterol).
In sum, there has been a steady increase in the use of both attributive
adjectives and nouns as premodifiers over the past three centuries. However,
the most dramatic change has been the extremely recent increase in the use of
nouns premodifying common nouns.

4.2 Postmodification patterns

In Section 2 we identified prepositional phrases and relative clauses as the two
most productive types of postnominal modifier in Present-day English.
Prepositional phrases are in fact many times more common than relative
clauses, but relative clauses are the most common type of clausal modifier.
Figure 10 displays the overall patterns of change for these two features, and
confirms the greater reliance on prepositional phrases as postmodifiers. For
relative clauses, there has been relatively little change in use over these periods.
58 Douglas Biber and Victoria Clark

35

30
Freq per 1,000 words

25

20

15

10

0
drama fiction news medical
Register
Rels; 18th C. Rels; 20th C. PPs; 18th C. PPs; 20th C.

Figure 10.
Frequency of relative clauses and prepositional phrases as postmodifier.

35

30

25
Freq per 1,000 words

20

15

10

0
17th C. 18th C. 19th C. 20th C.
Period

OF phrases Prep phrases ED clauses

ING clauses Relative clauses

Figure 11.
Postmodifier types in Drama across periods.

Prepositional phrases were already more common as postmodifiers in the eigh-

teenth century, especially in registers like news and medical prose. However, the
 Historical shifts in modification patterns with complex noun phrase structures 59

35

30

25
Freq per 1,000 words

20

15

10

0
17th C. 18th C. 19th C. 20th C.
Period
OF phrases Prep phrases ED clauses
ING clauses Relative clauses

Figure 12.
Postmodifier types in Fiction, across periods.

difference in the use of these structures has increased dramatically over the past
two centuries, so that prepositional phrases are many times more common than
relative clauses by the twentieth century.
To more fully document the patterns of change for postnominal modifiers,
we plot the full range of modifiers for each register in Figures 11–14. These
figures show the frequencies of the three major types of postnominal clauses
(relative clauses, -ing clauses and -ed clauses) across all periods, together with
the frequencies of postnominal prepositional phrases in the eighteenth century
and twentieth century.1 These figures show that clausal modifiers are rare in all
registers when compared to the frequencies of prepositional postmodifiers.
Further, these clausal modifiers show little change in their frequency of use
across periods.
In contrast, prepositional phrases are common in all periods and show a
marked increase in more recent periods. Because of the heavy reliance on
of-phrases in sixteenth-century English (see Raumolin-Brunberg 1991, dis-
cussed in Section 3), we distinguish between of-phrases and other prepositional
phrases in these figures. This distinction is especially important in news and
60 Douglas Biber and Victoria Clark

40

35

30
Freq per 1,000 words

25

20

15

10

0
17th C. 18th C. 19th C. 20th C.
Period

OF phrases Prep phrases ED clauses

ING clauses Relative clauses

Figure 13.
Postmodifier types in News, across periods.

medical prose (see Figures 13–14): of-phrases are overwhelmingly the dominant
type of nominal postmodifier in eighteenth-century English. However, there
has been a dramatic increase in the use of other prepositional phrases over the
past two centuries (coupled with a decrease in the use of of-phrases, at least in
news reportage), resulting in the present-day pattern with of-phrases and other
prepositional phrases being used with about the same frequencies.
Figure 14 shows that this shift has not come about gradually. Rather, other
prepositional phrases continued to be relatively rare in medical prose through
the nineteenth century, showing an extreme increase in use only in the past 100
years. Figure 15 shows that this increase is due largely to a marked increase in
the use of prepositional phrases headed by two specific prepositions: in and
with. (We have unintentionally illustrated the use of both structures in the title
to our paper.)
Text Samples 1–3 (in Section 4.1) illustrate these patterns of use. Of-phrases
are shown in BOLD CAPS, while other prepositional phrases are shown in bold
italics. Text Sample 1 illustrates the dense use of of-phrases in eighteenth-
century newspaper reportage (e.g. the Custody OF the Seals OF the Dutchy and
 Historical shifts in modification patterns with complex noun phrase structures 61

40

35

30
Freq per 1,000 words

25

20

15

10

0
17th C. 18th C. 19th C. 20th C.
Period
OF phrases Prep phrases ED clauses
ING clauses Relative clauses

Figure 14.
Postmodifier types in Medical Prose, across periods.

County Palatine OF Lancazster; the Chamber OF Financaes; the Manner OF

raising the extraordinary Contribution OF a Million OF Ducats). This frequent
use of of-phrases continues in twentieth-century news, but Text Sample 2
additionally illustrates the frequent use of other prepositional phrases (e.g. the
incoming Bush administration in America; pressure for retaliation; the first
difficulties in her relationship with the new President). Finally, Text Sample 3
illustrates the dense use of other prepositional phrases (in addition to
of-phrases) in twentieth-century medical prose (e.g. myocardial infarction on the
ECG; patients with normal coronary arteriograms). Note especially the use of in
to identify group membership in addition to the literal meaning of physical
location (e.g. the 14 patients in the ‘non-occlusive’ group; the 36 patients in the
‘occlusive’ group; the pattern OF coronary artery disease in young men in North
East Scotland).
62 Douglas Biber and Victoria Clark

30

25

20
Freq per 1,000 words

15

10

0
18th C. 19th C. 20th C.
Period
OTHER TO FOR WITH IN
Figure 15.
Frequency of individual prepositions as postmodifier (excl. OF), in Medical
Prose.

5. Summary and conclusion

Four major patterns of use documented in the present study are especially
noteworthy. First, for nominal premodifiers, we have shown that nouns are as
productive as attributive adjectives, with a major historical change to favor
nouns modifying a common noun as head (rather than a proper noun). Second,
for postnominal modifiers, we have shown that prepositional phrases are many
times more common than clausal modifiers, with a major historical shift to
favor prepositional phrases beginning with ‘other’ prepositions (in addition to
of-phrases). Third, we have documented the extremely rapid rate of change in
the past 100 years (especially the past 50 years in the case of noun–noun
sequences). And finally, we have shown how these patterns of change have been
sharply stratified across registers, occurring primarily in the informational
written registers (like newspaper reportage and medical prose), while the
patterns of use in popular registers (like drama and fiction) have remained
relatively stable.
 Historical shifts in modification patterns with complex noun phrase structures 63

One major consequence of these historical shifts is the development of a

much more compressed style of presentation, which is at the same time less
explicit in the expression of relational meaning. That is, nominal modifiers
could be ranked along a cline of compression as follows:
COMPRESSED – premodifiers < phrasal < non-finite < relative – EXPANDED
EXPRESSION postmodifiers clauses clauses EXPRESSION

Over the past three centuries, nominal modifiers have been used with increas-
ing frequencies, with the largest expansion in use occurring at the ‘compressed’
end of this continuum (premodifiers and phrasal postmodifiers).
These increasingly compressed styles of expression are at the same time less
explicit in meaning. For example, noun–noun sequences can represent a
bewildering array of meaning relationships, with no overt signal of the intended
meaning. The following list illustrates only a few of these meaning relationships:
noun–noun sequence meaning relationship
air disaster N1 expresses the location of N2
reprisal raid N1 expresses the purpose of N2
baggage inspection N1 expresses the ‘patient’ of N2
airline officials N2 is employed by N1
blood pressure N2 is caused by N1
glass bottle N2 is composed of N1
Similarly, prepositional phrases headed with the prepositions of, in and with can
represent a wide range of different meaning relationships to the preceding head
noun. In contrast, relative clauses are much more explicit about the intended
meaning relationships (e.g. compare systems which provide feedback continuously
but have chaotic behaviour versus continuous-time feedback systems with chaotic
behaviour). Thus, the movement towards a more compressed style comes with
a high price: the loss of explicit signals expressing the meaning relations among
constituents.
Interestingly, these shifts have been restricted primarily to informational
written registers, accelerating only in the last 50–100 years. Two functional
factors have probably been influential in these developments. First, there has
been an increasing awareness of the production possibilities of the written
mode, offering almost unlimited opportunities for crafting and revising the
final text. Typewriters, and more recently word processors, have been techno-
logical developments that facilitate authors’ abilities to manipulate the language
in written texts. At the same time, we have witnessed an ‘informational explosion’,
64 Douglas Biber and Victoria Clark

resulting in pressure to communicate information as efficiently and economi-

cally as possible. Although there may be additional structural and social factors,
these two factors can be taken together to explain in part the rapid increase in
the use of compressed noun modification devices over the past 100 years.
Additional research could be undertaken to track these changes across
smaller time periods and additional registers. It will be especially interesting
over the coming years to observe whether the rate of these changes slows down,
similar to the S-curve patterns found with the lexical diffusion of sound changes
(McMahon 1994: 50–53). It will also be interesting to observe whether popular
spoken and written registers (such as conversation and fiction) follow this same
course of change, or whether a stable stratified pattern is maintained distin-
guishing ‘oral’ and ‘literate’ registers. At present, though, we appear to be in the
midst of a dramatic period of change for the preferred patterns of noun
modification.

Note

1. Prepositional phrases can serve either adverbial or postnominal functions. Because these
functions could not be reliably distinguished by computer, all prepositional phrases were
analyzed by hand in the present study. As a result, we report findings here for only two major
periods: eighteenth century and twentieth century. For medical prose, we also analyzed the
patterns of use in the nineteenth century, to identify the actual period of change with greater
accuracy.

References

Akinnaso, F. Niyi. 1982. “On the Differences between Spoken and Written Language”.
Language and Speech 25.97–125.
Biber, Douglas. 1988. Variation across Speech and Writing. Cambridge: Cambridge Univer-
sity Press.
Biber, Douglas & Edward Finegan. 1989. “Drift and the Evolution of English Style: A History
of Three Genres”. Language 65.487–517.
Biber, Douglas & Edward Finegan. 1997. “Diachronic Relations among Speech-based and
Written Registers in English”. To Explain the Present: Studies in the Changing English
Language in Honour of Matti Rissanen ed. by Terttu Nevalainen & Leena Kahlas-Tarkka,
253–275. Helsinki: Société Néophilologique.
Biber, Douglas, Edward Finegan & Dwight Atkinson. 1994. “ARCHER and its Challenges:
Compiling and Exploring A Representative Corpus of Historical English Registers”.
 Historical shifts in modification patterns with complex noun phrase structures 65

Creating and Using English Language Corpora ed. by Udo Fries, Gunnel Tottie & Peter
Schneider, 1–14. Amsterdam: Rodopi.
Biber, Douglas, Susan Conrad & Randi Reppen. 1998. Corpus Linguistics: Investigating
Language Structure and Use. Cambridge: Cambridge University Press.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999.
Longman Grammar of Spoken and Written English. Harlow, Essex: Pearson Education
Limited.
Brown, Gillian & George Yule. 1983. Discourse Analysis. Cambridge: Cambridge University
Press.
Chafe, Wallace L. 1982. “Integration and Involvement in Speaking, Writing, and Oral
Literature”. Spoken and Written Language: Exploring Orality and Literacy ed. by Deborah
Tannen, 35–54. Norwood, N. J.: Ablex.
Conrad, Susan & Douglas Biber, eds. 2001. Variation in English: Multi-Dimensional Studies.
Harlow, Essex: Pearson Education Limited.
DeVito, Joseph A. 1966. “Psychogrammatical Factors in Oral and Written Discourse by
Skilled Communicators”. Speech Monographs 33.73–76.
DeVito, Joseph A. 1967. “Levels of Abstraction in Spoken and Written Language”. Journal of
Communication 17.354–361.
Gumperz, John J., Hannah Kaltman & Mary Catherine O’Connor. 1984. “Cohesion in
Spoken and Written Discourse”. Coherence in Spoken and Written Discourse ed. by
Deborah Tannen, 3–20. Norwood, N. J.: Ablex.
Haan, Pieter de. 1989. Postmodifying Clauses in the English Noun Phrase: A Corpus-based
Study. Amsterdam: Rodopi.
Halliday, M. A. K. 1979. “Differences between Spoken and Written Language: Some Implica-
tions for Literacy Teaching”. Communication through Reading: Proceedings of the
Australian Reading Conference ed. by Glenda Page, John Elkins & Barrie O’Connor,
37–52. Adelaide: Australian Reading Association.
Halliday, M. A. K. 1988. “On the Language of Physical Science”. Registers of Written English
ed. by Mohsen Ghadessy, 162–178. London: Pinter.
Kay, P. 1977. “Language Evolution and Speech Style”. Sociocultural Dimensions of Language
Change ed. by Ben G. Blount & Mary Sanches, 21–33. New York: Academic Press.
Kroll, Barbara. 1977. “Ways Communicators Encode Propositions in Spoken and Written
English: A Look at Subordination and Coordination”. Discourse across Time and Space
(SCOPIL 5) ed. by Elinor O. Keenan & Tina Bennett, 69–108. Los Angeles, Calif.:
University of Southern California.
McMahon, April M. S. 1994. Understanding Language Change. Cambridge: Cambridge
University Press.
Ochs, Elinor. 1979. “Planned and Unplanned Discourse”. Discourse and Syntax ed. by Talmy
Givón, 51–80. New York: Academic Press.
O’Donnell, Roy C. 1974. “Syntactic Differences between Speech and Writing”. American
Speech 49.102–110.
O’Donnell, Roy C., W. J. Griffin & R. C. Norris. 1967. “A Transformational Analysis of Oral
and Written Grammatical Structures in the Language of Children in Grades Three, Five,
and Seven”. Journal of Educational Research 61.36–39.
</TARGET "bib">

66 Douglas Biber and Victoria Clark

Olson, David R. 1977. “From Utterance to Text: The Bias of Language in Speech and
Writing”. Harvard Educational Review 47.257–281.
Raumolin-Brunberg, Helena. 1991. The Noun Phrase in Early Sixteenth-century English: A
Study Based on Sir Thomas More’s Writings. Helsinki: Société Néophilologique.
Varantola, Krista. 1984. On Noun Phrase Structures in Engineering English. Turku: University
of Turku.
<TARGET "bri" DOCINFO

AUTHOR "Laurel J. Brinton"

TITLE "Grammaticalization versus lexicalization reconsidered"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Grammaticalization versus lexicalization

reconsidered
On the late use of temporal adverbs

Laurel J. Brinton
University of British Columbia

1. Introduction

Despite being the subject of intense study in recent years — and the topic of
several conferences — the process of grammaticalization is not yet fully
understood. In particular, the distinction between grammaticalization and
lexicalization (or degrammaticalization) remains vexed. In this paper I will
examine these concepts as they are defined in the literature, focusing on the
concept of lexicalization, about which there exists perhaps even more confusion
than there does about grammaticalization. I will then pursue a problem in the
history of English which presents a challenge for these concepts: the use of
temporal adverbs as attributive adjectives, as in my late father or the then
practice, a usage which was common in the Early Modern English period but is
now fairly restricted.

2. Grammaticalization vs. lexicalization

There are at least four different ways in which grammaticalization and lexical-
ization are thought to be related to one another. The most widely accepted view
is that they are opposite, or mirror image processes. In other words, lexical-
ization is a process of degrammaticalization. They move in different directions,
as shown in Figure 1.
However, recently it has been proposed that the two processes are comple-
mentary. Moreno Cabrera (1998) has argued that lexicalization is lexicotelic
68 Laurel J. Brinton

(grammaticalization)

lexical item grammatical item

(lexicalization)
Figure 1. Grammaticalization and lexicalization as mirror-image processes.

(‘leading towards lexical items’) and syntactogenetic (‘originating in syntax’),

that it involves metonymic concretion and that it feeds the lexicon but bleeds
syntax. In complementary fashion, grammaticalization is syntactotelic and
lexicogenetic, involves metaphorical abstraction and feeds syntax but bleeds the
lexicon. Giacalone Ramat (1998: 121) has likewise suggested that the two
processes may be “complementary or overlapping” in that both involve loss of
autonomy and univerbation. Hagège (1993: 171) sees the relation in a compara-
ble way: while both processes involve “partial or total loss of the ordinary trap-
pings”, lexicalization is mostly unconscious, the semantic results of it are
unpredictable and the newly formed units behave like lexical items, while
grammaticalization is always unconscious, it involves semantic broadening and
the newly formed units are increasingly subject to grammatical constraints.
A third view of the relationship is that the two are parallel processes
occurring on different levels, or in different domains. Traugott (forthcoming b)
points out that the processes of “conflation and opacity” found in lexicalization
have much in common with the processes of “bonding and coalescence” found
in grammaticalization, the former involving changes at the level of lexical
expression and the latter involving changes in morphosyntactic status. More
specifically, Wischer (2000) argues that both involve syntactic reanalysis,
demotivation, conventionalization, fossilization and phonetic reduction. She
represents the parallelism as in Table 1. The asterisk in the table denotes
Wischer’s belief that the change from grammatical item to lexeme is “highly
unlikely”. She notes, however, that the processes differ semantically:
grammaticalization is characterized by the loss of semantic content, lexicaliza-
tion by the addition of semantic content.
 Grammaticalization versus lexicalization reconsidered 69

Table 1. Lexicalization and grammaticalization as parallel processes (adapted from

Wischer 2000: 365)
Lexicalization Grammaticalization

syntagm becomes new lexical item syntagm becomes new grammatical item
* lexeme becomes grammatical item
lexeme becomes more lexical grammatical item becomes more grammatical

A fourth view of the relation between grammaticalization and lexicalization

is expressed by Lehmann (2002) as “orthogonal” ‘at right angles’. Both are
“reductive” processes, he says, but while grammaticalization reduces the
autonomy of an item and regulates it, lexicalization reduces the inner structure
of an item and adds it to an inventory. For Lehmann, the opposite of grammati-
calization is degrammaticalization (giving autonomy to a hitherto dependent
expression), while the opposite of lexicalization is folk etymology (the bestow-
ing of structure onto a hitherto opaque expression).

2.1 Complex prepositions

Giacalone Ramat (1998: 120) points out a further source of confusion between
grammaticalization and lexicalization, namely, that “in [the] linguistic literature
the same phenomena are sometimes cited as exemplary cases of either linguistic
process”. The formation of complex prepositions, such as ahead of, in case of or
on top of (see Quirk et al. 1985: 669–673), constitutes a striking example of this
observation.
According to Traugott (forthcoming a; cf. also Schwenter & Traugott 1995;
Tabor & Traugott 1998: 244–253), substitutive complex prepositions such as
instead of, in place of and in lieu of provide a “fairly uncontroversial example” of
grammaticalization. They involve decategorialization of the nominal, general-
ization to a larger class of complements and syntactic reanalysis, as well as
semantic changes, all of which are typical of grammaticalization. In contrast,
Huddleston (1984: 342) thinks that complex prepositions arise historically
through lexicalization; likewise, Lehmann (2002), discussing Spanish complex
prepositions such as desde < de ex de, bajo < baxo de and cabe < a cabo de, argues
that they undergo reanalysis characteristic of lexicalization in that they are
subtracted from the rules of syntax and their internal structure becomes no
longer relevant. Central to his argument is the claim that there are grammatical
and lexical members of every word class, so that the primary prepositions such
70 Laurel J. Brinton

as de are “grammatical” and the secondary prepositions such as desde are

“lexical”.
Ramat (1992: 553–554) suggests that the English complex prepositions
between < be tweonum, among < on gemang and beside < be sidan are part of a
spiral from the lexicon (via syntax) to the grammar and back to the lexicon.
While the change from the analytic source to the synthetic target can be seen as
grammaticalization, with the targets clearly belonging to the grammatical words
of English, he notes, “they are, however, part of the English lexicon and cannot
be considered on a par with other grammatical means like affixes and ablaut
(…)” (554).1 Ramat thus sees both processes at work, conceding that “the
boundary between lexical and grammatical units is not neat” (555).
Likewise, Quirk et al. (1985:1530) seem to equivocate in this respect. They
point out that in their grammar they have considered the formation of complex
prepositions (as well as of complex subordinators and phrasal and prepositional
verbs) “as a grammatical phenomenon (…) though without ignoring the lexicali-
zation aspect”. Their reason for treating these forms grammatically is that the unit
of lexicography is the word. However, they admit that a process of what they
call “phrasal lexicalization” may appropriately describe just such forms, which
often have semantic and grammatical integrity and signs of coalescence. Their
conclusion, much like Ramat’s, is that “this illustrates the gradience between
grammar and lexicon, including a gradience in lexicalization” (1985: 1530).

3. Lexicalization

Before we can understand the relation between grammaticalization and

lexicalization, we would do well to have a clearer sense of what is meant by
‘lexicalization’.2

3.1 Definitions
As Lessau notes in his dictionary on grammaticalization, “[t]his term [lexical-
ization] has acquired various, partly quite incompatible, uses within the
scholarly discourse” (1994: s.v. lexicalization). These uses are not entirely
discrete but overlap in a number of ways. Moreover, lexicalization, like gram-
maticalization, is used ambiguously to refer to both synchronic and diachronic
processes (Traugott forthcoming b). There is consequently “a great deal of
confusion” (Wischer 2000: 359) about lexicalization.3
 Grammaticalization versus lexicalization reconsidered 71

At least nine definitions of lexicalization can be found in the literature,

which I will discuss beginning with the broadest one.
1. Adoption into the lexicon
The “intuitive sense” of lexicalization, Lessau asserts, is “a process that leads
from something that is not a (or, one) lexeme to a lexeme, i.e. to something that
‘belongs in the lexicon’” (1994: s.v. lexicalization). A similar definition of
“synchronic” lexicalization as the adoption of a word into the lexicon as a usual
formation — as opposed to a nonce formation — is given by Bussmann (1996: s.v.
lexicalization). In this view, the result of lexicalization is an item which is stored
in the lexicon or becomes part of the “inventory” (Lehmann 2002).4
2. Falling outside the productive rules of grammar
According to Anttila (1989: 159), “[w]henever a linguistic form falls outside the
productive rules of grammar it becomes lexicalized”. A more extended definition
in this same vein is given by Bauer (1983); for him, lexicalization, which is the third
step following “nonce formation” and “institutionalization”, can be said to have
occurred when a lexeme “has, or takes on, a form which it could not have if it
had arisen by the application of productive rules” (1983: 48). He sees four types
of lexicalization: (a) phonological (e.g. the separation of hus in husband from
house, which results from phonetic change); (b) morphological (see definition
7 below); (c) semantic (see definition 8 below); and (d) syntactic (e.g. the
opaqueness of the verb–object relation in a compound such as pickpocket).
3. Ordinary processes of word formation
For Hagège (1993: 171ff.), lexicalization is the association of two free units
through derivation or compounding to yield a new complex unit. Quirk et al.
(1985: 1525–1530) classify not only derivation and compounding, but also
conversion and back formation as lexicalization.
4. Grammatical word (category) > lexical word (category)
This is the most widely used sense of the term ‘lexicalization’, though ‘degram-
maticalization’ is also used in this sense (see below). For example, in their
textbook on grammaticalization, Hopper & Traugott (1993: 49, 127) under-
stand lexicalization as the development of a fully referential lexical item from a
nonlexical, or grammatical item, such as the development of the verb up from
the homophonous particle up or the German verb duzen from the pronoun du.
Similarly, Ramat (1992: 550–551) defines lexicalization as “a process whereby
linguistic signs formed by rules of grammar are no longer perceived (parsed) in
this way but simply as lexical entries”, citing examples such as comparatives that
72 Laurel J. Brinton

have lost their grammatical status (elder, mayor) or participles that are no
longer part of the verbal paradigm (shorn, cloven).
5. Syntactic construction > lexeme
A common understanding of lexicalization is the univerbation of a syntactic
phrase or construction into a single word. For example, Hopper & Traugott
(1993: 127; see also Traugott 1994: 1485) speak of the “incorporation and fos-
silization of earlier independent grammatical morphemes into lexical material”.
Moreno Cabrera (1998: 214) says that lexicalization “obtains when a phrase or
a syntactically-determined lexical item (…) becomes a full-fledged lexical item
in itself”; it is “the process of creating lexical items out of syntactic units”
(1998: 214). This is the process of “desyntacticization” referred to by Wischer
(2000: 258). The classic example of this type of change is today < to + dæge.
Lehmann (2002) seems to have a similar sense of lexicalization. Beginning
with a complex unit, he claims, lexicalization destroys the regular syntactic
construction, renouncing its internal structure, leading to lost or irregular
internal relations.
6. Bound morpheme > lexeme
Anttila (1989: 151) notes that the change of a suffix to an independent word
(e.g. ism, ology, onomy, ocrasy, ade, itis) is “a clear case of lexicalization” (see
also Newmeyer 1998: 269). Likewise, Ramat (1992: 549–550) cites examples
such as ade (from lemonade), teen (from teenager) and gate (from Watergate) as
examples of degrammaticalization (the same as lexicalization by his definition).5
Note that the examples cited all involve derivational, not inflectional affixes.
Combining the two directions of change implied by definitions 5 and 6, we
have a process of lexicalization as shown below, with movement inward from
either below or above:
morpheme Æ lexeme ¨ syntactic construction
The focus of these processes is the end point, the lexeme.
7. Independent morphemes > monomorphemic form
The development of idiosyncratic lexical pairs due to phonological change and
morphological loss, such as lie/lay or foot/feet, is cited by Hopper & Traugott
(1993: 127, 223n) and Traugott (1994: 1485) as an additional type of lexical-
ization. Newmeyer (1998: 263–264) describes the fusion of affixes with roots to
create new, morphologically opaque lexical items, as in *drank + jan > drench,
as “lexicalization”.6 Similarly, Bauer (1983: 53–54) cites non-productive affixes
such as -th and word sets such as eat/edible or right/rectitude, where one
 Grammaticalization versus lexicalization reconsidered 73

member is productive and one non-productive (and hence lexicalized), as

instances of what he calls “morphological lexicalization”.
8. Idiomaticization
Lexicalization is often equated with idiomaticization, or the loss of semantic
compositionality (Lehmann 2002; Traugott forthcoming b). This is what Bauer
calls “semantic lexicalization” (1983: 55–59), instancing compounds such as
blackmail and butterfly or derivatives such as unquiet and inspector. Bussmann
considers this to be the diachronic element of lexicalization, which occurs when
“the original meaning can no longer be deduced from its individual elements”
(1996: s.v. lexicalization; see also idiomaticization). The motivation of the form
can only be deduced etymologically; it has undergone “demotivation” (see
Wischer 2000: 358).
9. Semanticization
Finally, Hopper & Traugott (1993: 223n) and Traugott (1994: 1485) suggest that
the term ‘lexicalization’ can also refer to what they prefer to call ‘semantici-
zation’, or the incorporation of conversationally inferred meanings into the
conventional meaning of the word.
How can these many senses of lexicalization be reconciled? Traugott (forthcom-
ing b) points out that several of them have in common the concept of “mor-
phosemantic opacity”. However, if one adopts the view that lexicalization is in
some sense the reverse of grammaticalization, as is traditionally assumed, it
might also be possible to arrive at a working definition of lexicalization by
reversing accepted definitions of grammaticalization. For example, one might
reverse Haspelmath’s (1999: 1044) definition (“grammaticalization shifts the
linguistic expression further towards the functional pole of the lexical-function-
al continuum”) to read: “lexicalization shifts the linguistic expression further
toward the lexical pole of the lexical-functional continuum”.
Or one might, as Wischer (2000:359) suggests, reverse Kuryłowicz’s (1965:69)
well-known definition (“grammaticalization consists in the increase of the range
of a morpheme advancing from a lexical to a grammatical or from a less grammati-
cal to a more grammatical status, e.g. from a derivative formant to an inflectional
one”) to read: “lexicalization is the process that turns linguistic material into
lexical items, i.e. into lexemes, and renders them still more lexical”.
Recently, Traugott (2000) has suggested a revision of Kuryłowicz’s defini-
tion of grammaticalization. She proposes a split between “primary grammatic-
alization”, which involves a functional change (the first half of the definition)
and “secondary grammaticalization”, which involves a formal change (the
74 Laurel J. Brinton

second half of the definition). A comparable move reconciles several of the

different definitions of lexicalization given above, viz.:
a. Primary lexicalization: the process that turns linguistic material into lexical
items. This is a functional change, the change of operants into lexemes, such as
pronoun > noun, auxiliary > verb (definition 4 above).
b. Secondary lexicalization: the process that renders lexical items still more
lexical.7 This is a formal change. It would refer to the loss of morphophonemic
texture, such as the change from a morpheme or clitic to an autonomous word
(definitions 6 and 7 above). It would also refer to the loss of syntactic texture,
such as the change from a syntactic construction to a word (definition 5 above).
It can be argued, I believe, that these changes may or may not result in semantic
opaqueness or non-compositionality (definition 8 above).8

3.2 Related concepts

Before turning to my test case, I would like briefly to examine several related
concepts.
1. Degrammaticalization
For many scholars, degrammaticalization is the same as lexicalization (e.g.
Kuryłowicz 1965; Ramat 1992), though for others it is not (e.g. Wischer 2000;
van der Auwera 2002). Similarly, while some examples of degrammaticalization
have been cited in the literature,9 for many scholars the phenomenon is still
considered extremely rare. For example, Heine, Claudi & Hünnemeyer find
instances of degrammaticalization to be “statistically insignificant” (1991: 5),
Lehmann finds “no cogent examples of degrammaticalization” (1995: 19) and
Haspelmath (1999: 1046) believes it to be “extremely restricted”.
Again, there are several uses of the word ‘degrammaticalization’. The most
common is identical with the fourth definition of lexicalization given above,
namely the process of converting a grammatical word to a lexical word, thus
reversing the first half of Kuryłowicz’s definition. For example, Chen (1998)
suggests that when a grammatical item has lost its grammatical function and
become “merely a lexical item”, it has been degrammaticalized. Ramat (1992)
argues that degrammaticalization occurs when an item (either a morpheme or
a word) becomes devoid of grammatical function and acquires concrete lexical
status. Wischer (2000: 359) stipulates that degrammaticalization involves a
grammatical item turning back into a lexeme. A related, but somewhat more
 Grammaticalization versus lexicalization reconsidered 75

specialized sense of degrammaticalization is that it involves a hitherto depen-

dent expression gaining autonomy (Lehmann 2002).
The phenomenon of bidirectionality implied by the concept of degram-
maticalization obviously poses a threat to the unidirectionality hypothesis of
grammaticalization (see, for example, Hopper & Traugott 1993: Chapter 5;
Haspelmath 1999). This hypothesis has come under attack on both philosophi-
cal and empirical grounds (see Newmeyer 1998: 260–275; Fischer 2000; Lass
2000), though space does not allow a full discussion of this controversy here.
A second sense of degrammaticalization is the movement from more to less
grammatical (Heine, Claudi & Hünnemeyer 1991: 4), thus reversing the second
half of Kuryłowicz’s definition of grammaticalization. Chen (1998) argues that
degrammaticalization can also refer either to a grammatical item acquiring a
lexical sense while retaining a grammatical function or to the disuse or extinc-
tion of a grammatical item; Allen (1995) suggests that degrammaticalization is
the loss of a word’s grammatical role.10
Van der Auwera (2002) distinguishes between “wide degrammaticalization”
and “narrow degrammaticalization”. The former includes both degrammatical-
ization and lexicalization, as in the evolution of if into a noun. The latter
includes only degrammaticalization, as in the evolution of the genitive ’s
inflection into a clitic (see below).
2. Decliticization/ demorphologization
Decliticization refers to the emergence or re-emergence of a clitic as an inde-
pendent word (Lehmann 1995: 18), or the movement from morphology into
the lexicon. A commonly cited example is the change in the English genitive:
from inflectional ’s (the King’s dog) to enclitic ’s (the King of England’s dog) to
independent his (the King his dog).11 Decliticization is usually understood as a
type of degrammaticalization (equivalent to definition 6 of lexicalization).
The term ‘demorphologization’ (or ‘demorphemization’), though it might
suggest a comparable change, is normally used to refer to the opposite direction
of movement, namely from morphology into phonology (see Joseph & Janda
1988: 196; Lehmann 1995: 13–14). It refers to the movement from agglutination
to inflection or from morphology to morphophonemics (cf. definition 7 of
lexicalization above). It represents a species of what Hopper (1994; cf. also
Hopper & Traugott 1993: 164) terms “phonogenesis”: a morpheme loses its
grammatical-semantic contribution to a word, but retains some remnant of its
original form and thus becomes an indistinguishable part of the phonetic
construction, as in the case of the dative plural inflection -um in seldom or the
76 Laurel J. Brinton

nasal infix in stand. Again, combining senses (1) and (2), we arrive at the
diagram given below:
morphophonemic element ¨ inflection ¨ agglutinative affix Æ clitic Æ word (Æ syntax)12

The focus of these processes is the starting point, the affix, with movement
outward in either direction, up or down.
For Hopper (1994: 31; cf. also Lehmann 1995: 14) phonogenesis or demor-
phologization is the end stage of the grammaticalization process, while van der
Auwera (2002) cites examples of phonogenesis as cases of degrammaticaliza-
tion. Moreover, Joseph & Janda (1988: 204) suggest that demorphologization
may involve movement towards “(greater) lexicalization”.13
3. Regrammaticalization
Regrammaticalization would seem to refer to three possible circumstances: (a)
a form without any function acquires a new grammatical function; (b) a form
is reinterpreted in a new grammatical function; and (c) a form which has lost its
function regains it (see Anttila 1989: 150; Heine, Claudi & Hünnemeyer 1991: 4,
262n; Lessau 1994: s.v. regrammaticalization; Allen 1995).

4. Test case: temporal adverbs as adjectives

The evolution of a set of temporal adverbs, such as late or then, into attributive
adjectives in English (as in my late husband or the then president) presents an
opportunity to explore, once again, the distinction between grammaticalization
and lexicalization.
The set of temporal adverbs used as adjectives fall into the following
semantic classes:
a. Anterior time reference: aforetime, before, erstwhile, hitherto, late, long-ago,
long-time, quondam, since, then, whilom, yesterday.
b. Present time reference: now, today.
c. Posterior time reference: after, hereafter, soon.
d. Durative/iterative time reference: ever, evermore, oft, often, once, one-time,
seldom, short-time, sometime, thrice, twice, two-time.
 Grammaticalization versus lexicalization reconsidered 77

4.1 Discussion of the phenomenon in synchronic grammars

Only brief mention of the use of temporal adverbs as adjectives is made in
traditional grammars, generally under the rubric of word formation (e.g.
Kruisinga 1932:128–129; Quirk et al. 1985:453). Both Jespersen (1948:353–355)
and Poutsma (1926:698–704) cite a number of examples of adverbs of time which
function as adjuncts; Jespersen (1948:290–291) points to their frequent use with
verbal or adjectival ideas.
Traditional grammarians have two main concerns in regard to these forms.
First, they question whether they are ‘really’ adjectives. Their opinion —
expressed quite decisively — is that they are not; rather, they see them as ad hoc
formations which are essentially still adverbial in nature:
the attributive use of adverbs cannot be said to have struck firm root in the
language. Although frequently seized on for the syntactic convenience it
affords, it mostly grates on our linguistic instinct (…) [The forms] have largely
preserved their adverbial character, owing to the fact that they do not suggest
any equivalent adjective (Poutsma 1926: 699; cf. also Quirk et al. 1985: 453).

Only Kruisinga expresses mild disagreement: pre-position, he argues, causes the

adverb to have “a somewhat more adjectival character, i.e. causes it to express
some quality rather than position in time or space” (1932: 232).
The second question concerns why these adverbial forms are used in this
context in place of ‘true’ adjectives. Jespersen (1948: 290, 353) suggests that
everyday adverbs are used when there are no corresponding adjectives, hence
explaining the frequency of then and rarity of now (since present exists).
Dictionaries seem to equivocate about the status and function of these
forms, variously describing them as “adjectiv[al]”, “quasi-adjectiv[al]”, “attrib-
utive”, or “passing into adjectiv[al]”. They sometimes note explicitly that the
forms derive from adverbs.

4.2 Historical data14

Most attributive uses of temporal adverbs appear for the first time in Early
Modern English, including after, before, ever/never, evermore, hereafter, hitherto,
late, now, oft, often, once, quondam, seldom, since, sometime, soon, then, thrice,
twice, whilom and yesterday. After is found commonly beginning in the seven-
teenth century with nouns denoting time (1a), verbal action (1b) and personal
role or title (1c). There are also a number of fully lexicalized forms with after,
such as afternoon, afterbirth, afterlife, aftertaste, aftereffects or after-dinner.15
78 Laurel J. Brinton

(1) after (OED s.v. after a., I.3)

a. surely when the after Age/Shall hither come in Pilgrimage / These
sacred Places to adore. (1681 Marvell, Miscellaneous Poems 77;
U of V)
dine with thee at noon, / Daunce all the after day, bring thee at
night/Into the wedding chamber (1609 Armin, The Two Maids of
More-clacke, 1634–36; CHVD)
b. or jump the after inquiry on your own peril. (1609–10 Shakespeare,
Cymbeline [Link].151–52; U of V)
He’s passing cunning to deceive himself, / But all the better for the
after sport. (1662 Haughton, Grim the Collier of Croyden [Link].311–12;
CHVD)
c. Any other after Tenant of the land (1641 Termes de la Ley 138; OED)
The after Lawyers whose hands it passed thorough (1710 Prideaux,
Orig. Tithes v. 268; OED)

Before is a rare form with the meaning ‘previous’ and occurs with nouns
denoting actions/events (2). Ever, in the sense ‘everlasting, constant, perpetual’,
appears in an adjectival function beginning in the sixteenth century; the OED
points out that it occurs “chiefly with agent-nouns or sbs. of action” (3).
Evermore is a rare form, not recorded in the OED as an adjective (4). Hereafter,
in the sense of ‘future’, occurs in the sixteenth century with nouns of time (5).
Hitherto, with the exception of one eighteenth-century example, seems to be of
modern vintage. It has the meaning ‘thus far, to this point’ (6).
(2) before (OED s.v. before adv., prep. and conj., D.1)
Men are punisht for before breach of the Kings Lawes (1599 Shakespeare,
Henry V IV.i.179; OED)
Before tea, it was rather a dull affair, but then the before tea did not last
long (1796–1817 Jane Austen, Jane Austen’s Letters to her Sister Cassan-
dra and Others 127; U of V)
(3) ever/never (OED s.v. ever adv., I.6)
A neuer writer, to an euer reader (1609 Epist. Shaks. Tr. & Cr; (Qq. 1,2)
179; OED)
But the time of my euer farewell approacheth. (1580 Sidney, Arcadia
(1622) 481; OED)
(4) evermore
And frantic-mad with evermore unrest (1593–1600 Shakespeare, Sonnet
147; cited in Jespersen 1948: 355)
 Grammaticalization versus lexicalization reconsidered 79

(5) hereafter (OED s.v. hereafter adv. (a., n.), B)

And that hereafter Ages may behold/What ruin happen’d in revenge of
him (1591 Shakespeare, 1 Henry VI [Link].10; U of V)
For I myself have many tears to wash/Hereafter time (1592–93 Shake-
speare, Richard III [Link].389–90; cited in Jespersen 1948: 354)
(6) hitherto (OED s.v. hitherto adv. (a.), B)
All his hitherto offences (1787 Mad. D’Arblay, Diary (1842) III.303; OED)
most precious bequest to American civilization from all the hitherto ages
(1891–92 Whitman, Leaves of Grass 663; U of V)
that all the hitherto experience of the States (1891–92 Whitman, Leaves of
Grass 1008; U of V)
Hugh, the “hitherto baby”, if that is a possible term, sat in one corner.
(1886 Kate Douglas Wiggin, The Birds’ Christmas Carol 10; U of V)

The frequent form late has three central senses. Applied to persons, it means
‘deceased’, as shown in (7a) or, when attached to a role, ‘that was recently but
is not now, former’, as shown in (7b).16 Note that there can be ambiguity as to
whether late applies to the life or the term of office.17 Applied to events, it
normally means ‘recent’ (7c). All of these uses arise in the sixteenth century.18
A common form in the Helsinki Corpus is a late (?< on late).
(7) late (OED s.v. late a.1, A.5–6)
a. Graunts made by your noble progenitors, confirmed also by the late
King of famous memory, your noble father (1500–1570 Thomas
Wolsey, Original Letters II.19; HC)
Lay in the same bed & appartment where the late Queene lay
(1620–1706 Evelyn, The Diary of John Evelyn 902; HC)
b. It apperith in the booke of the antiquitees of the late monasterie of
Bath (Leland, The Itinerary of John Leland in or about the Years
1535–1543 1, 143; HC)
Here comes the Lady Widow, the late wife /To the deceas’d Sir Ava-
rice Golden-fleece (1657 Middleton, No Wit/Help Like a Womans
I.i.176–77; CHVD)
When didst thou see Timentes, the late Generall? (1631 Wilson, The
Swisser II.i.218; CHVD)
c. noblemen, and chief men in countreys, of the late conspiracy
(1537–53 Edward VI, The Diary of Edward VI 359; HC)
My Lord, I abhorred both the Principles and Practices of the late
Rebellion (1685 The Trial of Lady Alice Lisle IV, 122C2; HC)
80 Laurel J. Brinton

He undergoes my challenge, and contemnes it, / And threatens me

with the late Edict made /’Gainst duellists (1632 Massinger, The
Maid of Honour [Link].1–3; CHVD)

Now is of frequent use in the seventeenth century with the meaning ‘present,
existing’; it can be found with nouns denoting roles or titles (8a), actions or
events (8b) and occasionally with other nouns (8c).19
(8) now (OED s.v. now adv., conj., n.1, and a., IV.16a)
a. and the now King/The quondam Mounsieur shall not desire me this.
(1653 Heminges, The Fatal Contract [Link].525–36; CHVD)
thinkeing the now wearer of the Helmett, asmuch his Enemy (1655
Boyle, Parthenissa, Part 3 2.1.7; CHP)
b. On the Death of Mr. William Cartvvright, and the now publishing of
his Poems (1651 Cartwright, The Lady-Errant; CHVD)
c. he might prevent all her designs and rambles, the now Joy of her
Heart. (1687 Behn, The Amours of Philander and Silvia 387; CHP)
if the now cause of my Death could but extinguish the just provoca-
tions (1655 Boyle, Parthenissa, Part 4 2.8.356; CHP)
six hundred ninety nine shall be drawne distilled or made in his now
Distillary from any Malt Corne or Graine. (1640–1710 The Statutes
of the Realm (VII) 460; HC)

Though oft is now normally prefixed to and lexicalized with a following

participial adjective (e.g. oft-occurring, oft-repeated, oft-quoted), in the EModE
period it could modify a noun of action (9). The form often was common in the
sixteenth and seventeenth centuries with deverbal nouns denoting action (10a),
less frequently with nouns denoting roles (10b).
(9) oft (OED s.v. oft adv., a., B)
I ascribe my safety to myne oft fastynges (1548 Udall, etc. Erasm. Par.
Mark 746; OED)
I was moued by the oft reading and perusing of them, with a restles and
loft desire. (1585 Jas. I, Ess. Poesie (Arb.) 20; OED s.v. perusing vbl.n., 3)
(10) often (OED s.v. often adv. and a., B)
a. the often remembrance of his late receiued ioyes (1573 Gascoigne,
The Adventures of Master F. I. 53; CHP)
yet the often pronounciation of his letters, will be a means to help his
speech. (1660 Hoole, A new Discovery of the old Art of Teaching
Schoole 2; HC)
 Grammaticalization versus lexicalization reconsidered 81

b. Any favours, that may worthily make you an often courtier (1601 B.
Jonson, Poetaster [Link]; OED)

Once in the sense of ‘former’ occurs frequently with nouns of action (11a) and
role (11b), later with other nouns (11c). The OED distinguishes the first use
from the others, claiming that once in this use “can be explained as still an adv.
qualifying the vb.”
(11) once (OED s.v. once adv., (conj., a., n.), D.1)
a. Then is ye once sacrifice of Christ utterly to be abandoned and dis-
authorized (1548 Gest Pr. Masse in H. G. Dugdale Life (1840) App.
90; OED)
b. Slue the once owner of this vn-ioyn’d scull (1613 Stephens, Cinthia’s
Revenge [Link].33; CHVD)
c. We had a view of the dons of the once Capitol of the U. S. (1862
James E. Beard: Augusta County: Diary of James E. Beard
1861–08–08; 1862–08–28; U of V)

Quondam ‘former’ is common in the seventeenth century with nouns of role/

function (12a) and nouns of action (12b).
(12) quondam (OED s.v. quondam adv., n. and a., C)
a. This is the quondam King; Let’s seize vpon him. (1623 Shakespeare,
Henry VI, Part III I.i.; CHVD)
resolves to revenge her of this Wrong, and to reward her quondam
Lover (1683 Anon., The Dutch Rogue “The Life, Rise, and Fall of the
Decayed Merchant” 102; CHP)
b. The heighth of their quondam pride and cruelty (1642 Vicars, God in
Mount (1644) 44; OED)
not forgetting my Mistresses quondam kindnesses (1665 Richard
Head, The English Rogue, Part 1 XII.142; CHP)

Seldom ‘rare, infrequent’ is found rarely in the seventeenth century in an

attributive function (13). Since is a rare adjectival form with the meaning ‘that
has been since’ (14).
(13) seldom (OED s.v. seldom adv. and a., B)
For blunting the fine point of seldome pleasure (1593–1600 Shakespeare,
Sonnet 52; cited in Jespersen 1948: 354)
We stand amased now confounded by our seldome aspects. (1584 Rich,
Don Simonides, Tome 2 99; CHP)
82 Laurel J. Brinton

(14) since (OED s.v. since adv., prep. and conj. (also a. and n.), A.3b)
Eldest sonne of the since Earle of Norwich (1620–1706 Evelyn, The Diary
of John Evelyn 28 July 1641; OED)
Who … might very possibly give me an account of the since carriage and
deportment of Bess Bridges (1662 Dauncer, The English Lovers, Part 1
70; CHP)

Sometime is used in the sense of ‘former’ from the sixteenth to nineteenth

centuries, as exemplified in (15). The meaning ‘occasional’ is modern.
(15) sometime (OED s.v. sometime adv. (and a.), 1d and 2f)
I wonder now of my sometime boldness to chide and quarrel Christ.
(1637 Rutherford, Lett. (1862) I.254; OED)
none appeared more odious to him then his sometime bosom-friend faithful
Lamachus. (1659 Brathwait, Panthalia, or the Royall Romance 275; CHP)

Soon is used adjectivally in the sense of ‘early, speedy’ (16); the meaning
‘former’ is archaic.
(16) soon (OED s.v. soon a.)
comforts himself with the soon coming to the end (1639 J. S., Clidamus
3; CHP)
Mr. Boyd to be spoken to about the soon scaling of the Barony Kirk on
Sunday afternoon (1651 in Z. Boyd Zion’s Flowers (1855) Introd. 53;
OED s.v. skailing vbl. n.)

Then is a common attributive form dating back to the seventeenth century,

occurring with roles (17a) and actions/events (17b). It means ‘that existed or
was so at that time’.20
(17) then (OED s.v. then adv. (conj., a., n.), [Link].9b)
a. susanna martin the then wif of Georg martin being brought to Court
for a wich (1692 The Salem Witchcraft Papers, Vol. 2 572; U of V)
I. S. was really and truly a Prisoner in the Custody of E. F. the then
Sheriff Goaler or Keeper of the said prison. (1640–1710 The Statutes
of the Realm VII 75; HC)
b. Predicting … the happy, future State of our Country; and that the
then Fermentation would be perfective to it (1682 Early Anglesey,
State Govt. in Somers Tracts II.196; OED s.v. fermentation)
In this privacy I lived till the then War expired. (1662 Anon., The Life
and Death of Mary Frith 156; CHP)
 Grammaticalization versus lexicalization reconsidered 83

It may be well to let you know the then state of my mind. (Benjamin
Franklin, The Autobiography of Benjamin Franklin; U of V)
Scotland, the then Refuge of Traiterous transfugers (1611 Speed,
Hist. Gt. Brit. [Link] (1623) 1170; OED s.v. transfuge)

Thrice and twice are rare forms; note that both occur with nouns denoting
verbal actions (18). Whilom, like late, can have either the meaning ‘former’,
especially when modifying a role or rank, or ‘deceased’ (19) (see also Brinton
1999: 186–188). While may also occur in this sense. Yesterday means ‘immediate
past, very recent’ in its attributive use (20).
(18) thrice, twice (OED [Link]. thrice adv., 4; twice adv. (n., a.), 4)
S. Peter … after his relapse with thrise denial and forswearing of him
(1600 Watson, Decacordon (1602) 44; OED)
We heard of the twice returne of the Paragon (1624 Capt. Smith, Virginia
239; OED)
(19) whilom (OED [Link]. whilom adv. (a.), conj., A.2b; while adv. (a.), conj.
(prep.), A.2b)
But it against my whilome Lord did fight / With thee sweet Boy I came
(1613 E. Carew, Mariam [Link].40–41; CHVD)
The master of the mint our whilome refresher and consolation, now tooke
part against vs. (1594 Nashe, The Vnfortvnate Traveller 50; CHP)
(20) yesterday (OED s.v. yesterday adv., n. and a., C)
His Judgment dictated, that Yesterday Writers are most proper for mat-
ters of Antiquity (1665 J. Webb, Stong-Heng (1725) 41; OED)

Of this set of adverbial forms, the only ones found with adjectival use in
contemporary English — if one may judge from their occurrence in the British
National Corpus — are late and then (still common), after, before, now, some-
time and soonest (infrequent) and quondam (rare).
In the modern period, a number of adverbial forms, including aforetime,
erstwhile, long-ago, long-time, one-time, short-time, today and two-time, have
acquired an adjectival use. However, these are different in kind from the forms
under discussion. They are all derivationally complex, with a noun or noun-like
form (time, while, day and ago) as their second element. With the exception of
sometime and yesterday, the older forms are short, derivationally simple forms
going back to Old English (e.g. now) or are combined forms derived from an
adverb plus a non-nominal element, such as a particle (e.g. hitherto), another
adverb (e.g. hereafter), or an inflectional suffix (e.g. once).21
84 Laurel J. Brinton

4.3 Features of the change

A few generalizations can be made about this historical change. First, by far the
largest number of temporal adverbs seem to have come to be used as attributive
adjectives at roughly the same time, in the Early Modern English period.
However, for many of the forms, this use was transient and they are found only
adverbially in Modern English.
Second, there seems to have been a common discourse environment for the
change; it occurred in the context of nouns denoting actions or events (de-
verbalized nouns), agent nouns, or nouns of rank, role, or relation. It should
also be noted that the adverbs appear frequently as attributive modifiers with
gerunds (see examples 8b, 9, 16, 18).22 Nouns from other semantic classes were
not impossible in this context, but they were quite rare.
Third, the shift from adverb to adjective is limited to temporal adverbs
and a few spatial adverbs,23 but manner adverbs and other types of adverbs are
not allowed.
Fourth, the semantic change is also remarkably consistent. In addition to a
high degree of specialization in terms of past time, the majority of forms acquire
the meaning of ‘former, previous’ when used attributively, including both
adverbs denoting anterior time (aforetime, before, erstwhile, late, long-ago,
quondam, since, then, whilom, yesterday) and those which are originally frequen-
tative (once, one-time, sometime). Semantic change in the other forms is not as
regular, though adverbs denoting posterior time tend to acquire the meaning
‘subsequent, later’ (after, hereafter).
Fifth, the forms continue to function primarily as temporal adverbs and
their use as adjectives is limited morphosyntactically. They may occur only in
attributive, not in predicative position:
(21) *My friend is sometime. ?His wife is now.

*My husband is late (= ‘former’). *The effects are after.

*The kingdom is then. *The picture is before.

In this function, the forms are not susceptible to inflection for degree, modifica-
tion by intensifiers, or the usual derivational processes expected of adjectives
(e.g. -ly, -ness, un-):
(22) *my later husband/*the oftener response
*my rather late husband/*the very then kingdom/*the quite after effects
*the thenness of the ruler/*the unthen ruler/*he ruled thenly
 Grammaticalization versus lexicalization reconsidered 85

The occurrence of comparative/superlative forms of oft, often, seldom and soon

(23) might suggest the morphological status of these forms as adjectives, but it
must be remembered that adverbs too inflect for degree.
(23) to breede occasion of ofter meeting of him and her (1568 Ascham,
Scholem. (Arb.) 85; OED s.v. oft adv., a., B)
to have the oftner accesse vnto Laurana (1598 Forde, Parismus, Part 1
32; CHP)
Liable to an oftner anger (1640 Bp. Reynolds Passions xii; OED s.v. often
adv. and a., B)
Yet amongst my manie volumes, I hope Gods booke hath not beene my
sildomest lectures. (Q. Eliz. In Holinshed’s Chron. (1587) III.1396/2; OED
s.v. seldom adv. and a., B)
that which is the most desirable and the seldomest found in Aristocracies
(1655 Boyle, Parthenissa, Part 3 2.3.248; CHP)
He that hath (…) his gold ready shall have a sooner dispatch, then the
best Scholar upon ticket. (1656 Heylin, Surv. France 147; OED s.v. ticket
n.1, 7)
make your soonest haste; So your desires are yours. (1606 Shakespeare,
Antony and Cleopatra [Link].27–28; U of V)
for the greatest flowe hath the soonest ebbe (1583 Greene, Mamillia 22;
CHP)

The most common of the adverbs are also found conjoined syndetically (24) or
asyndetically (25) with other adjectives. While such conjunction might likewise
suggest their full status as adjectives, it is not uncommon in earlier English to
find cases of asymmetric coordination.24
(24) The oft and frequent welcomes giuen my sonne (1633 Heywood, The
English Traveller III.i.7; CHVD)
the often and free chaunging of persons (1595 Sidney, Defence of Poesie; U
of V)
The often and much vse of lettuce … hindreth procreation, …and maketh
the body lumpish. (1620 Venner, Via Recta vii.141; OED s.v. lumpish a.)
Long, impertinent, and often Epithets (1679 Hobbes, Rhet. [Link] (1681)
105; OED s.v. often adv. and a., B)
there are no signs amongst them of sooner or later production (1677 Plot
Oxfordsh. 110; OED s.v. soon a., 2)
The greatest and oftenest laugher (1831 Carlyle, Sart. Res. i.v; OED s.v.
often adv. and a., B)
86 Laurel J. Brinton

A suppressed and seldom anger (1650 Jer. Taylor, Holy Living ii.§29
(1727) 59; OED s.v. seldom adv. and a., B)
the then and still owners of that happy periodical (1879 Trollope,
Thackeray 22; OED s.v. still adv., 4a)
(25) a late most learned Writer (1675–76 Boyle, Electricity and Magnetism 35;
HC)
the late most dreadfull storm (1640–1710 Letters of Sir Richard Haddock
45; HC)
the quondam laborious antiquary (1840 Penny Cycl. XVIII.167/1 [Pinker-
ton]; OED s.v. petit-maitre)

4.4 Accounting for the change

There would seem to be a number of possible ways to explain this phenomenon:
a. as a matter of word formation, i.e. conversion or functional shift,
b. as a process of lexicalization,
c. as a process of grammaticalization,
d. as a counterexample to grammaticalization, viz. degrammaticalization, or
e. as a matter of ‘recategorialization’.

4.4.1 Conversion or functional shift

While Bauer (1983: 226) considers conversion a “totally free process”, devoid of
constraints, in which “any lexeme can undergo conversion into any of the open
form classes as the need arises”, he does not discuss the conversion of adverbs
to adjectives, nor do other standard treatments of word formation (such as
Marchand 1969: 359ff.). Bauer notes — in regard to examples of N > A conver-
sions such as head teacher, model airplane — that the conversion to adjectives is
often problematic: “when the form is used attributively, criteria for concluding
that conversion has taken place must be spelled out with great care” (1983: 228).
Quirk et al. (1985: 1562) require that for such a conversion to have occurred, it
must be possible for the form to be used predicatively and there must be
inflectional (and derivational) evidence of the word’s status as an adjective. If a
suspected conversion fails these tests — as ours do — it should be accounted
for in syntactic terms.25
By “syntactic terms” they are suggesting the account favored by traditional
grammarians, namely, “X acting like a Y”, or “an adverb acting like an adjec-
tive”. Apart from the problems inherent in such an explanation — in what
sense is X really an X or really not an X, or a Y or really not a Y? (see
 Grammaticalization versus lexicalization reconsidered 87

Huddleston 1984: 93ff.) — this sort of description provides no diachronic

explanation of the shift from adverb to adjective.
Alternative accounts might be in terms of “intersective gradience”26 — a
theory currently being explored by Bas Aarts and David Denison (see Denison
2001), which argues that there may well be “gradience” or soft boundaries
between word classes such as N and A — or in terms of the hybridity of lexical
categories, or in terms of a supercategory such as Modifier, which would
encompass both adjectives and adverbs. Likewise, the fact that the adjectival and
adverbial classes were less well defined in Early Modern English than in Modern
English might explain the phenomenon.27 While these suggestions can account
for the varying degree of membership in the category Adjective of the forms
under discussion and the gradualness of the change, I do not think that they
address questions about the mechanism or directionality of the change.

4.4.2 Lexicalization
Perhaps the most obvious conclusion to reach is that the changes here are an
example of lexicalization, describable by the fourth definition of lexicalization
given in Section 3.1. That is, they show the functional shift from grammatical
word > lexical word, in the sense that adverb is more grammatical (member of
a more closed set) than adjective, adjective more lexical (member of a more
open set) than adverb. They could be treated as an instance of what above (see
Section 3.1) was termed ‘secondary lexicalization’, representing a change
toward greater lexical function, or categoriality.
However, it is not intuitively clear to me that there is an absolute, or even
a relative, difference in degree of lexicality between adjectives and adverbs. In
general, the distinction between grammatical and lexical is rather fuzzy; it is
clearly language dependent, with the same concept often being expressed
lexically in one language and grammatically in another (Lass 2000; Traugott
forthcoming b). Moreover, there would seem to be an even more serious
problem with viewing this change as an instance of lexicalization. Typically, the
process of lexicalization is idiosyncratic. It consists of unexpected changes, one-
off occurrences, following no general pattern and affecting individual forms.
Both the forms chosen for lexicalization and the type of lexicalization they
undergo are unpredictable. However, in the case of temporal adverbs used as
adjectives, the change is widespread; it affects an entire class of forms. More-
over, it is to a large extent regular, since temporal (and sometimes spatial)
adverbs, but not manner adverbs, for example, can undergo the change.
88 Laurel J. Brinton

The semantic changes involved in this change are also atypical of lexical-
ization. The semantic changes in lexicalization are thought to be “unrecover-
able” (Quirk et al. 1985) or “unpredictable” (Hagège 1993) and to involve the
addition of semantic content (Wischer 2000: 364–365). However, the semantics
of the forms discussed here involve either the widening of semantic scope (from
‘before’, ‘one-time’, ‘late’, etc. to ‘former’) or are recoverable from the temporal
meaning of the corresponding adverb.

4.4.3 Grammaticalization
A third possibility, given the apparent regularity and systematicity of the
change, is to view the change from freely occurring temporal adverb to highly
restricted attributive modifier as an instance of grammaticalization. Considered
in respect to the features of grammaticalization set out in Lehmann (1985, 1995:
Chapter 4), these forms exhibit the following:
a. semantic “attrition”, or desemanticization, undergoing restriction of
semantic content, with broadening or loss of meaning;
b. “condensation” in scope; changing from modifiers of the entire VP or S to
modifiers of the N within the NP; and
c. “fixation”, coming to occupy a fixed slot, pre-nominally.
While some of the features of grammaticalization noted by Lehmann, including
most importantly phonological “attrition”, or erosion of phonetic substance,
and morphological “coalescence”, or bonding, do not occur in this instance,
these developments are absent in standardly-accepted examples of grammatical-
ization, such as the development of auxiliaries.28
Considered in respect to the principles of grammaticalization set out by
Hopper (1991), this change is characterized by:
a. “divergence”, with the forms remaining as fully lexical temporal adverbs;
b. “persistence”, with the forms retaining traces of their original meaning in
their grammaticalized form and possibly
c. “decategorialization”, with the forms “los[ing] or neutraliz[ing] the mor-
phological markers and syntactic privileges” of their category (1991:22).
Although as a class adverbs have few morphological markers, they are character-
ized by syntactic mobility and wide (predicate or sentence) scope. The forms
under discussion here become fixed syntactically and reduced in scope. Further-
more, they do not acquire the morphological and syntactic characteristics of
adjectives. Thus, they can be seen as “decategorialized”.
 Grammaticalization versus lexicalization reconsidered 89

The semantics involved in the use of temporal adverbs as adjectives are also
more characteristic of grammaticalization than of lexicalization: first, the lexical
sources — rather non-specific anterior time adverbials — are sufficiently
general in sense and second, the meaning shifts move in the direction more
referential to less referential, or more concrete (position in time) to more
abstract (temporal quality). Both very general lexical sources and abstraction of
meaning are considered central to grammaticalization (see Traugott forthcom-
ing b). Finally, the fact that the change occurs in a highly constrained discourse
context — namely in the context of nouns of action, role, or rank — agrees
with current thinking about grammaticalization, as discussed, for example, in
Traugott (forthcoming a). Also, one could argue that the locus for the reanalysis
(which is central to grammaticalization) of the adverb as an adjective is the
position before the ‘hybrid’ gerund form, which would seem to permit either
adverbial or adjectival modification.

4.4.4 Degrammaticalization
There is, however, an important respect in which the forms discussed here do
not undergo Hopper’s central process of decategorialization. Decategoriali-
zation is typically understood as involving a unidirectional cline from more
major to more minor category or from more open to more closed form class.
Our forms do not exhibit the expected downgrading of categorial status; rather,
as suggested before, they seem to move from a more minor to a more major
word class (being upgraded from adverb to adjective). The shift from more to
less grammatical might in fact be viewed as an example of “categorialization”,
or degrammaticalization.
However, there would seem to be two reasons to question whether the shift
examined here is indeed degrammaticalization. If it were, it would be quite rare.
For example, while Newmeyer (1998: 272–274) cites numerous examples of
“upgrading” from functional category to lexical category, including the shifts
from Preposition > Verb, Pronoun > Verb, Preposition/Conjunction >
Adjective, Preposition/Conjunction > Noun and Pronoun > Noun, he gives no
examples of the upgrading from Adverb > Adjective, nor have I found any cited
in the literature. Furthermore, it is unclear to me from the definitions given in
3.2 whether degrammaticalization necessitates that there be some prior gram-
maticalization process that is undergoing reversal. In some of the cases that I
have examined, there is a preceding grammaticalization process: the old
inflected forms such as once, seldom, whilom have undergone “phonogenesis”,
which can be viewed as the end stage of grammaticalization (see Section 3.2
90 Laurel J. Brinton

above). The compounds sometime and yesterday, with a noun as second

element, or the compounds evermore, hereafter, hitherto (and older before,
since), with a particle as second element, all involve univerbation, which is often
a feature of grammaticalization (though here it may represent lexicalization).
However, in most instances the shift from adverb to adjective does not undo or
reverse in any direct way a prior grammaticalization process. In a number of the
cases (e.g. after, ever, oft/often, then, now, late, soon), there is no evidence that
the forms began as anything other than adverbs.29

4.4.5 Recategorialization/refunctionalization
There are two less obvious ways to account for the development of temporal
adjectives from adverbs. The first is “recategorialization”. According to Heine,
Claudi & Hünnemeyer (1991:213, 233–238), the losses brought about by
decategorialization may be compensated for by gains, in the form of “recategor-
ialization”: “a process whereby language tends to restore iconicity between form
and meaning. It has the effect that the ‘hybrid forms’ (…) resulting from
decategorialization develop into new, function-specific morphemes”
(1991:213). Recategorialization involves a form acquiring a new categorial status
(1991:238) or the acquisition of new grammatical structures, syntactic modes, or
morphological patterns. For example, when I think is grammaticalized as an
epistemic parenthetical, it loses status as a matrix clause but gains status as an
adverb; it also acquires greater syntactic freedom. However, recategorialization
would seem explicitly to require a prior grammaticalization process, which, as I
noted before, is not always the case here.
A second way to account for our shift is as “exaptation” (Lass 1990) or
“refunctionalization under conditions of discontinuity”, which, as Giacalone
Ramat (1998) explains it, involves an item sliding to an adjacent (and more
central) area of grammar (in contrast to lexicalization, where the item slides
into the area of the lexicon). The difficulty in explaining our process in this way
is that we do not begin — as Lass so colorfully describes it — with “junk”, or
linguistic material which has lost its grammatical function, but rather with fully
functional material.

5. Conclusion

In conclusion, in regard to the use of temporal adverbs as attributive modifiers,

most especially in the Early Modern English period, we would seem to be
 Grammaticalization versus lexicalization reconsidered 91

dealing with a shift which challenges the conventionally recognized processes of

change. We are clearly not witnessing the process of conversion and despite the
temptation to see it as lexicalization, the systematic nature of the change would
rule out lexicalization. The lack of a prior process of grammaticalization
excludes regrammaticalization and possibly degrammaticalization. If the change
is analyzable as grammaticalization it calls into question the clines proposed for
decategorialization or at least shows that, outside of the core categories of noun
and verb, atypical movements are possible.30 However, it might be the case that
degrammaticalization shares many of the features of grammaticalization but
simply with a different directionality. Such an insight might lead to a somewhat
better understanding of the degrammaticalization process.
This paper may have contributed to an understanding of grammatical-
ization, lexicalization and degrammaticalization — or simply muddied the
water. But it has shown, I think, that an extended example such as this calls into
question all three processes and begs for a clearer definition of each.

Notes

1. As Newmeyer (1998: 253) observes, “functional categories have a characteristic semantic

value regardless of whether they represent downgradings from lexical categories”.
2. For the purposes of this paper, I am assuming — most likely incorrectly — that the
meaning of ‘grammaticalization’ is generally agreed upon.
3. Some scholars use the term ‘relexicalization’ in the sense of lexicalization (Hagège
1993: 209), though for Ramat (1992: 554) relexicalization denotes movement back to the
lexicon in a spiral.
4. Van der Auwera’s (2002) definition of lexicalization (“the making of a lexical item out of
something other than a lexical item” ) would seem to belong to this category, but his example
of songwriter suggests the third definition.
5. This change might also be seen as ‘demorphologization’ (see below).
6. See also Traugott (forthcoming b), though she describes this change as “demorpholog-
ization” (a species of lexicalization for her; see below).
7. According to Traugott (forthcoming b), there is no change from less to more lexical in
lexicalization comparable to the change from less to more grammatical in grammaticalization.
8. Definitions 1 and 2 are overly general and will not be considered further. Moreover, I wish to
distinguish lexicalization from general processes of word formation (definition 3), which involve
the shift from lexical > lexical (not grammatical > lexical). The process of semanticization
(definition 9) seems rather different in kind and will also not be considered further in this paper.
92 Laurel J. Brinton

9. The number of such examples is growing. See especially Newmeyer (1998: 263ff.) for
examples of “upgradings”, which he feels are “rampant”. However, Newmeyer makes no
distinction among the different kinds of upgradings he discusses.
10. A somewhat more specialized sense of degrammaticalization is proposed by Páez
Urdaneta (1982), who argues that when a grammatical word loses function at the proposi-
tional or textual level and acquires new functions at the conversational level, that is,
“performative” or sociostylistic functions, then it can be said to be degrammaticalized. Such
a change would correspond to the development of pragmatic markers, which I have treated
as an instance of grammaticalization (Brinton 1996).
11. This sequence is cited for illustrative purposes only and probably does not correspond to
the actual development of these forms (see Allen 1997).
12. For Joseph & Janda (1988) demorphologization also involves the movement of morphol-
ogy into syntax.
13. Lehmann (1995: 14) points out that Givón refers to demorphologization as lexicalization.
14. Data for this paper were gathered from the Oxford English Dictionary (2nd edition on
CD-ROM; OED), from the Early English Prose and Verse Drama sections of Chadwyck
Healey (CHP, CHVD, respectively), from the University of Virginia Modern English
Collection (U of V), from the Helsinki Corpus (HC; see Kytö 1996) and from the British
National Corpus (BNC).
15. After seems to have been confused with aft ‘behind’ in expressions such as the after part,
the after deck, the after locker.
16. In a discussion of the semantics of what he calls “separative adjectives”, such as former or
late in Susie’s former husband lives in Senegal or Eddy’s old school keeps asking him for money,
Ferris (1991: 577–579) points out that the adjectives do not qualify the properties, or
descriptive content, of the nouns (husband, school), but rather “the RELATION between the
entity which is seen as a participant in the structure of the sentence communicated, and the
description which characterizes and identifies that entity” (578). Thus, by analogy, we could
say that in the case of late (which has the same meaning as former and old in Ferris’s
examples), as in the late wife (of Sir Avarice), it is the relation between the woman and the
description (wife of Sir Avarice) that is ‘former’.
17. Because of this possible ambiguity, Fowler (1983: [Link]. late et al.) suggests that late should
be reserved for the meaning ‘dead’. Webster’s (1989: s.v. late) reports that much ink has been
spilled in usage guides about how long one may use late after a person has died; it concludes
that the time span is unlimited since late means ‘the now dead’ and furthermore functions
as a sign of respect.
18. The attributive use of the adverb differs from the ‘true’ adjective late, which is of ancient
provenance (see OED s.v. late a.1, A). According to Jespersen (1948: 355), the ‘deceased’
meaning of late must originate in the adverbial use; likewise, the OED considers the ‘recent’
meaning apparently to have developed from the adverb late meaning ‘not long since,
recently’ (OED s.v. late adv., 4b).
19. In the 1960s attributive now acquired a second meaning, namely, ‘fashionable, current,
up-to-date’ (OED s.v. now adv., conj., n.1 and a., IV.16b), but this use seems to have been
 Grammaticalization versus lexicalization reconsidered 93

rather transient. Webster’s (1989: s.v. now) notes that while the older usage generated no
animosity, there was considerable objection to the newer use.
20. According to Webster’s (1989: s.v. then), some objection was raised to this usage in the
nineteenth century and again in the twentieth century, but “the controversy never got off the
ground”.
21. Examples of the modern forms are the following:
a. aforetime (OED s.v. aforetime adv.)
Believing not the aforetime unity Of the Divine and human. (1846 Grote, Greece
(1862) i.i.37; OED)
b. erstwhile (OED s.v. erstwhile adv. and a., B)
this incident leading to a further quarrel, the erstwhile friends parted. (1922, Stanley
H. Redgrove, Alchemy: Ancient and Modern 69; U of V)
c. long-ago (OED s.v. long-ago).
A book, the long-ago gift of his dead mother (1889 Chicago Advance 24 Jan.; OED)
d. long-time (OED s.v. long a.1, [Link].18)
Big Tim, “the long-time leader of the Sullivans” (1915 Anne Conway, Acres of Dia-
monds and His Life and Achievements; U of V)
e. oft-time (OED s.v. oft-time adv. (a.))
The oft-time Premier of the Colony (1896 Daily News 12 Sept. 5/1; OED)
f. one-time (OED s.v. one numeral a., pron., etc., [Link].35)
Old Lodge, we salute thee for thy venerable antiquity; but we owe thee no respect
as the one-time resort of the boasted virgin queen! (1850 W. Howitt Year-bk.
Country vi.179; OED s.v. one)
g. short-time (OED [Link]. short a., n. and adv., A.V.23; short time)
In the army they have *short-time soldiers and long-time soldiers (1877 Spurgeon,
Serm. XXIII. 130; OED s.v. short)
h. today (OED s.v. today adv., n. and a., C)
I’m a today writer. (1969 Harper’s Mag. Oct. 65/2; OED)
i. two-time (OED s.v. two-time a., 1)
i. Is a widder, even a two-time widder, got nothin’ else to do but … go about
grievin’ for them that’s gone? (1897 R.M. Johnston, Middle Georgia 113; OED)
ii. The loss of the two-time former Scottish champion is a severe blow (KSH 4275;
BNC)
iii. two-time pieces or *sea-watches (1767 Ann. Reg. X.I 141/1; OED s.v. sea n.)
Like quondam and whilom, erstwhile (b) has an archaic or pedantic quality (see Webster’s
1989: [Link]. erstwhile, quondam, whilom; Fowler 1983: [Link]. late et al.), so it is interesting to
note that it is a modern development. Oft-time (e) and one-time (f) occur with the sense of
‘former’. Today (h) is similar to the second meaning of now (see footnote 19), namely
‘modern’, rather than ‘present’. Two-time, according to the OED, is a non-standard usage,
presumably of American origin (i-i), but it is found currently in collocation with ‘winning’
and ‘losing’ in sportswriting, discussions of election results, etc. (i-ii), and it may be of even
earlier origin, judging from one example in the OED (i-iii).
94 Laurel J. Brinton

Teresa Fanego (p.c.) has pointed out that these forms may have a different derivation
from the forms discussed in this paper. She suggests that they may simply be attributive uses
of the corresponding NPs: e.g. a long-ago gift < a gift of long ago. This suggested derivation
works for some of the forms but is awkward with others (e.g. a one-time resort < ?a resort of
one time); nor could the semantic change in the latter (from ‘one-time’ to ‘former’) be
adequately accounted for by this derivation.
A different explanation of the origin of these forms is proposed by Poutsma (1926:701),
who suggests that sometime and onetime derive from the phrases at sometime and at onetime.
22. Teresa Fanego has argued (see, e.g. Fanego 1998) that the gerund in Early Modern
English was a hybrid form, showing increasingly verbal qualities, yet not having lost the
abstract nominal qualities of its origin. She suggests that it might therefore be possible to
consider the forms that I am discussing, when they precede gerunds, as adverbs rather than
as adjectives. From gerunds their use would have spread to other deverbal nouns.
23. Jespersen cites examples such as my hence departure, his downward progress, the hither
side of thirty, the above letter, the off side (1948: 355–358). The referees of the paper have
noted that other types of adverbs may undergo this shift, citing only and well. However, such
a use of only (e.g. the only refusal, his only dethe) would seem to be very rare (see Nevalainen
1991: 131) and according to the OED (s.v. well adv., VII.30a–b), well occurs attributively only
with gerunds.
24. I am indebted to Teresa Fanego for this observation.
25. Note that conversion is included in the regular processes of word formation labeled as
‘lexicalization’ in the third definition above. As such, it is subject to the same questions which
I will discuss below in respect to that process.
26. As a formalist, Newmeyer (1998: 247, 290) argues strongly against the possibility of
categorial gradience on the synchronic level. Though he admits that diachronically certain
aspects of categorial change may be gradual, he points out that not every step has “categorial
significance”, i.e. results in a change in categorial status.
27. The suggestions were made by Elizabeth Traugott and Ans van Kemenade during
discussion which followed this paper at 11 ICEHL and by Teresa Fanego, who bases her
hypothesis on the less consistent use of -ly in deriving adverbs from adjectives in Early
Modern English. However, while this might explain why derived adverbs (e.g. easy) are
identical to adjectives, it does not explain why derived adjectives are identical to adverbs.
28. There is also neither “paradigmaticization” nor “obligatorification”.
29. I agree here with Lass (2000) in rejecting what he terms “strong UD [unidirectionality]”,
the view that all grammatical items must originate in lexical items, that there are no
originally grammatical forms.
30. The atypicality might account for the transience of the grammaticalizations we see here.
 Grammaticalization versus lexicalization reconsidered 95

References

Allen, Andrew S. 1995. “Regrammaticalization and Degrammaticalization of the Inchoative

Suffix”. Historical Linguistics 1993: Selected Papers from the 11th International Conference
on Historical Linguistics ed. by Henning Andersen, 1–8. Amsterdam & Philadelphia:
John Benjamins.
Allen, Cynthia. 1997. “The Origins of the ‘Group Genitive’ in English”. Transactions of the
Philological Society 95: 1.111–131.
Anttila, Raimo. 1989. Historical and Comparative Linguistics. Amsterdam & Philadelphia:
John Benjamins.
Auwera, Johan van der. 2002. “More Thoughts on Degrammaticalization”. Wischer &
Diewald 2002. 19–30.
Bauer, Laurie. 1983. English Word Formation. Cambridge: Cambridge University Press.
Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse
Functions. Berlin & New York: Mouton de Gruyter.
Brinton, Laurel J. 1999. “‘Whilom, as Olde Stories Tellen us’: The Discourse Marker Whilom
in Middle English”. From Arabye to Engelond: Medieval Studies in Honour of Mahmoud
Manzalaoui on his 75th Birthday ed. by A. E. Christa Canitz & Gernot R. Wieland,
175–199. Ottawa: University of Ottawa Press.
British National Corpus, [Link]
Bussmann, Hadumod. 1996. Routledge Dictionary of Language and Linguistics, trans. & ed. by
Gregory Trauth & Kerstin Kazzazi. London & New York: Routledge.
Chadwyck-Healey, English Verse Drama, [Link]
Chadwyck-Healey, Early English Prose, [Link]
Chen, Guohua. 1998. “The Degrammaticalization of Address-Satisfaction Conditionals in
Early Modern English”. Advances in English Historical Linguistics ed. by Jacek Fisiak &
Marcin Krygier, 23–32. Berlin & New York: Mouton de Gruyter.
Denison, David. 2001. “Gradience and Linguistic Change”. Historical Linguistics 1999:
Selected Papers from the 14th International Conference on Historical Linguistics ed. by
Laurel J. Brinton, 119–144. Amsterdam & Philadelphia: John Benjamins.
Fanego, Teresa. 1998. “Developments in Argument Linking in Early Modern English Gerund
Phrases”. English Language and Linguistics 2.87–119.
Ferris, Connor. 1991. “Time Reference in English Adjectives and Separative Qualification”.
Linguistics 29.569–590.
Fischer, Olga. 2000. “Grammaticalization: Unidirectional, Non-Reversable? The Case of to
before the Infinitive in English”. Fischer, Rosenbach & Stein 2000. 149–169.
Fischer, Olga, Anette Rosenbach & Dieter Stein, eds. 2000. Pathways of Change: Gramma-
ticalization in English. Amsterdam & Philadelphia: John Benjamins.
Fowler, H. W. 1983. A Dictionary of Modern English Usage. 2nd ed. revised by Sir Ernest
Gower. Oxford & New York: Oxford University Press.
Giacalone Ramat, Anna. 1998. “Testing the Boundaries of Grammaticalization”. Giacalone
Ramat & Hopper 1998.107–127.
Giacalone Ramat, Anna & Paul J. Hopper, eds. 1998. The Limits of Grammaticalization.
Amsterdam & Philadelphia: John Benjamins.
96 Laurel J. Brinton

Hagège, Claude. 1993. The Language Builder: An Essay on the Human Signature in Linguistic
Morphogenesis. Amsterdam & Philadelphia: John Benjamins.
Haspelmath, Martin. 1999. “Why is Grammaticalization Irreversible?” Linguistics
37.1043–1068.
Heine, Bernd, Ulrike Claudi & Friederike Hünnemeyer. 1991. Grammaticalization: A
Conceptual Framework. Chicago: University of Chicago Press.
Hopper, Paul J. 1991. “On Some Principles of Grammaticization”. Approaches to Gram-
maticalization ed. by Elizabeth Closs Traugott & Bernd Heine, vol. I, 17–35. Amsterdam
& Philadelphia: John Benjamins.
Hopper, Paul J. 1994. “Phonogenesis”. Perspectives on Grammaticalization ed. by William
Pagliuca, 29–45. Amsterdam & Philadelphia: John Benjamins.
Hopper, Paul J. & Elizabeth Closs Traugott. 1993. Grammaticalization. Cambridge: Cam-
bridge University Press.
Huddleston, Rodney. 1984. Introduction to the Grammar of English. Cambridge: Cambridge
University Press.
Jespersen, Otto. 1948 [repr. London, 1961]. A Modern English Grammar on Historical
Principles. Part II: Syntax (First Volume). Copenhagen: Ejnar Munksgaard / London:
George Allen & Unwin.
Joseph, Brian D. & Richard D. Janda. 1988. “The How and Why of Diachronic Morpholog-
ization and Demorphologization”. Theoretical Morphology: Approaches in Modern
Linguistics ed. by Michael Hammond & Michael Noonan, 193–210. San Diego: Academ-
ic Press.
Kruisinga, Etsko. 1932 [5th ed.]. A Handbook of Present-Day English. Part II. English
Accidence and Syntax, vol. III. Groningen: P. Noordhoff.
Kuryłowicz, Jerzy. 1965. “The Evolution of Grammatical Categories”. Diogenes 51.55–71.
Kytö, Merja. 1996 [3rd ed.]. Manual to the Diachronic Part of the Helsinki Corpus of English
Texts: Coding Conventions and Lists of Source Texts. Helsinki: Department of English,
University of Helsinki.
Lass, Roger. 1990. “How to Do Things with Junk: Exaptation in Language Evolution”.
Journal of Linguistics 26.79–102.
Lass, Roger. 2000. “Remarks on (Uni)directionality”. Fischer, Rosenbach & Stein 2000.
207–227.
Lehmann, Christian. 1985. “Grammaticalization: Synchronic Variation and Diachronic
Change”. Lingua e stile 20.303–318.
Lehmann, Christian. 1995. Thoughts on Grammaticalization. (= Lincom Studies in Theoretical
Linguistics, 01). München & Newcastle: Lincom Europa.
Lehmann, Christian. 2002. “New Reflections on Grammaticalization and Lexicalization”.
Wischer & Diewald 2002. 1–18.
Lessau, Donald A. 1994. A Dictionary of Grammaticalization. (= Essener Beiträge zur
Sprachwandelforschung, 21). 3 vols. Bochum: Universitätsverlag Dr. N. Brockmeyer.
Marchand, Hans. 1969. The Categories and Types of Present-Day English Word-Formation. (=
Handbücher für das Studium der Anglistik). München: C. H. Beck.
Moreno Cabrera, Juan C. 1998. “On the Relationship between Grammaticalization and
Lexicalization”. Giacalone Ramat & Hopper 1998. 209–227.
</TARGET "bri">

Grammaticalization versus lexicalization reconsidered 97

Modern English Collection. University of Virginia Electronic Text Center,

[Link]
Nevalainen, Terttu. 1991. BUT, ONLY, JUST: Focusing Adverbial Change in Modern English
1500–1900. (= Mémoires de la Société Néophilologique de Helsinki, 51). Helsinki: Société
Néophilologique.
Newmeyer, Frederick J. 1998. Language Form and Language Function. Cambridge, Mass. &
London: MIT Press.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S. C. Weiner. Oxford: Oxford University Press.
Páez Urdaneta, Iraset. 1982. “Conversation pues in Spanish: A Process of Degrammatical-
ization?” Papers from the 5th International Conference on Historical Linguistics ed. by
Anders Ahlqvist, 332–340. Amsterdam & Philadelphia: John Benjamins.
Poutsma, Hendrik. 1926. A Grammar of Late Modern English. Part 2: The Parts of Speech,
Section 2: The Verb and the Particles. Groningen: P. Noordhoff.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive
Grammar of the English Language. London & New York: Longman.
Ramat, Paolo. 1992. “Thoughts on Degrammaticalization”. Linguistics 30.549–560.
Schwenter, Scott A. & Elizabeth Closs Traugott. 1995. “The Semantic and Pragmatic
Development of Substitutive Complex Prepositions in English”. Historical Pragmatics:
Pragmatic Developments in the History of English ed. by Andreas H. Jucker, 243–273.
Amsterdam & Philadelphia: John Benjamins.
Tabor, Whitney & Elizabeth Closs Traugott. 1998. “Structural Scope Expansion and
Grammaticalization”. Giacalone Ramat & Hopper 1998. 229–272.
Traugott, Elizabeth Closs. 1994. “Grammaticalization and Lexicalization”. The Encyclopedia
of Language and Linguistics ed. by R. E. Asher. Oxford: Pergamon Press.
Traugott, Elizabeth Closs. 2000. “From Etymology to Historical Pragmatics”. Plenary
Address at SHEL-1 (Studies in the History of the English Language 1), Los Angeles, 27
May 2000.
Traugott, Elizabeth Closs. Forthcoming a. “Constructions in Grammaticalization”. Hand-
book of Historical Linguistics ed. by Brian Joseph & Richard Janda. Oxford: Blackwell.
Traugott, Elizabeth Closs. Forthcoming b. “Lexicalization and Grammaticalization”.
Lexikologie — Lexicology ed. by A. A. Cruse, F. Hundsnurscher, M. Job & P. R. Lutzeier.
Berlin & New York: Mouton de Gruyter.
Webster’s Dictionary of English Usage. 1989. Springfield, Mass.: Merriam-Webster.
Wischer, Ilse. 2000. “Grammaticalization versus Lexicalization: ‘Methinks’ There is Some
Confusion”. Fischer, Rosenbach & Stein 2000.355–370.
Wischer, Ilse & Gabriele Diewald, eds. 2002. New Reflections on Grammaticalization.
Proceedings from the International Symposium on Grammaticalization (Potsdam,
Germany, 17–19 June 1999). Amsterdam & Philadelphia: John Benjamins.
<LINK "kas-n*">

<TARGET "kas" DOCINFO

AUTHOR "Dieter Kastovsky"

TITLE "The derivation of ornative, locative, ablative, privative and reversative verbs in English"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

The derivation of ornative, locative, ablative,

privative and reversative verbs in English
A historical sketch*

Dieter Kastovsky
University of Vienna

1. Introduction

Ornative, locative, ablative, privative and reversative verbs represent the most
productive domain of Modern English verb-formation, i.e. the derivation of
verbs by word-formation processes from nouns, adjectives and verbs. Modern
English examples illustrating these semantic categories are:
(1) ornative: to bedew, encrown, chlorinate, alcoholize, salt; locative: to encage,
bottle; ablative: to deplane, disbar, unsaddle; privative: to behead, defrost,
stone; reversative: to demilitarize, disengage, unlock, unbutton

The meanings of these verbs can all be related to a common underlying

semantic structure (cf. also Kastovsky 1973), viz.:
(2) [[agent]] cause theme (t) become [not] be in location (l)1

i.e. some Theme (= Object) is caused to be located in some Location (= Place)

or is removed from this Location (= Place). More generally, all these verbs
belong to the major semantic category of causative-inchoative verbs, which
need not necessarily be morphologically complex, cf. give ‘Agent causes
something to be in someone’s possession’ vs. take ‘Agent causes something to
be no longer in someone’s possession’; similarly lend vs. borrow, put/add vs.
remove/deprive and many others. It is not unlikely that this basic semantic
structure has a universal cognitive-conceptual foundation (cf. Kastovsky
1996a: 205f.) reflecting the basic human activity of moving objects around in
space. Thus, the Theme can be regarded as a Trajector, and the Location can be
100 Dieter Kastovsky

interpreted as a Landmark, to use Langacker’s terms, which by now have

become current in cognitive linguistics (cf. Langacker 1987). This spatial
relationship has undergone a metaphorical extension into more abstract
domains, where the Location represents a Status or State (S), into which the
Theme is transferred (= changed) or from which it is removed (= changed back
to its original Status or State), e.g.
(3) a. to enslave the population ‘convert the population into the status of
slaves’, to dramatize a novel ‘convert a novel into the status of a drama’
b. to demilitarize an area ‘to undo the militarized state of the area’, to
unbarbarize a person ‘to undo the barbar(ian) state of a person’, to
untie a shoe ‘to undo the tied status of the shoe’

Group (3a) is parallel to the locative type in (1), while group (3b) matches the
ablative and reversative verbs in (1), where a given state is changed into its
opposite, i.e. is undone. It should be noted, however, that it is not always clear
whether the state in question implies a pre-action by which it has been brought
about (as necessarily in to untie, unfasten, disjoin, unbutton), or not (as probably
in unbarbarize). It would seem that this is at least partly a pragmatic question,
i.e. has to do with our perception of extralinguistic reality and our general
extralinguistic knowledge. Therefore the borderline between reversative verbs
implying a pre-action and ablative or privative verbs not implying a pre-action
is somewhat fuzzy. Thus, it is not quite clear whether e.g. to disarm should be
treated as a privative or a reversative verb. Having arms normally presupposes
the pre-action of providing someone with arms, since one is not born with
them, and this would make the verb to disarm reversative. But to disarm could
also simply be interpreted as ‘remove arms’ with any presupposed pre-action
being backgrounded or obliterated, which would make it a privative verb,
although this alternative strikes me as less plausible than the first interpretation.
Inversely, to behead is clearly privative and not reversative; one hardly ever first
puts a head on and then removes it, since people and animals are normally born
with heads. But in linguistics it is of course conceivable that one first provides
a structure with a head (i.e. makes it headed) and then beheads it again, in which
case the verb has to be interpreted as reversative. It would seem that the
distinction between alienable and inalienable possession plays a role in the
decision as to whether a specific verb is interpreted as privative/ablative or as
reversative.
The difference between locative, ablative and reversative verbs on the one
hand, and ornative and privative verbs on the other is a matter of focus, of point
 Ornative, locative, ablative, privative and reversative verbs in English 101

of view (or topicalization, cf. Kastovsky 1973). With the first group, L/S is
focused on and becomes the basis of the derivation, while T is expressed as the
external (syntactic) Object of the derived verb; with the latter group, T is
focused on and functions as the basis of the derivation, while L is expressed as
external (syntactic) Object of the derived verb, cf. the following examples,
where the focused element is italicized:
(4) locative (‘cause T to come to be in L/S’ = ‘put T into L/S’): encage ‘to put
someone/something into a cage’, emplane; enslave, entangle; enfeeble,
embolden; prettify, legalize, clean; dramatize, gasify, bottle, bundle, cripple
(5) ablative (‘cause T to come to be not in L’ = ‘remove T from L’): deplane
‘to remove oneself/someone from a plane’, disbar, unearth
(6) reversative (‘cause T to come to be not in L/S’ = ‘undo the result S of a
pre-action’): desegregate ‘to cause something to be not segregated’, decen-
tralize, demilitarize, denazify; disjoin, disentangle; untie, unbutton
(7) ornative (‘cause T to come to be in L’ = ‘cause L to have T’, ‘provide L
with T’): encrown ‘to cause a crown to come to be with someone = to
cause someone to have a crown’, chlorinate, alcoholize, butter, salt
(8) privative (‘cause T to come to be not in L’ = ‘cause L not to have T’,
‘deprive L of T’): behead ‘to cause the head to come to not be with some-
one = cause someone not to have a head’, defrost, desulphurate, deoxy-
genate; disarm, unbalance, unnerve; bone, shell

Clearly the fact that this basic locative structure and its negation as well as its
metaphorical extensions underlie all these verbs is the reason for them to use
the same derivational means.
Let me now first review the situation in Modern English in somewhat
greater detail. I will then turn to a description of the morphological realization
of these semantic patterns in Old English, and in a third part point out the
major changes that have taken place between Old English and Modern English
during the Middle and Early Modern English periods, primarily as a result of
the influence of French and Latin (cf. also Kastovsky 1996b: 112–113 for some
more general remarks).

2. Modern English

As the examples have already demonstrated, all semantic types that theoretically
can be derived from the basic semantic-cognitive structure postulated in (2)
102 Dieter Kastovsky

exist and are productive. They are realized partly by prefixal formations, with or
without a suffix, and partly by suffixations alone, or by zero-derivation, with
zero-derivation being the most frequent option in many instances (or zero-
derivation combined with prefixation).2
(9) ornative:
be-: bewall, besnow, begrime, bejewel, begirdle (often only as participles)
en-: encrown, encloud, encolour (weak)
-ate: hyphenate, chlorinate
-ize: alcoholize, carbonize, ionize, oxygenize
-Ø: arm, flavour, butter, salt, pepper
(10) locative:
be-: bemonster, beslave (EModE only)
en-: encage, encapsule, emplane, endanger, enrapture, enslave, entangle;
enable, enlarge, enrich, embitter; embolden
-ify: prettify, humidify, intensify (State only)
-ize: dramatize, legalize, neutralize (State only)
-en: blacken, sweeten, whiten (State only)
-Ø: bottle, can, bag, catalogue, blacklist; bundle, cash, cripple; calm,
clean, dirty, empty
(11) ablative:
de-: dethrone (or privative?), debus, deplane, detruck, delist
dis-: dislodge, displace, disbar, dischurch (rather rare)
un-: unsaddle, unbale, uncage, unearth, unhook, unseat
(12) privative:
de-: debark, defat, delouse, defrost, deworm; decapsulate, dehydrate, de-
sulphurate; deodorize, deoxidize
dis-: disarm, discourage (unproductive)
un-: unburden, unfrock, unmask, unnerve
-Ø: bone, bark, fin, gut, scale, stone, shell, skin, weed, worm
(13) reversative:
de-: demilitarize, decentralize, depolarize, dehumanize; desegregate
dis-: disconnect, disarrange, disengage, disinfect
un-: untie, undo, unroll, unfasten; unbutton, unclasp, unlatch, unlock

As the examples illustrate, there is a mixture of native and non-native forma-

tions: the prefixes de-, dis- and en- as well as the suffixes -ate, -ify and -ize are
non-native, whereas the prefixes be- and un- as well as zero-derivation repre-
sent native formations.
 Ornative, locative, ablative, privative and reversative verbs in English 103

3. Old English

In Old English, most of these semantic categories are also present, with one
exception, however. Thus, I could not find any ablative verbs. The only derivat-
ional source for this semantic category would be formations with the prefixes
on-/un-, but they did not provide any ablative examples. This may be due, of
course, to the restricted material available in the existing texts. But it might also
represent a genuine gap in the OE word-formation patterns.3 As in Modern
English, we find prefixation, suffixation and zero-derivation or a combination
of these processes, but the number of affixes used is smaller than in Modern
English, and, with the exception of reversative verbs, the productivity of these
patterns also seems relatively restricted compared to Modern English:
(14) ornative: be-: beclæman (extension of clæman) ‘to plaster’, begyrdan
(extension of gyrdan) ‘to girdle’, betynan (extension of tynan) ‘to en-
close’, bebyrdan (extension of (ge)byrdan) ‘to fringe, border’, besmirian
(extension of smirian) ‘to smear, anoint’, bewæpnian (extension of
wæpnian) ‘to provide with weapons’; bedician ‘surround with a dike’,
behriman ‘cover with hoarfrost’, besmocian ‘to smoke, envelop with
incense’, besniwan ‘cover with snow’
on-: ongierwan (extension of gierwan) ‘to clothe’, onscrydan (extension of
scrydan) ‘to clothe’, geunmihtan ‘deprive of strength = provide with
weakness (unmiht)’
-s-(ian): geegsian ‘to terrify, inspire with fear’, metsian ‘to feed’
-Ø: byrman ‘to ferment, provide with bearm’, scrydan ‘to clothe’, arian
‘to provide with honour’, beagian ‘to provide with a ring’, cynehelmian
‘to crown’, pician ‘to provide with pitch’, piporian ‘to pepper’, beddian
‘to provide with a bed’, gryndan ‘to found (of a house), i.e. to provide
with a foundation’
(15) locative: -s-(ian): bliþsian ‘to make glad’, clænsian ‘to clean’, hlænsian ‘to
make lean, soft’, unclænsian ‘to soil = to make unclean’, untreowsian ‘to
defraud’
-Ø: cistian ‘put into a coffin’, gryndan ‘to set, sink (of the sun) = to come
to be at the ground’, husian ‘to house’; heapian ‘to make into a heap’,
clynian ‘to make into a ball’, munucian ‘to make into a monk’; byldan ‘to
make bold’, drygan ‘to make dry’, fyllan ‘to fill = to make full’, heardian
‘to make hard, bold’, blodigian ‘to make bloody’; unretan ‘to make sad’,
geunsoþian ‘to falsify, disprove’, unyþgian ‘to trouble’, unsyngian ‘to
exculpate = to make innocent (unsynnig)’
104 Dieter Kastovsky

(16) privative: be-: beheafdian ‘to behead’, bewæpnian ‘to deprive of weapons’
(or reversative, cf. ornative wæpnian above?), befotian ‘to cut off one’s
feet’, behorsian ‘to deprive of horses’, belandian ‘to deprive of land’
-Ø: heafdian ‘to behead’
(17) reversative: on-: onbindan ‘to untie’, onlucan ‘to unlock’, ongierwan ‘to
unclothe’, ongyrdan ‘to unbuckle’, onreafian ‘to strip off garments’,
onsælan ‘to unseal’, onscogan ‘to unshoe’, ontynan ‘to open’
un-: unfealdan ‘to unfold’, unwindan ‘to unwind’, unwyrcan ‘to undo,
destroy’, unspannan ‘to unfasten’, uncnyttan ‘to unbind’, geunfæstnian
‘to unfasten’, geungewlitigian ‘to disfigure’, ungierwan ‘to unclothe’,
unhadian ‘to divest of holy orders’, uninseglian ‘to unseal’, unscrydan ‘to
take off clothes’, unscogian ‘to unshoe’

With ornative verbs, the prefix-formations often are just extensions of zero-
derived denominal formations (both old class 1 and more recent class 2 weak
verbs occur, sometimes side by side as doublets). But there are also formations
such as bedician, besniwan, which seem to be direct denominal formations
without a zero-derived non-prefixed verb as an intermediary (unless the
unprefixed verbs also existed but are not recorded, which is also possible, of
course). Note that the prefix be- has preserved its productivity until Modern
English (despite the rivalling opposite meaning of ‘privativity’ as in to behead),
whereas the prefix on- has been lost (or has merged with the prefix un-; either
interpretation is possible). Incidentally, on-, like be-, also had two opposite
functions, i.e. besides ornativity it also denoted privativity. As in Modern
English, the most productive morphological pattern seems to have been zero-
derivation. This pattern is relatively old, cf. the occurrence of the numerous
class 1 weak verbs exhibiting i-umlaut (cf. scrud ‘dress’ :scrydan ‘to dress’, frofor
‘comfort’ : frefran ‘to comfort’), a morphophonemic process which was hardly
productive in the later OE period (cf. Kastovsky 1996b).
The locative type is not yet well developed, which seems to correlate with
the total absence of the ablative type, its mirror image. Zero-derivation clearly
dominates, and there do not seem to exist any prefixal formations. While the
concrete locative pattern is rather weak, the metaphorical deadjectival and
denominal ‘change of state’ pattern seems to have been reasonably productive.
A striking feature is the derivation of verbs from negative un-adjectives like
unretan ‘make sad’ < unrot ‘sad’ (cf. rot ‘glad’), a pattern which does no longer
exist in Modern English.
 Ornative, locative, ablative, privative and reversative verbs in English 105

Privative verbs also constitute a relatively weak pattern. Again we find

competition between prefixal be- and simple zero formations with occasional
doublets, cf. beheafdian : heafdian.
Thus, besides the ornative group, the reversative pattern was the most
productive. Looking at the examples, it would seem that the two prefixes on-
(which partly also occurred in ornative use) and un- were merging, probably on
account of their relatively weak phonetic distinctiveness, but also due to their
semantic similarity. Thus Marchand (1969: 205) argues that
Probably starting from second participle forms, the prefix on- had come to be
felt connected with the negative prefix un-. The idea of negativity is common
to both (…) What distinguishes unbound ‘not bound’ from unbound
‘loosened’ is only the additional idea of an action preceding the state of being
loosened, but the state itself is the same.

Moreover, many of the reversative verbs based on a denominal ornative verb,

e.g. ungeocian, unhadian, unhlidian, unscogian, might also be interpreted as
privative, i.e. with the implicit pre-action put into the background. This would
imply that the originally purely reversative meaning had developed a privative
variant (or an alternative privative interpretation). This eventually resulted in
the Modern English situation, where un- is the most productive reversative,
privative and ablative prefix (apart from other functions like the purely negative
one, cf. unkind, unbelievable).
Thus, with the exception of the ablative pattern, all the relevant semantic
groups are already well established in Old English. However, there are certain
problems as to the correlation between meaning and form, since both prefix
formations and zero-derivations can be ambiguous. Thus, the prefixes be- and
on- both occur in ornative and privative formations, i.e. in formations consti-
tuting a complementary opposition; and pure zero-derivation occurs with
ornative, locative and privative verbs. The prefix un- derives reversatives, some
of which might also be interpreted as privatives, apart from its function as a
negative prefix creating antonyms like brad ‘wide’ : unbrad ‘narrow’, or com-
plementaries like acenned ‘begotten’ : unacenned ‘unbegotten’, i.e. it has not yet
developed an ablative function.

4. Historical development
106 Dieter Kastovsky

The further development is characterized by the introduction of the Romance/

Latin prefixes de-, dis- and en-/em- in Middle and Early Modern English, which
increase the morphological possibilities considerably, although these prefixes
are predominantly restricted to combinations with Romance bases.4 The
starting point were usually analysable loans, i.e. instances where both the verbal
base and the prefixation were borrowed so that a derivational connection could
be established between the two (cf. Marchand 1969: 153, 158). Thus, loans such
as defy, declare, denote, despair, disdain could not act as models for English
formations, since -fy, -clare, -note, -pair (in the required sense), -dain did not
exist, whereas loans such as deplume, decipher, decanonize, demoralize, disallow,
disobey, disarm could, because the derivational base had also been borrowed.
The emergence of an ablative pattern with the prefix un- in Late Middle English as
in unsaddle ‘remove from the saddle’ (besides earlier privative ‘remove the saddle
from the horse’), or unearth, unhouse, unship may well be related to this develop-
ment, since it may have arisen under the influence of ablative dis-formations
borrowed from French such as dislodge (1450), displace (1551), disbar (1631).
The French prefix en-/em- (going back to Latin in-) has both a locative and
an ornative meaning and both may be concrete or metaphorical. The former is
dominant. The first instances are loans such as enamour, enchain, encircle,
endamage, enfeeble, enrich, which came into the language before the fifteenth
century. From c. 1400 onwards we find the first English formations, e.g. locative
embow ‘bend into a bow’ (1400), embliss ‘make happy’ (c. 1450), endanger ‘put
in danger’ (1477), and ornative encrown ‘provide with a crown’. The most pro-
ductive period was the sixteenth century. The locative-metaphorical deadjecti-
val type enfeeble, enable, ennoble, embitter also has a French model and becomes
productive in the fifteenth century, while the denominal pattern enslave,
enthrall, encaptive does not have a French equivalent and is an independent
English extension of restricted productivity, going back to the sixteenth century.
Besides suffixless derivatives we also find forms with additional suffixation, e.g.
embolden, enharden, enliven, enhearten coined between 1500 and 1650. Many of
these formations have become obsolete, however.
The prefix dis- is a combination of French des-/dé- and Latin dis-. The
earliest loans have either a purely negative (=‘not’) meaning as in disallow ‘not
allow’ (1377), disavow (1393), disobey (1393) or a reversative meaning as in
dishonour ‘deprive of honour’ (1300), disarm (1314), disclose (1393). Some of
the latter, however, might also be interpreted as privative, if the presupposed
pre-action is backgrounded. The pattern becomes productive at the end of the
fifteenth century, and receives a boost by the pattern disentangle, disembowel,
 Ornative, locative, ablative, privative and reversative verbs in English 107

disenthrone at the end of the sixteenth century. A number of denominal French

loans were also analysable as privative, cf. some of the above examples like
dishonour, and later loans such as dismember, disfigure, discharm. This pattern
becomes productive outside the domain of loans in the second half of the
fifteenth century, although many of the coinages never really gained general
currency. Finally, some French loans were based on an ablative relation, e.g.
dislodge (1450), displace (1551), disparish (1593), disbar (1631), and it seems
that these were the first verbs of this kind, triggering the extension of the prefix
un- to this semantic domain.
The prefix de- seems to be the youngest of the three borrowed prefixes. ME
denominal privative loans such as deplume ‘deprive of plumes = feathers’
(1420) < Anglo-French déplumer, decipher (1528) < F déchiffrer were imitated
by formations such as depasture ‘deprive of pasture’ (1586) anglicizing Latin
depascere, dethrone (1609), but it was only in the nineteenth century as a result
of the popularity of the suffixal reversative type demilitarize (1883) that this
privative and eventually also ablative type became really productive. There are
some Late Middle English/Early Modern English loans such as depopulate
(1545), depilate (1560), deglutinate (1609) going back to corresponding Latin or
Neo-Latin formations, which were imitated especially in scientific parlance, cf.
desulphurate (1757), deoxygenate (1799), deoxidate (1799), denitrate (1863). But
it was only during the late eighteenth and the nineteenth century that de-verbs
really caught on thanks to the dramatically increased productivity of this
pattern in French, cf. loans such as demoralize (1793), cf. F. démoraliser,
decatholicize (1794), deoxidize (1794), dechristianize (1834). Apparently, this
prefix + suffix pattern also started out with the privative meaning, as was the
case with the pattern deplume, but since in some instances privativity and
reversativity were not always quite distinct, as we had already noticed with some
other patterns, originally privative formations could be reinterpreted as
reversative by adding the presupposition of a pre-action; as a consequence, de-
developed an independent reversative pattern, which is now also very strong
outside the scientific-technical domain, whereas privative and ablative forma-
tions kept their more scientific-technical character.

5. Conclusion

The derivational patterns investigated here share a common cognitive-semantic

basis, which accounts for the fact that they use the same formal derivational
<DEST "kas-n*">

108 Dieter Kastovsky

means. They are already very well established in Old English, with the exception
of ablative verbs, and a relatively weak representation of the locative type. This
gap is closed by the introduction of the French/Latin prefixes en-/em-, de- and
dis-, which generally strengthened this whole derivational set and made it one
of the most productive areas in verbal word-formation.

Notes

* I would like to thank the referee for the many helpful suggestions.
1. [[AGENT]] is not part of the semantic structure involved in the derivation of the verb
itself, but part of the semantic structure into which it is embedded by virtue of the element
CAUSE, which implies a CAUSER (= AGENT, INSTRUMENT).
2. For a more detailed description and some problems related to the morphological analysis
of prefixal verbs, cf. Marchand (1969, 1971, 1973) and Kastovsky (1986).
3. The material of this survey is based on an exhaustive search in Bosworth & Toller (1898/
1921) and Clark Hall (1962). It is of course possible that with the help of the Microfiche
Concordance to Old English (Toronto, 1985) some ablative examples could be discovered by
a detailed semantic study of the examples. Moreover, the lists given are not fully exhaustive,
but only include the most typical examples.
4. It should be noted that, although the actual source of these prefixes were French loans,
they were strengthened also by loans from Latin, especially in the EModE period, cf. also
Nevalainen (1999, esp. 358ff.).

References

Bosworth, Joseph & T. Northcote Toller. 1898. An Anglo-Saxon Dictionary. 1921. Supplement
by T. Northcote Toller. London: Oxford University Press.
Clark Hall, John R. 1962. A Concise Anglo-Saxon Dictionary. 4th ed. with a supplement by
Herbert D. Meritt. Cambridge: Cambridge University Press.
Kastovsky, Dieter. 1973. “Causatives”. Foundations of Language 10.255–315.
Kastovsky, Dieter. 1986. “Problems in the Morphological Analysis of Complex Lexical
Items”. Acta Linguistica Academiae Scientiarum Hungaricae 36.93–107.
Kastovsky, Dieter. 1996a. “The Representation of Space in English Derivational Morpholo-
gy”. The Construal of Space in Language and Thought ed. by René Dirven & Martin Pütz,
197–208. Berlin & New York: Mouton de Gruyter.
Kastovsky, Dieter. 1996b. “Verbal Derivation in English: A Historical Survey. Or: Much Ado
about Nothing”. English Historical Linguistics 1994 ed. by Derek Britton, 93–117.
Amsterdam & Philadelphia: John Benjamins.
</TARGET "kas">

Ornative, locative, ablative, privative and reversative verbs in English 109

Langacker, Ronald W. 1987. Foundations of Cognitive Grammar. Vol. I: Theoretical Prerequi-

sites. Stanford, Calif.: Stanford University Press.
Marchand, Hans. 1969 [2nd rev. ed.]. The Categories and Types of Present-day English Word-
Formation. München: Beck.
Marchand, Hans. 1971. “Die deadjektivischen reversativen Verben im Deutschen, Englischen
und Französischen: entmilitarisieren, demilitarize, démilitariser”. Interlinguistica.
Sprachvergleich und Übersetzung. Festschrift zum 60. Geburtstag von Mario Wandruszka
ed. by Karl-Richard Bausch & Hans-Martin Gauger, 208–214. Tübingen: Niemeyer.
Repr. in Marchand 1974, 397–405.
Marchand, Hans. 1973. “Reversative, Ablative and Privative Verbs in English, French, and
German”. Issues in Linguistics. Papers in Honor of Henry and Renée Kahane ed. by Braj
B. Kachru, Robert B. Lees, Yakov Malkiel, Angelina Pietrangeli & Sol Saporta, 636–643.
Urbana: University of Illinois Press. Repr. in Marchand 1974, 405–415.
Marchand, Hans. 1974. Studies in Syntax and Word-formation. Selected Articles by Hans
Marchand. On the Occasion of his 65th Birthday on October 1st, 1972 ed. by Dieter
Kastovsky. Munich: Fink.
Nevalainen, Terttu. 1999. “Early Modern English Lexis and Semantics”. The Cambridge
History of the English Language. Vol. III: 1476–1776 ed. by Roger Lass, 332–458.
Cambridge: Cambridge University Press.
<LINK "kor-n*">

<TARGET "kor" DOCINFO

AUTHOR "Lucia Kornexl"

TITLE "From gold-gifa to chimney sweep?"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

From gold-gifa to chimney sweep?

Morphological (un)markedness of Modern
English agent nouns in a diachronic perspective*

Lucia Kornexl
University of Greifswald

1. An outline of the problem

In Modern English, Agent is primarily a derivational category, the majority of

formations being marked by the ubiquitous -er suffix.1 However, this category is
also said to include a number of formally unmarked nouns derived from a verbal
base such as cheat, flirt, sneak, spy, coach, cook, guide, chimney sweep and shoeblack.
The corresponding Old English type is allegedly represented by formations such
as andetta ‘confessor’, bora ‘bearer, supporter’, gefēra ‘companion’, spreca
‘speaker’, winna ‘opponent’, gold-gifa ‘gold-giver, lord’ and mere-fara ‘seafarer’.
Following Marchand’s line of reasoning (1969: 76f., 359f., 376),2 Kastovsky
(1985: 225, 246f.; 1992: 392f.) classifies these nouns as ‘zero derivatives’, arguing
that the assumed agentive zero-morpheme parallels overt agentive marking in
derivatives such as OE bæcere > ModE baker and OE folgere > ModE follower.
On the basis of this ‘overt analogue criterion’3 more or less clear claims have
been made as regards the diachronic continuity of a suffixless agentive pattern
from Old English times onwards (cf. below, Sections 2 and 5). Yet a closer look
at the Modern English examples that serve to illustrate this pattern reveals that
only one of the items — namely cook — can actually be traced back to Old
English. This word, though, entered the language as a loan (cf. OED s.v. cook n.)
and can thus neither be used to prove the existence nor the persistence of a
word-formation pattern.4 The apparent lack of old formations need, however,
not per se be notable if we take into account that during its long history the
English lexicon has been subject to quite extensive loss and restructuring and
that the available diachronic records exhibit all sorts of limitations. Besides,
112 Lucia Kornexl

what writers usually claim is the continuity of a word-formational type, not the
survival of specific formations. Still, the apparent difficulty in producing
examples of zero-derived agentive nouns which were already current in Old
English and continue to be in use beyond the Early Middle English period may
well be indicative of problems that go beyond the retrieval of a sufficient
number of long-lived attestations.
It is the aim of this paper to try and test scholarly assumptions about the
diachronic continuity of a morphologically unmarked agentive pattern in
English on the basis of the lexical evidence presented in the literature. Taking
up the chief arguments on which the notion of ‘zero-derivation’ rests, we shall
have to examine the relevant historical and modern data both with regard to
their structural and semantic properties: is there a formal continuity between
the suffixless agentive pattern claimed for Old English and the unsuffixed
agentive type that has been postulated for Modern English, and do the respec-
tive forms also meet the content criterion by being equivalent in meaning to the
overtly marked instantiations of the Agent category? In pursuing these
questions special attention will be devoted to the situation in the Medieval
English period, as the Early Modern and Modern English material cited in the
handbooks has already been analysed in some detail for the purposes of a
previous article that took a cognitive approach to the matter (Kornexl 1998).
Trying to adduce additional proof for the line of argumentation developed
there, the present paper issues a challenge to claims that English has ever
possessed a morphologically unmarked agentive pattern with a sufficient degree
of productivity and semantic homogeneity to justify the establishment of a
word-formation type parallel to and competing with suffixed agentives.

2. A summary of previous research

A large part of the relevant Old English material has been collected and analysed
in detail by Dieter Kastovsky in his dissertation on Old English Deverbal
Substantives Derived by Means of a Zero Morpheme (1968). In a further investi-
gation (1985) Kastovsky examined the whole spectrum of deverbal nominal-
izations in Old and Modern English with regard to the structural properties and
the semantic functions of each derivational pattern. He arrives at the conclusion
that “[d]erivation without an explicit suffix is already extremely common in
Old English and is thus by no means a characteristic feature only of Modern
English, as some handbooks claim” (1985: 246). Semantically, it is action and
 From gold-gifa to chimney sweep? 113

agentive nouns that dominate among the Old English affixless derivations
under scrutiny (1985: 253), with zero showing “a somewhat greater affinity to
Action than to Agentive nouns” (1985: 255).5 This tendency seems to have
subsequently increased, for in Kastovsky’s table charting the overall system of
deverbal nouns in Modern English (1985: 226), agentive is no longer marked as
a dominant semantic type within zero-derivatives.
Most of the pertinent agent nouns in Old English belong to the weak
n-declension and carry the masculine ending -a, e.g. dēma ‘judge’, hunta
‘hunter’ and wiga ‘fighter’. Within this declensional class there are also a few
feminines in -e such as hore ‘whore’, widuwe ‘widow’ and cild-fōstre ‘nurse’,
which due to their scarcity, restricted analysability and apparent lack of produc-
tivity are only of minor relevance in our context.6 They will not, therefore, be
separately discussed in this article.
The exact morphological status of the final exponent in Old English and its
functional range are a matter of dispute. Nevertheless, there has scarcely been
any scholarly debate on this question. In older publications, the indiscriminate
terminology as regards the inflexional or derivational nature of the -a suffix, in
combination with pronounced statements about its agentive meaning, tends to
suggest word-formational properties.7 This interpretation still has its proponents.
Dennis Baron, for example, informs the readers of his study on Grammar and
Gender that “Old English employed the agentive suffix -a to indicate a person
performing a particular act, function, or profession” (1986: 116; my italics).
This view on the typological structure of the Old English n-stems has
repeatedly been refuted by Dieter Kastovsky. In his opinion a classification that
regards -a as an agentive suffix not only blends synchronic and diachronic
paradigms but in interpreting grammatical endings as derivational suffixes also
fails to observe the distinction between inflexion and word-formation. Kastov-
sky’s line of argumentation (1968: 13, 81–83; 1985: 246f.; 1992: 392f.; 1997: 72)
may be summarized as follows:
The patterns underlying the Old English n-stems originated in the addition
of nominal stem formatives (mostly Germ. *-an-, *-ōn- and — with a j-exten-
sion causing i-mutation in the root syllable — *-jan-, *-jōn-) to roots, in the
case of agent nouns predominantly verbal ones. Due to their merger with the
inflexional endings already in Germanic, these derivational affixes can no longer
be isolated in the Old English attestations of this type.8 The fact that the -a of
the Nominative Singular is not represented throughout the whole declensional
paradigm9 clearly marks it as an inflexion.10 Thus lacking an overt derivational
morpheme, the pertinent Old English formations have to be classed as affixless.
114 Lucia Kornexl

Kastovsky therefore postulates the following analysis: wig/Ø-a ‘fighter’, man-

slag/Ø-a ‘man-slayer’.

3. The ‘meaning’ of the Nominative Singular marker of the Old English

personal n-stems

From a form-oriented and strictly synchronic point of view Kastovsky’s analysis

of the Old English agentive n-stems seems fully justifiable. As we are dealing
with a pattern that as a consequence of phonological processes in pre-Old
English times underwent a significant change in morphological structure, it
may, however, be worth asking to what extent this modified pattern in its Old
English shape still fits discrete linguistic categories or — viewed from the
opposite angle — how adequate these categories are to capture the change in
morphological make-up. There are in fact indications that on their way to Old
English the personal n-stems did not fully cross the ‘boundary’ between
derivation and inflexion, thus in a way reflecting the categorial indeterminacy
inherent in the term ‘morphology’ itself.11
With regard to pre-Old English conditions it is commonly assumed that the
various stem-formatives of n-declensional nouns in Germanic apart from their
grammatical role also served a semantic function. According to Wolfgang
Meid’s Wortbildungslehre these n-suffixes carried primarily an individualizing
force, deriving personal designations, above all agentives, as well as names for
other living beings and anthropomorphized objects (Krahe-Meid 1969: 91, §91;
cf. also Kluge 1926: 7, §12). In Old English, the deverbal n-stems cover a variety
of semantic categories12 and — as has been outlined in the previous section —
they no longer contain a formal element that can be assigned a clear derivative-
semantic function. As the agentive nouns within this declensional class still
reflect what was apparently the original meaning of the derivational pattern in
Germanic, it seems, however, likely that their Nominative Singular marker,
which “probably started life as a deverbal suffix”, after having been reanalysed
as an inflexion retained “something of the semantics of its original” (Lass
1994: 134 fn.16). This seems the more plausible as in Old English the n-stem
agents formed part of a semantic category that was otherwise characterized by
explicit derivational marking. Thus their inflexional -a or -e may at least to
some extent have been regarded as a functional equivalent to word-formation
suffixes such as OE -end and -ere13 not only for semantic reasons. Exhibiting a
basic analysability and being attested in highly frequent agentive formations
 From gold-gifa to chimney sweep? 115

(see next section), the n-stem nouns under consideration presumably exhibited
a sufficient amount of relevant categorial traits and were notionally salient
enough to be categorized with suffixal agents by speakers of Old English. With
their final marker being a prime candidate for phonetic attrition and loss, these
typological crossbreeds could, however, not really stand a chance of successfully
competing with ‘real’, i.e. formally marked derivatives.

4. The question of productivity in Old English

The claim that on their way from Germanic to Old English the agentive n-stems
(together with a number of other word-formation patterns) changed their
morphological status from suffixation to zero-derivation (Kastovsky 1992:382f.,
392f.) appears to be problematic for yet another reason. Whether this process
is seen as a clear-cut typological shift or — as has been suggested in the previous
section — as a structural change of a more indeterminate nature, it no doubt
led to a redefinition of the place of already existing formations within the
overall system of Old English derivatives. Yet Kastovsky’s claim extends to the
derivational pattern per se and thus carries further implications: it suggests that
we are dealing with a productive word-formation pattern. The peculiarities of
the Old English material are, however, of such a kind that this notion definitely
needs qualification.
At this point, a practical problem concerning the acquisition of relevant
data may briefly be addressed. Technically speaking, the formations discussed
here represent a particularly evasive type of vocabulary. Words ending in highly
frequent inflexions or carrying no marker at all are rather unsuitable candidates
for automated searches in electronic dictionaries and corpora. This applies not
only to the earlier stages of the English language but is of course equally true for
Present-day English material. Thus exact frequency counts for unmarked forms
tend to be conspicuously absent from word-formation studies and other
relevant literature, even where the existence of affixless derivations is explicitly
acknowledged.14 The general line of development may, however, be quite
confidently reconstructed on the basis of existing collections.
That the Old English personal n-stems represent an established pattern of
long standing is evidenced by the fact that there are quite a number of obvious
linguistic relics among them. Nouns such as guma ‘man, lord, hero’ from a
synchronic point of view completely lack morphosemantic analysability.
Derivatives like scytta ‘shooter, archer’ and hlytta ‘soothsayer’ betray their
116 Lucia Kornexl

advanced age by the presence of i-mutation in the root. It is also commonly

assumed that derivation from strong verbs based on non-infinitival stems
marks a formation as pre-Old English.15 Yet the productive potential of a
pattern is usually not inhibited by the presence of some fossilized forms among
existing formations that on the whole exhibit a sufficient degree of analysability
and transparency.16
Though definite proof is impossible, it seems nevertheless reasonable to
assume a certain amount of productivity for the agentive n-stems in Old
English. This is above all suggested by a specific pattern, i.e. combinations of a
substantive plus a zero-derived deverbal substantive such as man-slaga ‘man-
slayer, murderer’ and bēag-gifa ‘ring-giver, lord, king’. Basically, these so-called
‘synthetic compounds’ or ‘verbal nexus combinations’ (cf. Marchand
1969: 15ff., 31ff.; Kastovsky 1992: 364f.) represent a very old compositional type
that was highly productive in North and West Germanic (Kluge 1926: 9, §15;
Krahe-Meid 1969: 26, §31.1a). Judged by the large number of attestations that
Kastovsky lists in his dissertation (1968), the type must have been quite prolific
in Old English, too. Their frequent presence in poetic texts, where lexical
variation was a major stylistic device, makes at least part of these formations
likely candidates for linguistic innovation in Anglo-Saxon times.
The mere size of the Old English material seems to contradict the notion of
a residual paradigm.17 At first sight, these compounds suggest themselves as
first-rate evidence for the productivity of a zero-derived agentive type. It is,
however, a peculiar kind of productivity which these compounds exhibit. As the
classified word lists in Kastovsky’s dissertation (1968) show, there are a number
of personal n-stems such as gifa ‘giver’ and boda ‘messenger’ which frequently
serve as the determinatum in combinations with varying first elements. In many
cases the determinatum does not even occur as an independent lexeme outside
the compound: cf. e.g. -breca as in ǣw-breca ‘adulterer’ or lah-breca ‘lawbreak-
er.’18 Because these compositional elements are possible verbal derivatives with
full lexical meanings, no convincing case can be made for treating them as
suffixes.19 But since compounds of the gold-gifa or ǣw-breca type use prefabri-
cated formations for their determinatum, they can neither be regarded as first-
rate evidence for the productivity of agentive zero-derivation in Old English.
What is likely to be new in an Old English coinage of this kind is the particular
combination of two compositional elements both of which belonged to the
common word stock and were presumably stored as lexical units in the mental
lexicon of their Anglo-Saxon creators.
 From gold-gifa to chimney sweep? 117

The concentration of the Old English evidence on compounded formations

in -boda, -bora, -gifa, etc. was probably not only of numerical significance.
Frequency of attestation and use must have given such agentives a notional
saliency that also encompassed the -a exponent as a morphosemantic marker.
Yet viewed against the -end and -ere suffixations, the formal inadequacy of the
Old English n-stems in their phonetically reduced and structurally diluted shape
must have become even more obvious.20 The ‘suffixal extension’ of many
agentive n-stem formations in Old and Middle English21 and — in many cases
— their eventual replacement by derivations in -er(e) (Marchand 1969: 76, 79;
Sauer 1992: 201) may be described as a systematic process of structural repair or
‘re-iconization’.22

5. The post-Conquest development and status of affixless agentives:

A critical reassessment

5.1 Compounded formations

Thanks to Hans Sauer’s extensive study on nominal compounds in Early
Middle English (1992) we get a fairly clear picture of the development of
affixless agents used as compositional elements subsequent to the Old English
period. Within his corpus, unsuffixed formations of this type are still relatively
well-represented, but a substantial proportion of the evidence comes from Old
English and is no longer semantically transparent. By contrast, his agent
compounds in -ere, though smaller in number, originate almost exclusively in
Middle English and thus exhibit a much higher productive potential (Sauer
1992: 198ff., 212ff.).
Sauer’s material proves instructive in yet another respect. His lists of Early
Middle English compounds contain several instances of an apparent homonym-
ic clash or of scribal confusion between cognate agent and action nouns such as
eu-bruche ‘fornicator’, ‘fornication’ (OE ǣw-breca, ǣw-bryce) and mon-slage
‘murderer’, man-slege ‘murder’ (OE man-slaga, man-slege) (Sauer 1992:203–206,
209). Polyfunctionality though, as Kastovsky (1985: 227, 253) reminds us, has
been a general characteristic of deverbal noun-forming suffixes including zero
throughout the history of English and can thus scarcely by itself account for the
weakening of the agentive type diagnosed above. Apparently the restrictions on
the productivity of n-stem agentives were not only a matter of insufficiently
explicit marking.23 There seems to have been a trend towards regular derivation
118 Lucia Kornexl

by word-formational elements for the category Agent that was probably

reinforced by a tendency to maintain a formal difference between agent and
action nouns. This question no doubt merits further investigation.
Assertions that — despite a significant reduction in productivity — the old
n-stem type of zero-derived agents lived on through Middle to Modern English
usually make use of compounded examples.24 Marchand’s line of argumenta-
tion concerning the biography of this pattern is rather inconsistent though. He
states that “[w]ith the exception of grasshop (…) the OE combinations appear
to have died out by the ME period” and that “most PE combinations have been
coined during the MoE period” (1969: 76), but subsequently seems to suggest
a continued productivity for the Old English agentive n-stems:
Morphologically speaking, watchmaker is a suffixal extension of an older type
that may be typified by PE chimney sweep. The second component is a zero
derived deverbal agent substantive which in Old English had the ending -e, as
in gærs-hoppe ‘grasshopper’ or -a as in gold-gifa ‘gold-giver’, man-slaga ‘man-
killer’. (Marchand 1969: 79)

However, chimney sweep — the standard example in the field — is first attested
in 1611 in the wake of an older chimney-sweeper (dating from around 1500). If
we are to believe the OED, chimney sweep owes its existence not to a word-
formation process but to a semantic one — i.e. a metonymic shift by which the
cry of “chimney sweep!” was turned into an occupational designation.25 Both
on account of its origin and its comparatively young age chimney sweep appears
to be a rather poor candidate for ‘typifying’ an affixless pattern of allegedly Old
English descent for which only a few examples — none of them dating earlier
than the second half of the seventeenth century — can be adduced. According
to Marchand (1969:78) “barkeep, bellhop, bootblack, cardsharp, carhop, lifeguard,
shoeblack, and soda jerk are the only more or less common words to denote
individuals” (cf. also Marchand 1965: 64).
An analysis of the information provided by the OED reveals that with the
exception of carhop and bootblack, which were probably coined on the model of
bellhop and shoeblack,26 all these derivationally unmarked compounds have an
-er suffixation beside them. This is scarcely remarkable as occupational terms
are natural candidates for agentive -er formations. However, the fact that in
each case the -er form is the one which is attested earlier seems significant in
two ways: viewed against the already existent suffixed combinations that served
to denote the same referent, the deverbal nature of the corresponding affixless
designations and therefore their formal status as instances of zero-derivation
 From gold-gifa to chimney sweep? 119

(or, viewed from a different theoretical angle, of conversion) is open to question.

In the light of the interpretation offered in Section 5.3 below, another observa-
tion is probably still more important, namely that Marchand’s collection
mainly consists of terms for unqualified jobs that have low prestige (cf. Sauer
1992: 201).27 A closer look at the uncompounded attestations of the alleged
zero-derived type strongly suggests that the key to an adequate theoretical
description of these formations lies on the semantic side.

5.2 Simplex forms

Marchand does not postulate the existence of a continuous productive pattern
for the derivationally unmarked simplexes of his so-called “cheat group”
(1963b: 178f.), because, as he says in a different context, “if we check their
history, we see that this type is not older than the 16th century” (1969: 383).
This “Subject group”
is illustrated by bore, flirt, flunk AE, gossip, grind, pry, scold, shirk, sneak, snoop,
soak, sponge, spy, tease, tramp (all derogatory), coach, cook, guide, judge (the few
non-derogatory words of the group of substantives denoting persons).
(Marchand 1969: 376)

Diachronic continuity is, however, assumed for such uncompounded forms by

Kastovsky (1985: 225). In his illustrative section on Modern English zero-
derivatives including agentives (cheat, flirt, sneak, spy, coach, cook, guide) he
declares the type to be “of Old English origin”, but does not account for its
conspicuous lack of vitality throughout the Middle English period and its
alleged resurgence in Early Modern English. Matters are further complicated by
the fact that genetically the items classified as zero-derived deverbal nouns do
by no means form a homogeneous word-formational ‘type’ or ‘group’.28 Strictly
speaking, loans such as cook, guide and judge, and instances of metaphorization
or metonymic shift such as sponge and bore do not belong to the field of word-
formation at all. Marchand though would have sharply refuted this argument,
for in determining the derivational status of lexical items and the derivative
relations between them “[s]ynchronic linguistics will not regard historical
evidence as a solution. History is one thing, grammar is another” (1963b:179).29
It seems, however, clearly counterproductive if diachronic information is
deliberately overlooked where, as in our case, it is apt to reveal the heterogeneity
of an assumed word-formation type.
120 Lucia Kornexl

5.3 A recategorization of the ‘cheat group’

Even if we follow Marchand’s demand for a strictly synchronic approach, we are
bound to run into difficulties, for the material he presents in one respect
definitely beats his own theoretical assumptions. As has already been pointed
out, the concept of a zero-morpheme presupposes a semantic parallelism with
an explicitly derivative pattern (cf. above, Section 1). This is usually claimed
with reference to agentives such as cook vs. baker and spy vs. observer. A closer
look at the semantics of the ‘cheat group’ items reveals, however, that most of
them cannot be categorized as prototypical agents equal in meaning to agentive
derivatives in -er. Marchand’s comment following the bulk of his examples
(1969: 376) — “(all derogatory)” — is in fact far more important than the
bracketing suggests. Significantly, the four prototypical agentives among his
sample — coach, cook, guide and judge — are conversely qualified as “(the few
non-derogatory words of the group of substantives denoting persons)”. What
according to Marchand’s classification scheme appears to be an additional
semantic feature of most of his agentive zero-derivations seems more likely to
constitute a categorial boundary. On closer inspection the members of
Marchand’s ‘cheat group’, which he presented as a single word-formational
type, may be allotted to two types that differ significantly in meaning:
1. A handful of agent nouns such as cook, guide, judge and spy — all of them
acquired by way of borrowing from Latin or French. As purely semantic agents
without any overt marking they indeed have an exceptional status within this
morphologically marked category. But to treat them as instances of agentive
zero-derivation or — adopting a different theoretical stance — of conversion
assumes the existence of a regular and productive word-formational pattern
with a sufficient number of attestations for which no convincing evidence can
be adduced.30
2. The second type, which consists of Marchand’s “derogatory” items, usually
also contains an agential component, but in terms of notional saliency other
features clearly prevail — features of a primarily evaluative nature that attest to
a distinct speaker-relatedness and a pronounced pragmatic function. This may
be demonstrated by a simple test. If, for example, we look up cheat in the LDCE
(1995), we find the following definition: “someone who is dishonest and
cheats”. Sneak is explained as “a child who is disliked because they tell adults
about bad things that other children have done wrong”, and bore as “someone
who is boring, especially because they talk too much about themselves”.
Likewise, the Comprehensive Grammar of the English Language in its section on
 From gold-gifa to chimney sweep? 121

“Conversion to noun” paraphrases the “Subject of V[erb]”-type bore by

“someone who or that which bores/is boring” (Quirk et al. 1985: 1560). The fact
that in all these cases in addition to or instead of the typical agent definition ‘a
person who V-s’ a characterizing description is provided indicates that these
lexical items have a peculiar semantic profile and — as their frequent labelling
as ‘slang’ or ‘colloquial’ suggests — exhibit also a specific stylistic and functional
range. Focussing primarily on semantic or pragmatic aspects, such nouns may
be variously classed as ‘attributive’, denoting a “person/object that is A/has the
property N” (Szymanek 1988: 179) or as ‘attitudinal’, reflecting “the speakers’
attitudes towards certain people or things” (Schmid 1999: 219).
As Szymanek uses the term ‘attributive’ to describe an exclusively derivatio-
nal category (including zero-derivation), it needs to be stressed that the
adoption of this term in our context carries no specific claim with regard to the
origin and structural make-up of the lexemes in question. The formal charac-
teristics of this expressive category definitely demand further study. According
to Dalton-Puffer (1996: 144f., 146, 157–159) there is only scant evidence for
expressive word-formation in Middle English, and the same applies to Modern
English. Let me tentatively suggest that this may at least to some extent be due
to the fact that — in marked contrast to agents — attributives in English
represent a predominantly non-derivational category.
By virtue of their mixed origin the members of Marchand’s ‘cheat group’
can hardly be said to constitute a word-formational type. This is, however, not
to say that structure is of no relevance for these items. On the contrary, their
simplex, monosyllabic shape seems as essential as a categorial marker as are the
expressive, predominantly pejorative meaning, a colloquial, often slangy tinge,
and occasional features of phonetic symbolism (compare e.g. flirt, flunk, sneak
and snoop).
That this expressive type draws on a variety of productive sources is
confirmed by an analysis of John Ayto’s recently published Twentieth-Century
Words (1999). Looking specifically for unmarked formations based on verbs —
i.e. those items which according to the traditional view most likely exhibit an
agentive meaning — I could find only five whose deverbal origin is, however,
not always fully clear: addict (1909) ‘someone who is addicted to a drug’, swank
(1913) ‘someone who swanks [i.e. behaves pretentiously, puts on airs]’ (British
slang), twit (1934) ‘a fool; a stupid or ineffectual person’ (slang, originally
British) stemming perhaps from to twit (< OE ætwı̄tan) ‘to reproach’, git (1946)
‘a worthless person’ (British slang), a southern variant of northern and Scottish
122 Lucia Kornexl

get ‘illegitimate child, brat’ which is ultimately derived from to get in its archaic
sense ‘to beget’, and freak (1967) ‘someone who freaks out, a drug addict’
(slang, originally US). All these terms are attributive or attitudinal in character
rather than agentive, and what seems decisive in our context: there is not a
single one among the whole range of Ayto’s derivationally unmarked neolo-
gisms denoting persons that due to a complete lack of evaluative components
clearly falls into the agent category.

6. Conclusion

Contrary to assumptions that — like other semantic categories — agentive

zero-derivation can boast a continuous history, the material analysed here
strongly suggests that English has never developed a productive type of zero-
derived agentives. The prime candidates in the field, the Old English personal
n-stem nouns with an agentive meaning exhibit too many traits of structural
ambiguity and functional restriction to have been able to compete successfully
with derivationally marked agent nouns. Even in their much more frequently
attested compounded forms they do not seem to have survived beyond the
Early Middle English period, whereas the Modern English examples for agentive
derivatives — Marchand’s ‘cheat group’ — according to his own statement
belong to a type that cannot be traced back further than the sixteenth century.
Besides, the ‘cheat group’ items are semantically and genetically inhomoge-
neous, i.e. only a few of them (being all loans) can pass as prototypical agents
and the group does not constitute a uniform word-formation pattern produc-
tive as such.
The richness in pragmatic components exhibited by those lexemes which
Marchand classed as a “derogatory” subgroup of his zero-derived agents seems
to merit the establishment of a special category of formally unmarked personal
‘attributives’ (from a semantic point of view) or ‘attitudinal nouns’ (in prag-
matic terms) in contradistinction to the prototypically derivational agentive
formations. To try and follow Marchand’s line by regarding these attributives
as a subtype of unsuffixed agents seems no convincing solution, because the
postulated main pattern itself — i.e. prototypical agents produced by way of
zero-derivation — is not sufficiently attested, and though the two types share a
number of semantic traits, native speaker intuition tends to make a clear
distinction between agentive -er derivatives and ‘attitudinal’ nouns. The latter
category deserves further study as to its development and composition. Among
<DEST "kor-n*">

From gold-gifa to chimney sweep? 123

other things such an investigation will have to account for the apparently
antidiagrammatic structure of the items in question in contrast to derived agent
nouns, which lend themselves to a straightforward analysis in terms of diagram-
matic iconicity.31
If for the reasons given above an immediate historical connection of the
attributive type with Old English affixless agentive formations is denied, we
must necessarily look for other possible affiliations. Venturing a hypothesis, let
me surmise that rather than with the deverbal n-stems as more distant cousins
this affiliation lies with the Old English representatives of a word class that is by
nature characterizing, i.e. adjectives. In their substantivized form they also
belonged to the n-stem pattern, and with words such as OE witega ‘wise man,
prophet’ (OHG wizag ‘knowing, foreseeing’) and wædla ‘poor man, beggar’
(OHG wadal ‘poor’) such nominalized adjectives counted among their num-
bers some lexicalized formations which as quality concepts seem to fit well into
our category.

Notes

* I would like to thank Claudia Claridge, Ursula Lenker, Dirk Schultze, the anonymous
referee and the editors of this volume for their valuable criticism and their helpful comments
on an earlier version of this paper.
1. For the almost unlimited productivity of ModE -er, which serves as a kind of default
option for forming agentives, cf. Bauer (1983: 285–291).
2. “By derivation by a zero-morpheme I understand the use of a word as a determinant in a
syntagma whose determinatum is not expressed in phonic form but understood to be present
in content, thanks to an association with other syntagmas where the element of content has
its counterpart on the plane of phonic expression” (Marchand 1969: 359).
3. This term is used by Sanders (1988), who provides a critical analysis of the assumptions
underlying the theory of zero-derivation. For an overview of this approach and the major
alternative concept of ‘conversion’ see Don, Trommelen & Zonneveld (2000); for a survey
of earlier research cf. Pennanen (1971). As Valera (2000: 145) points out in his review article
on Štekauer’s A Theory of Conversion in English (1996), this process “remains a challenge to
descriptive linguistics”. Major contributions in the field during the second half of the
twentieth century are also documented in Štekauer (2000; cf. Subject Index [Link]. ‘Conver-
sion’, ‘Zero-derivation’, ‘Zero-morpheme’, ‘Zero-suffixation’).
4. For a brief critical discussion of Marchand’s decision to classify loanwords such as OE cōc
(< late Lat. cocus) from a strictly synchronic, i.e. Modern English point of view, as instances
of zero derivation, see below, Section 5.2.
124 Lucia Kornexl

5. Zero-derived action nouns that have survived into Modern English may be exemplified by
OE drinc ‘drink’, stenc ‘stench’, hopa ‘hope’, lufu ‘love’, feoht(e) ‘fight’ and weorc ‘work’ (cf.
Kastovsky 1985: 247f.). For Old and Modern English unsuffixed agentive nouns as cited by
Kastovsky (1985: 225, 246f., 248f.) cf. above, Section 1.
6. For the Old English declensional paradigms cf. Campbell (1959:248–251, §§615–619) and
Sievers-Brunner (1965: 221–225, §§276–279). The Germanic n-stems are dealt with in some
detail in Kluge (1926: 7–10, §§12–17), Krahe-Meid (1969: 90–100, §§91–92) and Bammes-
berger (1990: 163–187). For a short discussion of the feminine pattern within the spectrum
of female formatives in Old English see von Lindheim (1958: 490–494).
7. See for example Knutson (1905: 15), who starts the relevant chapter entitled “German
endings (suffixes)” as follows: “In OE. there were a great many words ending in -a which
denoted a person or agent (nomina agentis). Such are e.g. cuma (a person who comes),
wilcuma (a welcome person) (…)”.
8. Bammesberger (1990: 165) offers the following tentative reconstruction of the n-stem
paradigm for ‘animate’ nouns (“die ‘belebten’ Substantiva”) in Proto-Germanic (cf. OE hana
‘cock’ = ModGerm Hahn): Sg. Nom. *hanō, Gen. *han-(e)n-az, Dat. *han-en-i, Acc. *han-
an-un; Pl. Nom. *han-an-ez, Gen. *han-(a)n-ōn, Dat. *han-(u)n-miz, Acc. *han-an-unz.
9. Cf. hunt-a (Nom. Sg.), hunt-an (Gen./Dat./Acc. Sg., Nom./Acc. Pl.), hunt-ena (Gen. Pl.),
hunt-um (Dat. Pl.).
10. As Kastovsky (1997: 72) points out, the reductive processes operating on the Old English
n-stems were comparatively less radical than the ones that affected other declensional classes.
He argues, however, that due to a reinterpretation of the old stem-formatives that had been
preserved throughout much of the paradigm as case/number endings the result is very much
the same.
11. Hermann Paul (1920: 349, §242) doubts the validity of this distinction for the pre-Old
English period: “Auf die gleiche Weise wie die Ableitungssuffixe entstehen Flexionssuffixe.
Zwischen beiden gibt es ja überhaupt keine scharfe Grenze”. Questions concerning basic
discreteness and functional autonomy are also raised in modern linguistic theory. Thus
Wolfgang Dressler (1990: 86) emphasizes that Natural Morphology (as developed by Dressler
et al. 1987; see also Dressler 2000) “assumes neither a discrete universal boundary between
compounding and derivation nor between derivation and inflection (…), but rather proto-
typical properties of prototypical inflection, derivation, and compounding respectively”.
12. A few examples taken from Kastovsky (1985: 247ff.) may suffice here: plega ‘play, fight’
(Action), hweorfa ‘spindle’ (Objective), bita ‘bit, morsel’ (Factitive), āga ‘owner’
(Benefactive), sceafa ‘plane’ (Instrumental), stı̄ga ‘path’ (Locative).
13. These two highly productive agent suffixes have been studied in detail by Kärre (1915)
and Kastovsky (1971). For Old English personal noun suffixes in general cf. the section on
“Nominal suffixes” in Kastovsky (1992: 384–389).
14. See for example the new Longman Grammar of Spoken and Written English, which in its
chapter on noun formations provides a list of zero-derived substantives but does not include
this type in the frequency counts and productivity ratings for common noun derivations
(Biber et al. 1999: 319, 323).
 From gold-gifa to chimney sweep? 125

15. But cf. the questioning remarks by Kastovsky (1985: 260 fn.17).
16. For some fundamental considerations about the synthetic (i.e. productive) and analytic
(i.e. descriptive) aspects of word-formation cf. Kastovsky (1985: 229). An overview of the
various aspects of morphological productivity is provided by Koefoed & van Marle (2000).
17. For an explication of this concept cf. Pilch (1985: 423).
18. Weinreich (1971), who observes a similar impairment of productive potential for the
corresponding Old High German pattern, takes this as a sure sign of imminent death: “Ein
Suffix, das zunehmend nur noch in der Komposition fruchtbar ist (…), muss dem Unter-
gang geweiht sein” (1971: 103f.). He rather inappropriately calls the final exponent -(e)o of
the second constituent — which just like in Old English quite frequently represents a
formation that is only attested in compounds — a “compositional suffix” (‘Kompositions-
suffix’) (1971: 102).
19. Cf. Kastovsky’s refutation of such arguments (1992: 365). For a discussion of the status
of these ‘potential words’ and for Old and Early Middle English attestations see also Sauer
(1992: 18, 200, 211f. and fn.134).
20. The subsequent fate of the -end suffix proves though that phonological substance alone
is no guarantee of survival.
21. The process may briefly be exemplified by the frequently attested OE dēma ‘judge, ruler’,
which survives the Norman Conquest but dies out around the middle of the thirteenth
century (cf. MED s.v. dēme n.1). Already in Old English times dēma has two suffixed
formations beside it — i.e. dēmend and dēmere (cf. DOE [Link].; MED [Link]. dēmere, dēmend;
OED s.v. deemer). In Middle English (before 1300) a further derivative in -estre is formed
that occurs in a number of variants (cf. MED s.v. dēmester(e; OED [Link]. deemster, dempster).
All these formations have ultimately been replaced by the French loan judge, which according
to the OED entered the English language at the beginning of the fourteenth century.
22. On diagrammatic iconicity see Haiman (2000) and the literature cited there. For an
attempt to describe the formal and conceptual structure of prototypical agent nouns in terms
of Peircean iconicity cf. Kornexl (1998: 64ff.).
23. In Sauer’s material, which covers the twelfth and thirteenth centuries (1992: 4f.), the Old
English final exponents, which originally represented different vowel qualities, have merged,
the resulting schwa being usually represented by ·-eÒ. Its morphological status remains unclear:
“Die ae. Endungen -a, -e, -u sind me. unter -e zusammengefallen, das aber wohl nicht mehr als
[Link].-Endung zu beschreiben ist” (Sauer 1992:202). For recent discussions of this question
cf. Kastovsky (1997: 72; 2000a: 720f.), who claims that “these endings lost their inflectional
function and were reinterpreted as part of the word itself, which therefore changed its status
from a stem to a free form (word) without any inflectional ending” (2000a: 721).
24. Cf. Sauer (1992: 201). Pilch (1985:422f.) seems undecided as regards the Modern English
output: “The tatpuruşa with zero-suffix such as chimney sweep, presumably, belong to a
residual paradigm which was productive in Old English, forming both masculine and
feminine n-stems”.
25. Cf. OED s.v. chimney-sweep senses 1 and 2. For a detailed analysis of the evidence cf.
Kornexl (1998: 56f.).
126 Lucia Kornexl

26. By referring to bell-hop the OED explicitly suggests such a connection for car-hop (s.v.
car n.).
27. Lifeguard may be considered an exception, but with guard having been borrowed from
French, the formation is not a natural member of the family anyway.
28. As pointed out in Section 1, Marchand’s material has been analysed in detail in Kornexl
(1998) so that a few examples may suffice here.
29. Marchand (1963a; 1963b; 1964) advocates a content-based and strictly synchronic
method of establishing the derivational relationship between pair members in cases of zero-
derivation, whose results may run counter to the historical evidence. For a brief discussion
of Marchand’s “principle of semantic-pragmatic dependency” (Sanders 1988: 173) from the
perspective of historical lexicography see Kastovsky (2000b: 121f.).
30. It has to be stressed that — were it not for certain systemic forces such as the above-
mentioned opposing tendencies with regard to the productivity of unsuffixed action and agent
nouns — agentive simplexes of the cook-type irrespective of their origin could indeed have
‘caught on’ and established a productive pattern. After all, the English lexicon contains also a
number of inherited agents without formal marking such as smith (OE smiþ) and thief (OE þēof).
31. In general, nouns “involving fewer complexities in the cognitive system are also likely to
be morphologically less complex than sophisticated ones” (Schmid 1999: 224). Compared
with prototypical agent derivatives, attributives no doubt exhibit a much more complex
conceptual structure. According to the above hypothesis this should actually result in a
greater degree of formal complexity which these monemes, however, completely lack. For a
first approach to this problem cf. Kornexl (1998: 64–70).

References

Ayto, John. 1999. Twentieth Century Words. Oxford: Oxford University Press.
Bammesberger, Alfred. 1990. Die Morphologie des urgermanischen Nomens. Heidelberg: Carl
Winter.
Baron, Dennis. 1986. Grammar and Gender. New Haven & London: Yale University Press.
Bauer, Laurie. 1983. English Word-Formation. Cambridge: Cambridge University Press.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999.
Longman Grammar of Spoken and Written English. Harlow, Essex: Pearson Education
Limited.
Booij, Geert, Christian Lehmann & Joachim Mugdan, eds. 2000. Morphologie — Morphology:
Ein internationales Handbuch zur Flexion und Wortbildung — An International Hand-
book on Inflection and Word-Formation. Berlin & New York: Walter de Gruyter.
Campbell, Alistair. 1959. Old English Grammar. Oxford: Clarendon.
Dalton-Puffer, Christiane. 1996. The French Influence on Middle English Morphology: A
Corpus-based Study of Derivation. Berlin & New York: Mouton de Gruyter.
DOE = Cameron, Angus, Ashley Crandell Amos & Antonette diPaolo Healey, eds. 1986–.
Dictionary of Old English. Toronto: Pontifical Institute of Mediaeval Studies.
 From gold-gifa to chimney sweep? 127

Don, Jan, Mieke Trommelen & Wim Zonneveld. 2000. “Conversion and Category Indeter-
minacy”. Booij, Lehmann & Mugdan 2000. 943–952.
Dressler, Wolfgang U. 1990. “The Cognitive Perspective of ‘Naturalist’ Linguistic Models”.
Cognitive Linguistics 1.75–98.
Dressler, Wolfgang U. 2000. “Naturalness”. Booij, Lehmann & Mugdan 2000. 288–296.
Dressler, Wolfgang U., Willi Mayerthaler, Oswald Panagl & Wolfgang U. Wurzel. 1987.
Leitmotifs in Natural Morphology. Amsterdam & Philadelphia: John Benjamins.
Fisiak, Jacek, ed. 1985. Historical Semantics, Historical Word-Formation. Berlin, New York &
Amsterdam: Mouton.
Haiman, John. 2000. “Iconicity”. Booij, Lehmann & Mugdan 2000. 281–288.
Kärre, Karl. 1915. Nomina Agentis in Old English. Part 1: Introduction. Nomina Agentis with
l-Suffix. Nomina Agentis in -end, with an Excursus on the Flexion of Substantival Present
Participles. Uppsala: University Press.
Kastovsky, Dieter. 1968. Old English Deverbal Substantives Derived by Means of a Zero
Morpheme. Esslingen: Langer.
Kastovsky, Dieter. 1971. “The Old English Suffix -er(e)”. Anglia 89.285–325.
Kastovsky, Dieter. 1985. “Deverbal Nouns in Old and Modern English: from Stem-formation
to Word-formation”. Fisiak 1985. 221–261.
Kastovsky, Dieter. 1992. “Semantics and Vocabulary”. The Cambridge History of the English
Language. Vol. I: The Beginnings to 1066 ed. by Richard M. Hogg, 290–408. Cambridge:
Cambridge University Press.
Kastovsky, Dieter. 1997. “Morphological Classification in English Historical Linguistics: the
Interplay of Diachrony, Synchrony and Morphological Theory”. To Explain the Present:
Studies in the Changing English Language in Honour of Matti Rissanen. (= Mémoires de
la Société Néophilologique de Helsinki, 52) ed. by Terttu Nevalainen & Leena Kahlas-
Tarkka, 63–75. Helsinki: Société Néophilologique.
Kastovsky, Dieter. 2000a. “Inflectional Classes, Morphological Restructuring, and the
Dissolution of Old English Grammatical Gender”. Gender in Grammar and Cognition
ed. by Barbara Unterbeck, Matti Rissanen, Terttu Nevalainen & Mirja Saari, 709–727.
Berlin & New York: Mouton de Gruyter.
Kastovsky, Dieter. 2000b. “Words and Word-formation: Morphology in OED”. Lexicography
and the OED: Pioneers in the Untrodden Forest ed. by Lynda Mugglestone, 110–125.
Oxford: Oxford University Press.
Kluge, Friedrich. 1926. Nominale Stammbildungslehre der altgermanischen Dialekte, 3rd ed.
by Ludwig Sütterlin & Ernst Ochs. Halle (Saale): Max Niemeyer.
Knutson, A. 1905. The Gender of Words Denoting Living Beings in English and the Different
Ways of Expressing Difference in Sex. Lund: Håkan Ohlsson.
Koefoed, Geert & Jaap van Marle. 2000. “Productivity”. Booij, Lehmann & Mugdan 2000.
303–311.
Kornexl, Lucia. 1998. “Nomina agentis und die sog. agentiven Nullableitungen im Englisch-
en: eine semantisch-kognitive Neubewertung”. Kognitive Lexikologie und Syntax. (=
Rostocker Beiträge zur Sprachwissenschaft, 5) ed. by Friedrich Ungerer, 49–75. Rostock:
Universität Rostock, Philosophische Fakultät.
Krahe-Meid = Krahe, Hans. 1969. Germanische Sprachwissenschaft. III: Wortbildungslehre,
7th ed. by Wolfgang Meid. Berlin: Walter de Gruyter.
128 Lucia Kornexl

Lass, Roger. 1994. Old English: A Historical Linguistic Companion. Cambridge: Cambridge
University Press.
LDCE = Summers, Della et al., eds. 1995 [3rd ed.]. Longman Dictionary of Contemporary
English. Harlow, Essex: Longman.
Lindheim, Bogislav von. 1958. “Die weiblichen Genussuffixe im Altenglischen”. Anglia
76.479–504.
Marchand, Hans. 1963a. “On Content as a Criterion of Derivational Relationship between
Words Unmarked by Derivational Morphemes”. Indogermanische Forschungen
68.170–175.
Marchand, Hans. 1963b. “On a Question of Contrary Analysis with Derivationally Connect-
ed but Morphologically Uncharacterized Words”. English Studies 44.176–187.
Marchand, Hans. 1964. “A Set of Criteria for the Establishing of Derivational Relationship
between Words Unmarked by Derivational Morphemes”. Indogermanische Forschungen
69.10–19.
Marchand, Hans. 1965. “The Analysis of Verbal Nexus Substantives”. Indogermanische
Forschungen 70.57–71.
Marchand, Hans. 1969 [2nd rev. ed.]. The Categories and Types of Present-Day English Word-
Formation. A Synchronic-Diachronic Approach. München: C.H. Beck.
MED = Kurath, Hans, Sherman M. Kuhn & Robert E. Lewis, eds. 1952–2001. Middle English
Dictionary. Ann Arbor, Mich.: University of Michigan Press.
OED = Simpson, John A. & Edmund S.C. Weiner, eds. 1989 [2nd ed.]. The Oxford English
Dictionary. Oxford: Oxford University Press.
Paul, Hermann. 1920. Prinzipien der Sprachgeschichte. 5th ed. Tübingen: Max Niemeyer
[10th, unaltered ed. (= Konzepte der Sprach- und Literaturwissenschaft, 6) 1995; first ed.
1880].
Pennanen, Esko V. 1971. Conversion and Zero-derivation in English. (= Acta Universitatis
Tamperensis Ser. A, Vol. 40). Tampere: Tampereen Yliopisto.
Pilch, Herbert. 1985. “The Synchrony-Diachrony Division in Word-formation”. Fisiak 1985.
407–433.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive
Grammar of the English Language. London & New York: Longman.
Sanders, Gerald. 1988. “Zero Derivation and the Overt Analogue Criterion”. Theoretical
Morphology: Approaches in Modern Linguistics ed. by Michael Hammond & Michael
Noonan, 155–175. San Diego, Calif.: Academic Press.
Sauer, Hans. 1992. Nominalkomposita im Frühmittelenglischen. Mit Ausblicken auf die
Geschichte der englischen Nominalkomposition. Tübingen: Max Niemeyer.
Schmid, Hans-Jörg. 1999. “Towards a Functional-Cognitive Lexicology of Nouns”. Words,
Lexemes, Concepts — Approaches to the Lexicon: Studies in Honour of Leonhard Lipka ed.
by Wolfgang Falkner & Hans-Jörg Schmid, 213–226. Tübingen: Gunter Narr.
Sievers-Brunner = Brunner, Karl. 1965 [3rd ed.]. Altenglische Grammatik. Nach der angel-
sächsischen Grammatik von Eduard Sievers. Tübingen: Max Niemeyer.
Štekauer, Pavol. 1996. A Theory of Conversion in English. Frankfurt am Main: Peter Lang.
Štekauer, Pavol. 2000. English Word-Formation: A History of Research (1960–1995). Tübingen:
Gunter Narr.
</TARGET "kor">

From gold-gifa to chimney sweep? 129

Szymanek, Bogdan. 1988. Categories and Categorization in Morphology. Lublin: Katolickiego

Uniwersytetu Lubelskiego.
Valera, Salvador. 2000. “Conversion and Onomasiological Theory”. Journal of Linguistics
36.145–155.
Weinreich, Otto. 1971. Die Suffixablösung bei den Nomina agentis während der althoch-
deutschen Periode. Berlin: Erich Schmidt.
<LINK "kru-n*">

<TARGET "kru" DOCINFO

AUTHOR "Manfred Krug"

TITLE "A path to volitional modality"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

A path to volitional modality*

Manfred Krug
University of Freiburg

1. Aims, methodology, previous research

This paper investigates from a construction grammar perspective the semantic and
syntactic development of an English verb that has become more grammatical in
the history of English. Construction grammar (cf. Fillmore 1985; Croft 2001;
Traugott forthcoming) takes every construction to possess both form and (at
least one) meaning. want, the verb under investigation, has undergone radical
changes. I shall show, however, that the different stages it has gone through do
not represent an arbitrary succession but that all of them can be motivated
syntactically, semantically or pragmatically. I shall further show how construct-
ional alternatives and changes bear on the text frequency of want. The present
paper focuses in particular on the transitional period from Early Modern to
Late Modern English, i.e. on the eighteenth century, which has been neglected
in previous research, even though it is during this period that modal semantics
(necessity, volition) and modal syntax (want + to infinitives) emerge.
A few definitional remarks seem necessary before embarking on the analysis
proper. Bolinger (1980b: 297) once famously observed that “the moment a verb
is given an infinitive complement, that verb starts down the road of auxiliari-
ness”. This statement encapsulates the crucial fact that ‘auxiliarihood’ and
‘modalhood’ are graded concepts — a view which was embraced already by
Visser (1969: §1727) and which has become commonplace in more recent work
(e.g. Bolinger 1980a; Quirk et al. 1985: 135–148; Palmer 1989: 4f.; Biber et al.
1999: 483; Krug 2000: passim). Depending on their terminological preferences,
some researchers might thus prefer to call in particular certain early want
constructions ‘incipient’ or ‘emerging auxiliaries’ (or ‘modals’), ‘auxiliary-like
catenatives’, ‘quasi-modals’ or ‘semi-auxiliaries’. For simplicity and clarity, I
shall call an occurrence of want a ‘modal auxiliary use’ in the remainder of this
paper if it meets the following five criteria:
132 Manfred Krug

(1) a. want is a verb;

b. want governs a to infinitive;
c. want has modal semantics;
d. the grammatical Subjects of the want clause and the infinitival
clause are coreferential;
e. the to infinitive is not the verb in a purposive clause.

Most of this definition is uncontroversial. Criteria (d) and (e) exclude struc-
tures like Who do you want to win? and I want my glasses to read the paper,
respectively. As far as the third criterion is concerned, relevant in the present
study are the domains of necessity, volition and obligation (see Palmer
1986: Chs. 2f.; 1989: passim for detailed discussion). The one potentially
controversial criterion is perhaps the second. After all, on standard assumptions
(e.g. Quirk et al. 1985: 137) it is precisely the absence of infinitival to that
distinguishes the nine central modals (can, may, must, etc.) from more periph-
eral members of the modal category in English. While this is true synchron-
ically, it impedes our understanding of the diachronic development of want.
As is shown in Krug (2000: Ch. 5), verbs taking bare infinitives in English form
a closed class. Unless there is early variation between to and bare infinitives (as
for help or dare), the modern path to modal status is for verbs to take to
infinitives. The infinitival marker may be obscured as in wanna or gonna, but
typically at least traces of it are retained.
This is not the first study to sketch the historical development of want.
Bertschinger (1941) offers a detailed semantic study of the verbal uses; Krug
(2000: Ch. 4) is a quantitative study of the entire lexeme and thus includes, for
instance, nominal uses. The present investigation complements these studies in
a number of ways. Its primary goal is to identify the focal senses of the verb in
the history of English. It is thus akin, but not identical to Bertschinger’s
approach, who carves up the semantic space for want in ways that are not
always easy to apply to actual data. I shall not, for instance, follow Bertschinger
in distinguishing between ‘desire’ and ‘will’ readings (his use ‘desire’ is meant
to refer to unattainable objects, whereas ‘will’ presupposes their attainability).
The distinction I make is syntactic in nature and has no implications concern-
ing the attainability of an object or action: ‘desire’ will be used for want + NP,
while ‘volition(al)’ refers only to constructions with a following to infinitive. By
the same token, ‘need’ refers to NP complements, whereas ‘necessity’ is reserved
for infinitival complements alone.
The historical data for the present investigation are drawn from two
diachronic corpora (see the Appendix for further detail): the historical part of
 A path to volitional modality 133

the Helsinki Corpus (HC, Old English to Early Modern English texts) and the
drama and fiction components of A Representative Corpus of Historical English
Registers (ARCHER, 1650–1990). Furthermore, I checked the fictional compo-
nents of the Lancaster-Oslo-Bergen Corpus (LOB, texts from 1961) and its
modern Freiburg analogue, FLOB (texts from 1991). In addition, I consulted
standard reference works on historical syntax (Visser 1963–1973; Mustanoja
1960), the OED (1989 edition on CD-ROM)1 and a number of literary works
(drama and prose). Why, one might ask, the focus on drama and fiction? If one
wants to establish, as this paper does, the focal senses of a given construction,
one has to determine ‘focal in what variety’, since language change proceeds at
differential speeds in different registers (Biber 1998; Hundt & Mair 1999).
Ideally, of course, we would like to know when a given change occurred in
spoken English, of which we have unfortunately no historical records. It is
desirable therefore to approximate as closely as possible the spoken facts. As
Biber (1998) has shown, the historical registers that resemble spoken English
most are drama and fiction.2
More specifically, my approach was as follows. For the Middle English
period I analysed the historical component of the Helsinki Corpus, which
yielded a clear picture as far as the dominant senses and construction types are
concerned (there was essentially only one sense, viz. ‘lack’). For the time since
Early Modern English, my approach was two-fold. First, I compared the corpus
examples in the fiction and drama components of ARCHER (c. 700,000 words)
to the Helsinki Corpus (c. 550,000 words). Since, however, contextual informa-
tion is less readily accessible in such balanced corpora consisting of relatively
short text samples, I then analysed in more detail the following literary works
(e-texts from Project Gutenberg):
a. William Shakespeare: Complete Works (c. 1600; c. 880,000 words);
b. Daniel Defoe: The Life and Strange Surprising Adventures of Robinson Crusoe
and The Farther Adventures of Robinson Crusoe (1719; totalling c. 220,000
words; henceforth: Robinson Crusoe I+II);
c. Charles Dickens: Oliver Twist (1837/1838; c. 160,000 words);
d. Jerome K. Jerome: Three Men in a Boat — To Say Nothing of the Dog! (1889;
c. 70,000 words).
For Present-day English, finally, I analysed the fictional components of LOB
and FLOB (1961 and 1991; each containing approximately 250,000 words).
Even though the use of individual literary works entails the disadvantage of not
being representative, it was deemed necessary for the present study because
134 Manfred Krug

analysing decontextualized corpus examples from a historical period can prove

difficult. Present-day intuitions may easily lead a researcher to read meanings
into want examples that are not identical to those of the speaker/writer’s
mental lexicon. Once, however, we look at individual works in some detail, a
host of examples enables the analyst better to identify with the historical
conceptualizer, thus creating a certain awareness of the intuitions prevalent at
the time of composition. Needless to say, the literary works must be known to
the analyst. To give but one example, in Defoe’s (1719) Robinson Crusoe, we
find the following passage:
(2) robinson meets natives
I then made signs to them for some water, and held out one of my jars to
them, turning it bottom upward, to show that it was empty, and that I
wanted to have it filled.

On the face of it, this sentence presents no syntactic or semantic difficulty for
the present-day reader. This might induce the analyst to interpret this instance
of want against the backdrop of present-day intuitions as a straightforward
case of volitional semantics, since this is today’s default reading (Burchfield
1996: 832). On second thoughts, however, this seems to be a less than ideal
analysis for robinson meets natives. If we compare the bulk of verbal want
examples in Defoe’s two novels entitled Robinson Crusoe, we immediately realize
that most examples are better paraphrased by ‘need’ and that this is probably also
the foregrounded meaning in (2). And of course it is true that the protagonist
not only wished to have water but that he actually needed water to survive.

2. The history of want

As a reference point for the remainder of this paper, Figure 1 encapsulates the
major results of the following discussion of the semantic and syntactic stages of
want constructions. It goes without saying that such a chart is a simplification
of the actual facts. First of all, more constructions have been recorded. For the
verb alone, the OED (s.v. want) lists five main senses with as many as 28
subsenses and constructions, some of which were short-lived. I have focused on
those that help us understand the overall development of verbal want, paying
particular attention to modal constructions. This requires concentrating on
constructions that became dominant at some point in time. In addition, I
include constructions that reflect necessary intermediate steps, cognitively as
 Construction Example(s) 1200 1300 1400 1500 1600 1700 1800 1900 2000
(sense)
I impersonal hym (obj) wanted audience *
(lack) Experiencer (dat/obj) WANT stimulus (subj)

II transitive + NP And alle-þogh þai (3PL subj) in helle want light,Yhit sal þai of alle
(lack) payns haf sight,...

III transitive + NP I wante monaye, argent me fault.

(need) This room wants cleaning.

dominant verbal usage at a given time].

You’re exactly the one we wanted!

IV AUX + to infinitive I then made signs to them for some water, and held out one of my jars
(necessity) to them, turning it bottom upward, to show that it was empty,
and that I wanted to have it filled.

V transitive + NP She wants a red four-door convertible.

(desire)

VI AUX +to infinitive I want to/wanna go to the cinema.

(volition)

VII AUX +to infinitive You want to turn left at the next corner.
(obligation) You want to/wanna be careful.

Figure 1. The syntactic and semantic development of verbal want [*bold lines indicate
A path to volitional modality
135
136 Manfred Krug

well as diachronically, on the path to a new dominant construction.

Another potential issue with Figure 1 is — as with all such charts — that
starting points and endpoints of the arrows representing the existence of certain
constructions are notoriously debatable. It is clear, for instance, that as a new
sense or construction enters the language, its use will be sporadic. Likewise it is
trivially true that before a construction becomes obsolete, it necessarily becomes
rarer and obsolescent, and may continue to be used in special registers or
regional dialects (cf. fnn. 3 and 5).
Finally, when exactly a new sense becomes dominant in a given variety can
sometimes not be pinpointed, in particular when constructions coexist whose
senses are related or indeed overlapping. An allowance of 50 years should
therefore always be made. An additional problem (what register to concentrate
on) has been discussed in the introductory section of this paper.

2.1 Middle English and Early Modern English (1100–1700)

As is standard practice, Middle English is taken to end in 1500. With Barber
(1997: 1) I take Early Modern English to be the period from 1500 to 1700. The
somewhat disputed beginning of the Middle English period (figures vary from
1066 to 1150) is irrelevant in the present study since want entered the English
language only at the turn of the twelfth century. It is most probably a loan from
Old Norse, which had a verb vanta ‘lack’ that took two accusative arguments.
Like its Scandinavian cognate, English want had ‘lack’ semantics.3 The details
for the early period are given in Bertschinger (1941: 6–20) and Krug
(2000: 118–127). I shall therefore confine myself to a very brief sketch of the
Middle English period. Below, two early attestations of want are given, both
dating from about 1200. They are the oldest examples of verbal want in English
that I have been able to trace (glosses and translations are my own).
(3) happy life i
(Middle English component of the Helsinki Corpus, SAWLES 174:6;
c.1225, ?c.1200)
ne schal ham (dat pl) neauer wontin.
neg shall them never lack
“They shall never lack (anything)”
(4) happy life ii
(Middle English component of the Helsinki Corpus, HALI 131:4M;
c.1225, ?c.1200)
for as seinte pawel seið. alle þing turneð þen gode. to gode.
 A path to volitional modality 137

ne mei na þing (subj) wonti þe (obl).

neg may no thing want thee
“You shall lack nothing”

Both constructions are impersonal and have a dative Experiencer (i.e. the one
who lacks). Notice that (3) wants a Subject (also called Cause or Stimulus in
impersonal constructions; here: the argument lacked). It is only such subjectless
constructions that are covered by narrow definitions of ‘impersonal verb’ such
as that given by Denison (1993), which encapsulates Fischer & van der Leek’s
(1983) position:
An impersonal construction is a subjectless construction in which the verb has
3SG form and there is no nominative NP controlling verb concord; an imper-
sonal verb is a verb which can, but need not always, occur in an impersonal
construction (1993: 62).

Others (such as Jespersen 1927 or Allen 1986) have adopted a wider definition
to include all non-personal uses (see also López-Couso & Méndez-Naya 1997
on terminological issues):
(5) kingly taste
þam kynge (dat) licodon (pl) peran (nom)
to the king liked pears
“The king liked pears”

Unlike (3) — but like (5) — the sentence in (4) has a Subject even though it is
only a (grammaticalizing) pronoun ‘nothing’, rather than a full NP.4 Imperson-
al constructions of both types (3) and (4) represent the first dominant use of the
verb. They decrease in frequency in the second half of the Middle English
period and become obsolete round about 1500. Personal, transitive want (still
meaning ‘lack’) followed by an NP, is historically the next dominant use (see
Krug 2000: Ch. 4 for the details of the development). According to Mustanoja
(1960: 435, quoting van der Gaaf 1904), it is first attested in c.1350. It must have
spread rapidly because by about 1400, transitive want is the most common use.
An early example is given below.
(6) (Middle English component of the Helsinki Corpus, PRICK 253:15,
1425; a.1400)
And alle-þogh þai (3pl, subj) in helle want light, / Yhit sal þai of alle
payns haf sight,
“And although they lack light in hell, yet they shall see all pains,”
138 Manfred Krug

The meaning ‘lack’ fell into disuse by roughly 1900 (the last OED citation is
from 1876; later ones are from highly specialized registers, e.g. palaeography).5
Nevertheless, the syntactic construction itself has arguably continued to be the
most frequent one to the present day, that is, if one considers noun phrase
complements and infinitival complements as functionally equivalent, as is
indicated by the following bracketing:
(7) I want [light]
(8) I want [to go]

One can, however, consider the later extension from nominal NP Objects to
infinitival complements a radical syntactic step, which I do (see next section).
If one did not, one would subscribe to the position that subsequent changes are
solely pragmatic or semantic in nature.
Towards the end of the Middle English period, a meaning extension from
‘lack’ to ‘need’ is observable. The new sense, however, comes to be the domi-
nant one only 200 years later, i.e. in the middle of the seventeenth century. An
early example is:
(9) money wanted
I wante monaye, argent me fault. (1530, Palsgr. 771/1, quoted from OED
s.v. want v. 4.a)

As is strongly suggested by the French gloss, example (9) is a case of ‘need’

semantics. This sense has survived to the present day, even though as the sole
meaning it is rare now. Consider two Present-day English examples, which have
both a ‘desire’ and a ‘need’ reading:
(10) job interview
She’s exactly the one we wanted!
(11) outperforming one’s opponent
This was exactly what Angel One wanted. For his plan to succeed, he
needed to create over-confidence in his opponent. (BNC HJD 2598)

As is stated in Burchfield (1996: 832) and as is borne out by the present investi-
gation, too, the focal sense in the late twentieth century is volitional. By
contrast, (exclusive) ‘need’ readings are now largely confined to constructions
in which want takes a gerundial Object:
 A path to volitional modality 139

(12) dirty carpet

The mess on that carpet wants cleaning. Go and get some water, go and
get the cloth quickly. (BNC KR0 2388)

Let us summarize the situation in Early Modern English. Almost until the end
of the period, the dominant use of verbal want is still ‘lack’ as in (13).
Throughout the period, however, one can in many cases argue for at least a
backgrounded ‘need’ reading, as in (14). As was mentioned earlier on, want in
Early Modern English occurs exclusively in personal constructions, typically
taking a nominal Object. Infinitives are not yet among the complements of
want. There are, however, some instances that can be argued to be precursors
of the modal construction. These are transitive verbs that are followed by
purposive to clauses, e.g.:
(13) speechless
but Sir William, thinking the foole wanted wit to tell his griefe (though not
wit to play the thiefe) had the barber depart, asking Jacke what he would
eate? he sayd, nothing. What he would drinke? he sayd, nothing; which
made Sir William doubt much of his health, (HC, FICT2A 13:16; 1608)
(14) poverty
So everything stands still for money, while we want money to pay for some
of the most necessary things that we promised ready money for in the
heighth of our wants — (HC, DIAR3A VIII, 315:7; 1666)

While the ‘lack’ example (13) is still a far cry from the modal construction, (14)
reveals a possible path to modalhood. On the semantic side, the modal notion
of ‘need’ is an equally plausible analysis as nonmodal ‘lack’. One can even argue
that in (14) faint traces of ‘desire’ are foreshadowed (which comes to be the
dominant meaning more than a century later): the Subject argument — the one
who needs the money — probably also wishes to pay for his debts. Syntactically,
too, this is a highly interesting construction. The underlying agent of the
infinitival to clause is also the speaker (and his wife), irrespective of whether the
meaning is ‘lack’, ‘need’ or ‘desire’. This entails that, in addition to the follow-
ing to infinitive, another criterion of modalhood mentioned above is met, viz.
subject identity of main and to clause.
To summarize, in Early Modern English the typical meaning is still either
the original (Old Norse) ‘lack’ or the newly developed ‘need’ (first attested
according to the OED in 1470). In the second half of the seventeenth century,
the two meanings enjoy similar frequencies. The analysis of ARCHER data from
1650–1699 splits the uses into three equal parts: one third has ‘lack’ semantics,
140 Manfred Krug

another third ‘need’ semantics and the last third is ambiguous between the two
senses. Thus, at the very end of the period, in 1700, the two meanings of the
verb are still competing, with ‘need’ just having taken over the dominant
position from ‘lack’. Obviously these two senses are very closely related. The
extension from the older to the newer sense, therefore, is not entirely arbitrary
(see Section 3 below on the motivations). They are contiguous, indeed overlap-
ping senses. It is not surprising, then, that a full third of the verbal uses in
ARCHER drama and fiction from the period 1650–1699 are ambiguous.
Witness, for instance:
(15) proposal: wife and husband wanted
Thus ’tis (Sir) in short: your Daughter (do ye conceive me) wants a Hus-
band; and I want a Wife (do ye conceive me;) Now what are we born for
in this world, but to supply one another’s wants? (1671, ARCHER
Caryll.D1)
(16) praise of a young lady
She is one, who has received no improvement from Education; Nor does
she want it: For, Nature has left her so well finished, that Art has little to
do. (1671, ARCHER Caryll.D1)

While the semantics of want in wife and husband wanted leans perhaps
more towards ‘need’, it certainly possesses traces of ‘lack’ and ‘desire’. In praise
of a young lady, however, no decision in favour of ‘lack’ or ‘need’ is possible,
one reason being that the latter sense entails the former. Grice has noted that “it
may not be impossible for what starts life, so to speak, as a conversational
implicature to become conventionalized” (1989: 39). Given the observed
semantic contiguity, it is highly plausible to assume that the change from ‘lack’
to ‘need’ is motivated by the conventionalization of the following implicature:
(17) +> If somebody lacks something, he or she will usually also need it (oth-
erwise stating the lack would be unnecessary or odd).6

2.2 From Late Modern English to Present-day English (1700–2000)

2.2.1 The rise of the modal auxiliary construction want + VP
The end of the Early Modern English period marks a convenient section
boundary for the present study since it is roughly in 1700 that as many as three
new syntactic constructions (and one new meaning) enter the language and that
the ‘need’ sense properly takes over the pole position. More important still for
the rise of volitional modality, two of these new constructions are auxiliary uses
 A path to volitional modality 141

in which want is followed by a to infinitive. In other words, the crucial period

for the emergence of modal want constructions is the transition from the Early
Modern to the Late Modern period, both from a syntactic point of view and also
semantically, because with the dominance of ‘need’ semantics, want is now
firmly situated within notional modality. In order to work out the details of the
change, we need to consider some textual evidence from the transitional period.
Previous accounts have left the intricacies of the extension from nominal to
verbal complements largely unexplained. In order to understand better the
mechanisms that have led to the development of the modal construction, we
will look at some constructions which are akin to modalhood, be it from a
syntactic or semantic point of view. A text which proved instructive in spotting
the types of changes that have affected want + to infinitive is Defoe’s Robinson
Crusoe from 1719, i.e. the beginning of the Late Modern period (the examples
in the remainder of this section are all taken from the e-texts of the two
volumes).7 The most frequent meaning of verbal want in this work is ‘need’,
but many times it is the older ‘lack’, and rather often a backgrounded ‘desire’
reading is possible. With so many relatively frequent coexisting senses, it comes
as no surprise that sometimes the meaning of want is indeterminate, in
particular between the two senses ‘lack’ and ‘need’, which are doubly motivated
by entailment and, conversely, by invited inferences (see the discussion of (16)
above). Compare:
(18) tools wanted i
When it [the corn] was growing, and grown, I have observed already
how many things I wanted to fence it, secure it, mow or reap it, cure and
carry it home, thrash, part it from the chaff, and save it.

The text goes on as follows:

(19) tools wanted ii
Then I wanted a mill to grind it, sieves to dress it, yeast and salt to make it
into bread, and an oven to bake it; but all these things I did without, as
shall be observed; and yet the corn was an inestimable comfort and
advantage to me too.

These examples have an interesting syntactic structure which may provide

important clues for identifying the extension of want to infinitival comple-
ments. As in (13) and (14), we have to do with purposive (‘in order to’)
constructions following want + NP. In (18), want and to are even contigu-
ous, which is the typical word order in modal constructions. And, as in (14), the
142 Manfred Krug

Subjects in the main and the subordinate clause are coreferential. It seems very
likely, then, that such structures have facilitated the extension from nominal to
infinitival complements, that is, proper auxiliary constructions that meet all
criteria laid out in (1). In fact, subject identity is also in evidence for (19), even
though some analysts might consider it to be a case of [want sth./sb. to do
sth.], since it is the mill which grinds the corn, etc. However, an underlying
human agent is present, too, namely the operator of the devices (mill, sieves,
oven). In addition, there is an interesting semantic facet to be found in (18) and
(19), which points to the semantic development of volitional modality: when
the speaker/writer says: “I wanted many things”, he ‘lacked and needed’ them,
as the highlighted passage but all these things I did without proves. Crucially,
however, an inference to ‘volition’ is possible. Robinson Crusoe, the Subject,
also actually desired to fence the crop, etc. Given the currency of such construc-
tions in the early eighteenth century, transitive verbs that are followed by
purposive to clauses whose Subjects are identical to that of the superordinate
clause are a likely factor in the development and propagation of modal want
constructions.
Purposive clauses apart, there is an alternative or, rather, additional path to
the extension from nominal to infinitival complements: by co-ordination of NP
and VP. Example (20) is a case in point:
(20) conversation between crusoe and priest about atkins
“Oh”, said the priest, “tell him there is one thing will make him the best
minister in the world to his wife, and that is repentance; for none teach
repentance like true penitents. He wants nothing but to repent, and then
he will be so much the better qualified to instruct his wife;”

In this example, the meaning of want is ‘need’. Syntactically, we observe the

co-ordination of a noun phrase complement (pronoun nothing), representing
the older construction, with a to infinitive. This might reflect the actual histori-
cal path of development: extension through co-ordinating, thus equating
syntactically the infinitive with a noun phrase. On this analysis, the noun phrase
and the action expressed by the infinitive are both perceived as needed or
desired entities. There is, however, an interesting alternative syntactic analysis
of (20). Obviously the pronoun nothing is not a prototypical NP. It has less
nominal force, in particular when immediately followed by but. If one analyses
the sequence nothing but as a complex adverb (it can be paraphrased by ‘only’
or ‘exclusively’), then this would seem to pave the way smoothly for infinitival
complements.
 A path to volitional modality 143

Early examples of auxiliary-like uses of a verb are often debated. The

following example is instructive in this respect because it has been classified in
different ways in the literature:
(21) education wanted
Madam: Inteed Matam, to say de trute, he wanted leetel good breeding.
Lady Fan: Good breeding! He wants to be caned, Madamoiselle. An
insolent fellow! (1697 Vanbrugh Provoked Wife II, ii, quoted from OED
s.v. want v. 4.b and Visser 1969: §1727)

The OED interprets the highlighted wants as ‘needs’. Visser, by contrast,

paraphrases it as ‘ought to’, classifying it along with his type You want to mind
your step:
Since in expressions like ‘you want to be careful’, ‘you want to watch your step’
want to semantically and structurally approximates ought to, it has the status
of an auxiliary. It frequently occurs in colloquial English. Although in modern
grammars it is treated as a rather recent innovation, the 1697 quotation (…)
[(21) above] shows that it was not unknown in the beginning of the seven-
teenth century. (It was apparently banned from the written language until the
twentieth century.) Whether the 1690 quotation from Dryden [viz. “they were
pedants; and for a just reward of their pedantic pains, all their translations
want to be translated into English.”] belongs here is rather doubtful.
(Visser 1969: §1727, emphasis added)

While I agree with Visser that such constructions enjoy auxiliary status — and
even though Visser’s ‘ought to’ is a perfectly acceptable gloss — I believe that
the OED paraphrase ‘need’ for (21) is superior to classifying it with the type You
want to mind your step, which is attested with a second person Subject for the
first time as late as 1913 (see (23) below).8 What Visser fails to realize is that all
the examples he quotes as belonging to this type can be classified as commands
or, according to more recent taxonomies (e.g. Huddleston 1984: 351 and Quirk
et al. 1985: 804, who follow essentially Searle 1976) as directives — except for
the rejected Dryden example from 1690 and example (21) from 1697. These two
are therefore likely to belong to a different type. Furthermore, of the remaining
22 relevant examples quoted by Visser, 21 have second person Subjects, giving:
You (don’t) want to + VP. The one exception is a third person directive:
(22) appointment
How does he know he’ll see me in the morning? He doesn’t want to be so
jolly sure about it. (1930, Priestley Angel Pavement, 383)
144 Manfred Krug

This shows that, except for the Dryden quotation, all examples invoked by
Visser have animate, human Subjects. There are, then, both structural, prag-
matic and semantic reasons to assume that, pace Visser (1969: §1727), the
obligation sense (or more narrowly, the use in directives, i.e. his type You want
to mind your step) was probably indeed an innovation from the early twentieth
century and not “banned from the written language” for two centuries. Witness
the earliest obligation example quoted in Visser (1969: §1727):
(23) “You want to bring an old coat and leave it here.” (…)“now look here.
You want to copy these letters in here.”
(1913, D. H. Lawrence Sons & Lovers, 130)

It makes therefore sense to distinguish the older necessity reading (stage IV in

Figure 1, which arises around 1700) from the later use of want in directives.
The fact that directives are first recorded in dialectal usage (see English Dialect
Dictionary 1961 [1905] or references in Kirchner 1940) and in Visser’s time are
considered “frequent in colloquial English” (ibid.) — a statement which is less
true at the turn of the millennium, though (see the quotations in Krug
2000: 147f.) — strongly suggests that it is a change from below, in Labovian
terms (e.g. Labov 1994: 78; 2001: 76f., 196f., 517f.). And even today this usage is
not fully established in the written mode. Nor has it become the dominant
usage of verbal want.9
The only way to save Visser’s analysis would be to assume one super-
category of ‘necessity’ (see Palmer 1989: 31f. on the different types of this
concept) which has a subcategory for all obligation readings (including the
debated example (21)), which in its turn contains the subset of all directives.10
This can be depicted schematically as follows:
2nd person
(largest group)
(24) necessity obligation directives
other persons

Let us now turn to the rise of ‘desire’ and ‘volition’ at the beginning of Early
Modern English.11 It will be remembered that this study distinguishes between
the two notions on syntactic grounds as follows: ‘desire’ is used for want taking
nominal complements, whereas ‘volition’ is reserved for the modal construc-
tion, i.e. desiderative want with infinitival to complements. The earliest
relevant attestation in the OED is a modal construction from 1706:
 A path to volitional modality 145

(25) hurry
All such as want to ride in Post-haste from one World to the other.
(1706, E. Ward Wooden World Diss.)

Another clear example of early volitional modality can be found in ARCHER:

(26) dangerous lover
He caused the Words to be set to Notes, and then sung them himself in
all Companies where he came: His flatterers, who were numerous, and
did not now want to learn his weak Side, gave him the Title of the dan-
gerous Swain, which he prided himself in; till his Mistress grew down
right uneasy, and would have him visit Rivella no longer. (1714, AR-
CHER MANL.FC2)

While (26) is a clear case of volitional modality, examples (27) and (28) from
Robinson Crusoe demonstrate the possibility of semantic overlap between
‘necessity’ and ‘volition’. They show how difficult it can be to disambiguate
between these senses.
(27) bigger barns wanted
And now, indeed, my stock of corn increasing, I really wanted to build
my barns bigger; I wanted a place to lay it up in, for the increase of the
corn now yielded me so much, that I had of the barley about twenty
bushels, and of the rice as much or more;
(28) oven wanted
But for an oven I was indeed in great pain. At length I found out an
experiment for that also, which was this: I made some earthen-vessels
very broad but not deep, that is to say, about two feet diameter, and not
above nine inches deep. These I burned in the fire, as I had done the
other, and laid them by; and when I wanted to bake, I made a great fire
upon my hearth, which I had paved with some square tiles of my own
baking and burning also; but I should not call them square.

In (27), the protagonist both needs and wishes to build his barns bigger. The
meaning is therefore indeterminate. Example (28) is even more interesting
because here the hero does not only need and wish to bake; the context betrays
that he actually will bake later on. In other words, we can even see shades of
‘intention’ in that example.
If one focuses on infinitival complements, it transpires that want to has
typically expressed ‘volition’ since the nineteenth century, even though in
Dickens’ Oliver Twist (from 1837) the verb is still often ambiguous between
146 Manfred Krug

‘volition’ and ‘necessity’. More generally, it is in the late nineteenth century (e.g.
Jerome’s Three Men in a Boat from 1889) that the wishing branch becomes
clearly the dominant use. This situation has not changed to the present day. But
while in the nineteenth and early twentieth century NP complements (i.e.
‘desire’ readings) constitute the dominant want construction, in the second
half of the twentieth century it is volitional modality (i.e. infinitival to comple-
ments) which accounts for the majority of verbal uses at large. This is evidenced
by the figures from LOB and FLOB and the British National Corpus (cf. also
Figure 3 below).12

Table 1. Text frequency of want (to) in journalistic and fictional British writing (LOB,
FLOB) and the British National Corpus
LOB FLOB LOB FLOB Spoken Written
press press fiction fiction BNC BNC
1961 1991 1961 1991

All forms of want 95 129 365 389 26,289 68,622

[incidence per 10,000w] [5.4] [7.3] [14.4] [15.4] [26.3] [7.6]

Modal want to 49 68 192 207 13,182 38,420

[incidence per 10,000w] [2.8] [3.9] [7.6] [8.2] [13.2] [4.3]
Modal want to as % of all want 51.6% 52.7% 52.6% 53.2% 50.1% 56.0%

3. Motivations

In order to consider potential motivations for the observed changes, let us

summarize the history of verbal want. The progression of the dominant
meanings is as follows:
(29) (i) lack (1200) > (ii) need (1700) > (iii) desire (1800) > (iv) volition (1950)

The motivations for the semantic changes from one stage to the next are not
difficult to identify. All dominant meanings that are diachronically adjacent are
highly related and can be motivated in the following fashion by the gradual
conventionalization of inferences:13
a. What is lacked by somebody will often be needed by him or her (otherwise
stating the ‘lack’ would not be noteworthy).
b. When somebody needs something, he or she will also desire it.
 A path to volitional modality 147

c. If nominal Objects (typically physical objects) can be desired, so too can

actions and states of affairs.
The progression of dominant meanings in (29) is interesting on a more global
level, since it ties in neatly with grammaticalization theory. Intriguingly but not
unexpectedly, the dominant meanings become increasingly abstract in the
history of the language. Much of this path can be subsumed under a specific
type of pragmatic inference, ‘subjectification’ (as conceptualized by Traugott),
which tends to be strengthened as grammaticalization proceeds. Witness
Traugott’s Subjectification Tendency III (1989: 35):
Meanings tend to become increasingly based in the speaker’s subjective
belief /state/attitude toward the proposition.
For illustration, imagine the following scenario in which an ethnologist
observer takes the following notes (cf. Quine 1960: Chs. 1–3):
(30) anthropologist meets unsocked natives
(i) The natives lack socks.
(ii) The natives need socks.
(iii) The natives desire socks.
(iv) The natives desire to put on socks.

Other things being equal, i.e. without differing utterances or gestures from the
subjects, the ethnologist has to draw ever more inferences for noting (ii) to
(iv).14 Only (i) is the statement of a relatively uncontroversial fact since lack is
typically an objectively testable concept: the object lacked (if physical) is absent
from the subject. need, by contrast, is already grounded in the ethnologist’s
cognition since it is her who draws the inference from lack to need, perhaps
on the basis of outer circumstances such as cold temperature. In (iii) the
ethnologist would need to have access to the subjects’ cognition since it is only
they themselves who can positively assert whether or not they desire something.
Volition, that is the desire for an unfulfilled action as in (iv), finally, is next to
impossible to observe since it is even more abstract than the desire for physical
objects. In other words, historically the dominant meanings of want have
become increasingly inferential and abstract, and thus more grammatical.15
148 Manfred Krug

4. Verbal and auxiliary want: Senses and increase in

discourse frequency
This section gives a frequency sketch of the rise of want and its modal auxiliary
uses. Figure 2 depicts the text frequencies for the entire lexeme and the modal
construction, where the former includes all word forms of want, i.e. want,
wants, wanted and wanting, while the latter includes want to, wants to, wanted
to, verbal wanting to and wanna.
The graphs for both the lexeme and modal want to have roughly the shape
of S-curves. Both reach their saturation stage at the turn of the nineteenth
century. It is worth noting that the graph for the modal uses is a more perfect
S-curve than that for the entire lexeme. In fact, the latter approximates the
S-curve pattern only because it includes the modal uses. As is obvious from the
graph for the nonmodal uses, the incidence of lexical uses is more erratic, as it

18

16

14

12

10

4 all uses of WANT per 10,000 words

modal WANT TO per 10,000 words
2
nonmodal uses per 10,000 words
0
Oliver Twist (1837)

LOB fiction (1961)

FLOB fiction (1991)

Shakespeare

Robinson Crusoe

Three men in
a boat (1889)
(c. 1600)

I+II (1719)

Figure 2.
The rise of want and want to.

 A path to volitional modality 149

were, than the incidence of the auxiliary, i.e. grammatical, construction. It is an

important finding of recent work in cognitive linguistics that regularity can
often be observed when grammatical items undergo semantic change.16 If we
can generalize from the present findings, then grammatical change is also more
regular in terms of frequency gains. Thus, grammatical change would seem to
be easier to model quantitatively than lexical change. If this is correct, then
linguists can formulate more confidently and more accurately likely scenarios
of a quantitative development in the middle of an ongoing grammatical change.
This would represent a modest step towards predicting linguistic change, which
has generally been considered an unattainable task.
Let us, however, return to the level of linguistic description. The steepest
increase in the use of the lexeme and the modal construction occurs in the
nineteenth century, as the wishing branch becomes dominant. This is true both
in absolute terms (see Figure 2) and in relative terms (Figure 3). For the modal
construction, this is the time when two auxiliary constructions (necessity and
volition) exist side by side, with volition on the increase. Measured proportion-
ately against the entire lexeme, necessity never reached as high a share as
volition (cf. Figures 1 and 3), perhaps due to the competing lexeme need (to).
Even the lexical ‘need’ sense (+ NP) of want was relatively short-lived as the
sole dominant use. The modal necessity construction in fact never came to
occupy the position of the dominant verbal meaning.17
A look at Figures 2 and 3 further reveals that the rise in discourse frequency
goes hand in hand with functional streamlining. The more frequent the lexeme
want becomes, the higher becomes the share of the modal construction: from
0% in 1600 to nearly two-thirds in Present-day English, with the steepest
increase in the nineteenth century. The increase in the nineteenth century is in
fact even more drastic than the graph in Figure 3 suggests, because the two
works listed for that century are only separated by a time span of some fifty
years as against 70 to 120 years between the remaining points adjacent on the
abscissa. In relative terms (Figure 3), the saturation stage seems to have been
reached only in the middle of the twentieth century, i.e. 50 years later than in
absolute terms (Figure 2).
It turns out, therefore, that the beginning of the Late Modern period (i.e.
the early eighteenth century) is the most interesting period on syntactic and
semantic grounds since it saw the rise of three new constructions, two of which
were modal. It is the second half of this period, however, that is most interesting
on frequency grounds, since it is during the nineteenth century that the rapid
spread of the modal constructions occurred.
150 Manfred Krug

60
Proportion of modal construction relative to entire lexeme

40

20

0
Oliver Twist (1837)

LOB fiction (1961)

FLOB fiction (1991)

Shakespeare

Robinson Crusoe

Three men in
a boat (1889)
(c. 1600)

I+II (1719)

Figure 3.
Modal uses of want (to) as proportions of all uses of the lexeme want.

5. Concluding remarks

It was seen that syntactic and semantic ambiguity are two prime factors in the
developments of verbal want. It was in particular pragmatic inferencing that
triggered such ambiguity and thus propagated the changes from one dominant
construction to another. The present study, then, lends full support to the
hypothesis that change typically arises in environments where structures can be
interpreted in novel ways, as has been stated by, for instance, Aitchison
(2001: Ch. 7) or Harris & Campbell (1995: Ch. 4). But while these works refer to
syntactic change, I hope to have shown that syntax and semantics are closely
intertwined in the development of the verbal constructions of want.
In the early twentieth century, Jespersen remarked that much language
change seems to be “nothing but a purposeless fluttering hither and thither”
(1941: 9). At the beginning of the twenty-first century, we need no longer be
that pessimistic, at least not as far as grammatical and grammaticalizing items
<DEST "kru-n*">

A path to volitional modality 151

are concerned. Sweetser therefore rightly assumes regularity for sense develop-
ments in modal constructions:
Words do not randomly acquire new senses, then. And since new senses are
acquired by cognitive structuring, the multiple synchronic senses of a given
word will normally be related to each other in a motivated fashion. (1990: 9)

The present investigation strongly supports the existence of such motivated

links between synchronic polysemy and historical change. More specifically, it
suggests that it is chiefly pragmatic inferences, i.e. context-driven metonymic
processes, that have led to the observed changes of want.

Appendix: Literary works and corpora consulted

A Representative Corpus of Historical English Registers (ARCHER, texts from 1650–1990,

totalling 1.7m words; see Biber et al. 1994 for details).
The British National Corpus (BNC, current British English, totalling some 100m words; see
Aston & Burnard 1998 for details).
Diachronic Part of the Helsinki Corpus of English Texts (HC, Old English to Early Modern
English texts, totalling 1.6m words; see Kytö 1991 for details).
Freiburg Version of the Lancaster-Oslo-Bergen Corpus (FLOB, texts from 1991, totalling 1m
words; see Hundt et al. 1998 for details).
Lancaster-Oslo-Bergen Corpus (LOB, texts from 1961, totalling 1m words; see Johansson et
al. 1978 for details).
E-texts from Project Gutenberg (available at [Link]
Defoe, Daniel. The Life and Strange Surprising Adventures of Robinson Crusoe and The Farther
Adventures of Robinson Crusoe (1719; totalling c. 220,000 words).
Dickens, Charles. Oliver Twist (1837/1838; c. 160,000 words).
Jerome, Jerome K. Three Men in a Boat — To Say Nothing of the Dog! (1889; c. 70,000
words).
Shakespeare, William. Complete Works (c. 1600; c. 880,000 words).

Notes

* For stimulating discussions inside and outside the Santiago lecture hall, my thanks are
owed to various participants in the 11th ICEHL conference, in particular Cynthia Allen,
Johan van der Auwera, Douglas Biber, Merja Kytö and Elizabeth Traugott. Furthermore this
paper has profited from the comments by Teresa Fanego, Bernd Kortmann and two
anonymous referees.
152 Manfred Krug

1. While work on the third edition (partly available on-line) is in progress, the entry for
want is not yet revised.
2. Due to the scarcity of examples, especially for Middle English, the entire Helsinki Corpus
was chosen without genre discrimination.
3. This sense has survived to the present day, even though it is now rare and largely confined
to nominal and adjectival uses (e.g. in the complex preposition for want of and be wanting;
cf. also fn. 5).
4. See Kahlas-Tarkka (1997) on the pronominalization of thing.
5. As always, exact datings for the obsolescence of a meaning must be taken with a grain of
salt, since meanings — and syntactic constructions too — can linger on in non-mainstream
varieties (e.g. jargon, dialects, etc.). It seems uncontroversial, for instance, that a sentence like
This construction wants (‘lacks’) a Subject represents rather marked present-day usage.
6. For instance, if the grammatical Object were something undesirable, the speaker would not
normally comment on the lack. Hence the anomaly of a statement like ??Our village lacks cholera.
7. Most readers will be familiar with the book. For those who are not: for the most part, the
setting is a lonely island and there are hardly any interlocutors. This facilitates discussing
meanings on the basis of relatively short text excerpts. Negotiation of meaning between
hearer and speaker is largely irrelevant for this book.
8. Leaving aside the early date and the different person of the Subject, one might add that
Söderlind (1951, 1958) shares the present view. He too (1951:192; 1958:19, 23) notes that the
verb want in Dryden always means either ‘lack’ or ‘need’, that is, never ‘desire’ or ‘ought to’.
9. Notice, however, that directives are far less common in written language than in face-to-
face interaction.
10. Palmer’s (1989: 123) classification of ‘necessity’ is akin to the one presented here.
11. It was seen in the previous section (cf. the discussion of (14)) that occasionally back-
grounded ‘desire’ readings can be found in the seventeenth century, too.
12. The criteria laid down in (1) are applied to the searches for Table 1. In addition, the
search was restricted to contiguous want to. This is to say that only immediately adjacent
forms of want and to were included in the count. This covers the overwhelming majority of
modal uses (notice that I want him to leave is not modal) even though, theoretically at least,
such a search excludes examples like He wants really to read. But since, first, adverb interpola-
tion between the verb and to is highly dispreferred (Krug 2000: 139) and, second, the same
procedure is applied to all corpora, the figures are reliable for the comparison at hand.
13. On the gradual vs. leap (i.e. inference or metonymy vs. metaphor) issue in semantic
change see Sweetser (1990); Heine et al. (1991: Ch. 3); Heine (1993: 2f.); Hopper & Traugott
(1993: Ch. 4); Bybee et al. (1994); Traugott (1996). More recent work such as Bybee et al.
(1994: Chs. 1, 8) or Traugott (1996) tends to acknowledge that both processes are fundamen-
tal mechanisms in semantic change and grammaticalization, but favours more strongly
inference-driven accounts, which would be consistent with the present findings.
14. In fact, statements (ii) to (iv) are increasingly ethnomethodologically unsound.
 A path to volitional modality 153

15. In addition, it is patent that the final dominant construction, which involves the
concatenation of two verb phrases as in want to see, is more grammaticalized from a syntactic
point of view than a transitive verb.
16. Relevant examples include the change from deontic to epistemic meanings, which is a
recurrent development found with — though by no means restricted to — modal verbs (see
e.g. Sweetser 1990; Heine et al. 1991; Bybee et al. 1994; Traugott 1996).
17. It is worth noting that for those periods that are investigated in the present study and in
Krug (2000: Ch. 4, which is based on ARCHER material), the results are remarkably
congruent. The observable differences are probably largely due to the fact that the individual
works looked at here are synchronic slices of the language, while ARCHER periods cover
works from 50-year periods.

References

Aitchison, Jean. 2001 [1981]. Language Change: Progress or Decay? Cambridge: Cambridge
University Press.
Allen, Cynthia L. 1986. “Reconsidering the History of Like”. Journal of Linguistics
22.375–409.
Aston, Guy & Lou Burnard. 1998. The BNC Handbook: Exploring the British National Corpus
with SARA. Edinburgh: Edinburgh University Press.
Barber, Charles. 1997 [1976]. Early Modern English. Edinburgh: Edinburgh University Press.
Bertschinger, Max. 1941. ‘To Want’: An Essay in Semantics. (= Swiss Studies in English, 13).
Berne: Francke.
Biber, Douglas. 1998. “Dimensions of Variation among 18th Century Speech-based and
Written Registers”. Paper presented at the ICAME Conference, Belfast (Northern
Ireland), 20–24 May 1998.
Biber, Douglas, Edward Finegan & Dwight Atkinson. 1994. “ARCHER and its Challenges:
Compiling and Exploring a Representative Corpus of Historical English Registers”.
Creating and Using English Language Corpora: Papers from the Fourteenth International
Conference on English Language Research on Computerized Corpora, Zürich 1993 ed. by
Udo Fries, Gunnel Tottie & Peter Schneider, 1–13. Amsterdam & Atlanta, GA: Rodopi.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999.
Longman Grammar of Spoken and Written English. Harlow, Essex: Pearson Education
Limited.
Bolinger, Dwight. 1980a. Language. The Loaded Weapon. London: Longman.
Bolinger, Dwight. 1980b. “Wanna and the Gradience of Auxiliaries”. Wege zur Universalien-
forschung: Sprachwissenschaftliche Beiträge zum 60. Geburtstag von Hansjakob Seiler ed.
by G. Brettschneider & Christian Lehmann, 292–299. Tübingen: Narr.
Burchfield, Robert W., ed. 1996 [1926]. The New Fowler’s Modern English Usage. Oxford:
Clarendon.
154 Manfred Krug

Bybee, Joan L., William Pagliuca & Revere D. Perkins. 1994. The Evolution of Grammar:
Tense, Aspect, and Modality in the Languages of the World. Chicago: University of
Chicago Press.
Croft, William. 2001. Radical Construction Grammar: Syntactic Theory in Typological
Perspective. Oxford: Oxford University Press.
Denison, David. 1993. English Historical Syntax: Verbal Constructions. London & New York:
Longman.
English Dialect Dictionary. 1961 [1905]. Ed. by Joseph Wright. Oxford: Oxford University
Press.
Fillmore, Charles. 1985. “Syntactic Intrusion and the Notion of Grammatical Construction”.
Proceedings of the 11th Annual Meeting of the Berkeley Linguistics Society, 73–86.
Fischer, Olga & Frederike van der Leek. 1983. “The Demise of the Old English Impersonal
Construction”. Journal of Linguistics 19.337–368.
Gaaf, Willem van der. 1904. The Transition from the Impersonal to the Personal Construction
in Middle English. (= Anglistische Forschungen, 14.) Heidelberg: Carl Winter.
Grice, Herbert Paul. 1989 [1975]. “Logic and Conversation”. Studies in the Way of Words
22–40. Cambridge, Mass.: Harvard University Press.
Harris, Alice & Lyle Campbell. 1995. Historical Syntax in Cross-linguistic Perspective.
Cambridge: Cambridge University Press.
Heine, Bernd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: Oxford
University Press.
Heine, Bernd, Ulrike Claudi & Friederike Hünnemeyer. 1991. Grammaticalization: A
Conceptual Framework. Chicago: University of Chicago Press.
Hopper, Paul J. & Elizabeth C. Traugott. 1993. Grammaticalization. Cambridge: Cambridge
University Press.
Huddleston, Rodney. 1984. Introduction to the Grammar of English. Cambridge: Cambridge
University Press.
Hundt, Marianne & Christian Mair. 1999. “‘Agile’ and ‘Uptight’ Genres: the Corpus-based
Approach to Language Change in Progress”. International Journal of Corpus Linguistics
4.221–242.
Hundt, Marianne, Andrea Sand & Rainer Siemund. 1998. Manual of Information to Accom-
pany the Freiburg-LOB Corpus of British English. Freiburg: English Department,
University of Freiburg.
Jespersen, Otto. 1927 [repr. 1961]. A Modern English Grammar on Historical Principles. Part
III: Syntax (Second Volume). Copenhagen: Ejnar Munksgaard / Heidelberg: Carl Winter.
Jespersen, Otto. 1941. Efficiency in Linguistic Change. Copenhagen: Ejnar Munksgaard.
Johansson, Stig, Geoffrey N. Leech & Helen Goodluck. 1978. Manual of Information to
Accompany the Lancaster-Oslo/Bergen Corpus of British English, for Use with Digital
Computers. Oslo: Department of English, University of Oslo.
Kahlas-Tarkka, Leena. 1997. “thing in English: A Case of Grammaticalization”. Language
History and Linguistic Modelling: A Festschrift for Jacek Fisiak on his 60th Birthday ed. by
Raymond Hickey & Stanislaw Puppel, vol. I, 281–291. Berlin & New York: Mouton de
Gruyter.
Kirchner, G. 1940. “‘To Want’ as an Auxiliary of Modality”. English Studies 22.129–136.
</TARGET "kru">

A path to volitional modality 155

Krug, Manfred. 2000. Emerging English Modals: A Corpus-based Study of Grammaticalization.

Berlin & New York: Mouton de Gruyter.
Kytö, Merja. 1991. Manual to the Diachronic Part of the Helsinki Corpus of English Texts:
Coding Conventions and Lists of Source Texts. Helsinki: Department of English, Universi-
ty of Helsinki.
Labov, William. 1994. Principles of Linguistic Change. Vol. I: Internal Factors. Oxford:
Blackwell.
Labov, William. 2001. Principles of Linguistic Change. Vol. II: Social Factors. Oxford:
Blackwell.
López-Couso, María José & Belén Méndez-Naya. 1997. “What is Really Meant by Imperson-
al? On Impersonal and Related Terms”. Atlantis 19.185–192.
Mustanoja, Tauno F. 1960. A Middle English Syntax. Part I: Parts of Speech. (= Mémoires de
la Société Néophilologique de Helsinki, 23.) Helsinki: Société Néophilologique.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S. C. Weiner. Oxford: Oxford University Press.
Palmer, Frank R. 1986. Mood and Modality. Cambridge: Cambridge University Press.
Palmer, Frank R. 1989 [1979]. Modality and the English Modals. London: Longman.
Quine, Willard Van Orman. 1960. Word and Object. Cambridge, Mass.: MIT Press.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive
Grammar of the English Language. London & New York: Longman.
Searle, John. 1976. “A Classification of Illocutionary Acts”. Language in Society 5.1–23.
Söderlind, Johannes. 1951–1958. Verb Syntax in John Dryden’s Prose. 2 vols. Uppsala:
Lundequist.
Sweetser, Eve E. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of
Semantic Structure. Cambridge: Cambridge University Press.
Traugott, Elizabeth Closs. 1989. “On the Rise of Epistemic Meanings in English: An Example
of Subjectification in Semantic Change”. Language 65.31–55.
Traugott, Elizabeth Closs. 1996. “Semantic Change: An Overview”. Glot International 2.3–7.
Traugott, Elizabeth Closs. Forthcoming. “Constructions in Grammaticalization”. Handbook
of Historical Linguistics ed. by Brian D. Joseph & Richard D. Janda. Oxford: Blackwell.
Visser, Frederikus Th. 1963–1973. An Historical Syntax of the English Language. 3 parts in 4
vols. Leiden: E.J. Brill.
<LINK "len-n*">

<TARGET "len" DOCINFO

AUTHOR "Ursula Lenker"

TITLE "Is it, stylewise or otherwise, wise to use -wise?"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Is it, stylewise or otherwise,

wise to use -wise ?
Domain adverbials
and the history of English -wise*

Ursula Lenker
University of Munich

1. Introduction

Considering that very few languages of the world exhibit a special morpholo-
gical device for sentence adverbials (Ramat & Ricca 1998: 203), it is remarkable
that English has recently developed a suffix for that matter, namely the suffix
-wise with the meaning ‘as regards, in respect of ’ or ‘as far as [the base] is
concerned’.1
(1) “We are both saving some money”, she says, “though there is no hope at
the moment of being able to rent or buy. Nothing is certain. Jobwise, I
don’t know where I will be in a year’s time.”
(The Independent, May 9, 1993)
(2) They begin (…) by disputing the idea that their fans are bashful under-
achievers who spend their lives on the net (…).“It’s too easy to put a
label on us. We get all sorts, agewise and professionwise. I see quite good-
looking people who’ve got girlfriends”, maintains Colburn. “I don’t look
out and see a bunch of geeky, speccy shy people.”
(The Guardian, July 16, 1999)

According to Houghton (1968) and, independently, the OED, these denominal

adverbs in -wise first appear in “colloquial American English” in the 1940s
(OED s.v. wise n.1, [Link]; see below (33)). As early as the 1950s and 1960s,
however, sentence adverbials in -wise gain a much wider currency and even
become fashionable.
158 Ursula Lenker

The evidence presented here suggests that the new sense of -wise has in a short
time made considerable progress toward establishing itself as a generally
accepted part of the language. (…) No one knows, of course, whether it will
become more widely used and accepted in the future or whether it is, as some
believe, a fad that will soon run its course and pass away. (Houghton 1968: 213)

At that time, not only some linguists and style critics considered sentence
adverbials in -wise a “gimmick” or “fad” and, accordingly, users of -wise were
often harshly criticized for being “trendy” speakers who are “insensitive to
language” (cf. Houghton 1968: 214). Some style critics even wanted to “outlaw
it from decent American usage” (Follett 1966: 361).
Yet in spite of the criticism and stylistic warnings of prescriptivists, sentence
adverbials in -wise have become an increasingly accepted part of the English
language. In 1985, Quirk et al. (p. 568) establish them as markers of “viewpoint
subjuncts”, though they still regard them as “more freely productive in AmE
than in BrE” and assert that their use is considered “informal” and that “many
people object to these formations” (1985: 568, 1557).
British Good Style Guides of the 1990s do not recommend the use of -wise
for more formal styles, but they do not really object to it either.
The habit began in America and continues to be commoner in AmE than in
other forms of English. Fastidious speakers treat it with mild disdain, or with
a shrug of the shoulders as if to say that its use in this way is inevitable, painful
or too clever by half though it is. (Burchfield 1996: 852)
Admittedly, -wise is overused trendy jargon, but that doesn’t mean it is not
useful sometimes: careerwise is much quicker than ‘in relation to my career’,
moneywise more direct than ‘as far as money is concerned’ (…) If you use -wise
carelessly it becomes a trendy linguistic gimmick, but occasionally -wise added
to a word can say something quickly and effectively, as long as you remember
it will make some people wince. New -wise words are probably not acceptable
yet in serious writing. (Howard 1993: 411)

By stressing that the adverbials in -wise are “inevitable” (Burchfield), “useful”,

“quicker” or “effective” (Howard), these authors now emphasize the functional
values of -wise as a morphological marker of sentence adverbials.
This paper also adopts a basically functional approach: after an examination
of the distribution and productivity of sentence adverbial -wise in today’s
Englishes (cf. 2), the paper concentrates on the syntactic and functional
properties of these new coinages in Present-day English (cf. 3). It then traces the
history of today’s nominal and adverbial functions from wise’s original use as an
 Domain adverbials and the history of English -wise 159

independent noun (cf. 4) and compares this development to the related

German forms Weise and the suffix -(er)weise (cf. 6). Since these diachronic and
contrastive analyses show that the new sentence adverbial function of English
-wise has to be kept apart from its older meanings, its emergence and diffusion
is compared to the fairly new sentence adverbials in German -mäßig (cf. 7) and
to English adverbs in -(c)ally. In a last step, the emergence of -wise as a suffix
marking sentence adverbials is connected to the history of viewpoint adverbials
in German and English in the nineteenth and twentieth centuries (cf. 8, 9).

2. Distribution and productivity of -wise in Present-day English

2.1 Data
The widespread use and productivity suggested in the statements quoted above
can be tested on data from computer-readable corpora documenting Present-
day English (PDE). The following data set of all tokens and types of wise-coin-
ages in the different corpora collected on the ICAME CD-ROM2 corroborates
the impression that the use of -wise is no longer restricted to American English.
A broad variety of wise-coinages is found in comparable numbers in all major
varieties of English — African, American, Australian, British, Indian, and New
Zealand English.3
(3) ICAME CD-ROM
a. Written English:
Australian English (ACE): elementwise
American English (Brown): balance-wise, marketwise, price-
wise
American English (Frown): progresswise
East African English (ICE_ea): bookwise, biologicalwise (2×),
contextwise, percent-wise,
scopewise
Indian English (Kolhapur): commoditywise (2×), costwise,
histrionics-wise, occupation-wise,
state-wise, storywise, sex-wise,
unit-wise
New Zealand English (WWC): security-wise, tenant-wise,
workwise
160 Ursula Lenker

b. Spoken English:
British English (LLC): educationwise
East African English (ICE_ea): healthwise, literaturewise,
moneywise, population-wise,
schoolwise, timewise
New Zealand (WSC): incomewise, pressurewise, sizewise,
staffwise, trainingwise,
unemploymentwise

Although the corpora comprising written British texts (LOB, FLOB) do not
contain any instances, wise-coinages are amply recorded in more recent British
newspapers of the 1990s, such as The Guardian, The Independent and The
Observer.4
(4) The Independent 1993; The Guardian 1999, 2000; The Observer 1999, 2000
The Independent 1993: jobwise (2×), matchwise, moneywise,
pricewise, textwise, timewise, workwise
The Guardian 1999, 2000: agewise, bedwise, bookwise, brainwise,
brandwise, dietwise, foodwise (2×),
footballwise, gloatwise, healthwise (3x),
indoorwise, joywise, pain-wise, presswise,
professionwise, teamwise, time-wise,
weatherwise
The Observer 1999, 2000: businesswise, healthwise (2×), moneywise,
plotwise (2×), policywise

2.2 Productivity
These data, which record the total number of instances found in the respective
corpora, allow the conclusion that -wise is a productive element in Present-day
English. Similar results are obtained from the British National Corpus (BNC),
which contains altogether 205 instances (137 types) of sentence adverbials in
-wise (Dalton-Puffer & Plag 2000: 236).
Even a quick look at the altogether 72 tokens of wise-formations listed
above reveals the considerable share of single instances of a particular word.
Out of the 56 types of wise-coinages in question, only 10 occur more than once,
namely biologicalwise (2),5 bookwise (2), commoditywise (2), foodwise (2), health-
wise (6), jobwise (2), moneywise (3), plotwise (2), timewise (3) and workwise (2).
 Domain adverbials and the history of English -wise 161

This distribution of types and tokens is — according to research done by

Baayen (1993) and Baayen & Renouf (1996) — tremendously important for the
question of the productivity of an element. Baayen & Renouf explain how the
number of words that occur only once in a given corpus (i.e. ‘hapax legomena’
with respect to the given corpus) correlates with the number of neologisms and
is therefore highly indicative of the productivity of a morphological element:6
[If] a word-formation pattern is unproductive, no rule is available for the
perception and production of novel forms. All existing forms will depend on
storage in the mental lexicon. Thus, unproductive morphological categories
will be characterized by a preponderance of high-frequency types, by low
numbers of low-frequency types, and by very few, if any, hapax legomena,
especially as the size of the corpus increases. Conversely, the availability of a
productive word-formation rule for a given affix in the mental lexicon guaran-
tees than even the lowest frequency complex words with that affix can be
produced and understood. Thus large numbers of hapaxes are a sure sign that
an affix is productive. (Baayen & Renouf 1996: 74)

From the large number of hapaxes in both the 100 million-word BNC (111 or
81% of the 137 types; see Dalton-Puffer & Plag 2000) and also in my very large
corpus (94% in the ICAME corpus of about 11 million words; 77% in the whole
corpus),7 we can therefore conclude that -wise is a productive suffix in Present-
day English, an accepted morphological device for marking sentence adverbials.
It certainly was not a “fad” that ran its course and “passed away” (cf. Houghton
1968: 213).

2.3 Research
This use of -wise has, however, only very rarely received the attention of
linguists. The three relevant studies on the topic were published in its early and
trendy years 1968 and 1969 (cf. Houghton 1968; Pulgram 1968; Rahn 1969). At
about the same time, Marchand (1969: 358) observes its increasing use and
comments on its functions and its possible future as a full-fledged suffix: “wise
is being used less and less as an independent word and may, as a semi-suffix,
one day come to reach the state of F[rench] -ment”. Other studies on word-
formation do not mention these wise-coinages meaning ‘as regards, in respect
of ’, except for Bauer (1983: 225) who lists -wise among the adverbial suffixes
without, however, discussing it.
More importantly, these studies mainly collect instances of wise-formations
and consider their morphological and, in most cases marginally, their stylistic
162 Ursula Lenker

properties. They completely disregard syntactic and functional considerations,

which are in my view essential for understanding the origin and spread of this
new function of -wise as a sentence adverbial.

3. Functional properties of -wise

3.1 Sentence position

The scope of these new adverbials in -wise is not the verb phrase, but the whole
clause or sentence, a fact which is reflected in the surface word order of the
relevant sentences, in which wise-adverbials mostly occur sentence-initially or
sentence-finally.
(5) Foodwise, Stanley has authentic northern tastes — his favourite is tripe.
(The Guardian, February 26, 2000)
(6) It was the only way I could keep alive, foodwise.
(The Guardian, February 4, 2000)

If sentence adverbials in -wise appear in the middle of the sentence, their

parenthetical character is marked by dashes (see example (16)) or commas.
(7) My (French) partner tells me she’s always being asked by her astounded
friends if, bedwise, I’m not a complete catastrophe. Fortunately, she’s an
accomplished liar. (The Guardian, June 13, 2000)

3.2 Sentence adverbials

More prototypical examples of sentence adverbials are the so-called ‘disjuncts’
— which in the terminology of Quirk et al. (1985: 478–653) are distinguished
from adjuncts (manner or time adverbs such as (to walk) slowly, (to come)
regularly) and conjuncts (connecting adverbs such as therefore, however).8
Disjuncts (Quirk et al. 1985: 612–631) express an evaluation of the speaker,
either with respect to the meaning of a sentence (content disjuncts) or with
respect to the form of the communication (style disjuncts). For speakers’
comments on or evaluations of the sentence (content disjuncts), cf.
(8) She wisely didn’t attempt to apologize.
(It was wise of her that she didn’t attempt to apologize.)
(9) The Yard’s wonder boy, appropriately, descends from the clouds.
 Domain adverbials and the history of English -wise 163

(It was appropriate that the Yard’s wonder boy …)

Speakers’ comments on the form of the communication can take the form of
style disjuncts, such as
(10) Frankly, I’m tired.
(Frankly speaking, I’m tired; Put frankly, (…); I’m frank when I say (…);
In all frankness, (…))

3.3 Domain or viewpoint adverbials

The new adverbial coinages in -wise investigated here are similar to style
disjuncts such as frankly in that they concern the form, or rather, the viewpoint
or perspective of communication. These wise-formations therefore belong to the
adverbial category of ‘viewpoint’ adverbials, a hitherto neglected category in
linguistic research (see Lenker in progress). In English, this subcategory of
sentence adverbials is — on adverbs formed from (Neo-)Latin roots9 —
marked by the suffix -(c)ally:
(11) Linguistically, this example is interesting.
Textwise, the characters are flat as traffic signs (…)
(The Independent, August 3, 1993)

Both style disjuncts and domain/viewpoint adverbials have a corresponding

participle clause with speaking, such as frankly speaking or linguistically speaking
(see also below, 8.3). They also share other syntactic properties: it is not
possible, for instance, to negate them. Unlike when they are used as adjuncts,
viewpoint adverbials are nongradable; hence they do not accept premodifi-
cation or comparison, a feature which also style disjuncts accept only in certain
restricted contexts (cf. Paraschkewoff 1976: 182–186; Quirk et al. 1985: 569).
domain/viewpoint adverbs style disjuncts
Linguistically speaking, …. Frankly/confidentially speaking, …
*Not linguistically (speaking), … *Not frankly/confidentially (speaking),
…
*Very linguistically (speaking), … *?Very frankly/confidentially (speaking),
…
Speakers employ adverbs such as linguistically or textwise because they want to
indicate that the proposition of the whole sentence or clause is only true in the
perspective chosen by the speaker, the given domain. Bellert therefore refers to
164 Ursula Lenker

these adverbs as “domain adverbs” (Bellert 1977: 347–348) and exemplifies the
fact that speakers “do not commit themselves to the truth of the proposition in
any other domain” via the help of sentences in which domain adverbials are
used contrastively (Bellert 1977: 348):
(12) Linguistically, this example is interesting, but logically it is not.
Logically, John is right, but morally he is wrong.

This inherent restrictive dimension of domain adverbials is reflected in

Bartsch’s term limitierende Adverbiale ‘limiting adverbials’ (Bartsch 1972).
Quirk et al. emphasize their semantics and refer to them as “viewpoint ad-
juncts” (1972) or, in 1985, as “viewpoint subjuncts”10 (Quirk et al. 1985: 568)
and give as typical examples:
(13) Architecturally, it is a magnificent conception.
(14) Morally, politically and economically, it is urgent that the government
should act more effectively on aid to developing countries.
(15) It could have been a serious defeat, not only militarily but psychologically
and politically.

Instances from my corpus show that adverbs in -(c)ally and -wise are not only
theoretically equivalent with respect to their semantics and syntax, but that they
are indeed used together in one phrase if speakers want to indicate that the
proposition of the sentence is only true in the given perspective.
(16) Peter White, chief executive elect of the new Bank of Ireland & Leicester,
therefore, must this weekend at once be an elated and troubled man.
Elated, because he — personally and businesswise — is the winner from
the planned £ 11bn marriage of Alliance & Leicester and Bank of Ireland.
(The Observer, May 30, 1999)
(17) Caroline Coyne (…) says the 12-month guarantee is critical. “What
attracted me is that you could try work, and if it didn’t work out you
could return to your benefit. This takes the pressure off, financially and
healthwise.” (The Guardian, September 8, 1999)

It has to be emphasized that this domain function is the only adverbial function
of -wise which is widely productive today. The next section on the etymology
and history of the form -wise shows that its new use as a domain adverbial has
to be kept apart from other meanings and productive patterns attested for its
history in English. Nowadays, most of these are generally regarded as obsolete
or archaic.11
 Domain adverbials and the history of English -wise 165

4. The history of PDE -wise

4.1 Homonym: Adjective compound in -wise

In a first step, our formations have to be kept apart from their homonyms,
adjective compounds in -wise such as streetwise or pennywise. These go back to
a different Germanic root, OE wı̄s ‘prudent’ (OED s.v. wise a.), and are quite
productive in today’s Englishes. Examples of adjective compounds occurring in
my corpus are all-wise, budget-wise, industry-wise, pennywise, poundwise,
streetwise, weather-wise and worldly-wise. Weatherwise is an interesting forma-
tion because it occurs in both uses, as a nominalized adjective compound and
as a viewpoint adverbial:
(18) The day of the party dawned with that odd summer haze which told the
weather-wise it would grow bright and warm. (ICAME, Frown, p14_99)
(19) Of course, the one time in 1987 we did have a bit of excitement, weather-
wise, Michael Fish got it wrong. For which we love him all the more.
(The Guardian, February 11, 2000)

This ambiguity is exploited in book titles such as Weatherwise: The Techniques

of Weather Study or Babywise and Drugswise (guides for parents). They are
probably deliberately ambiguous, i.e. ‘wise about the weather/babies/drugs’
(adjective-compound) or ‘as regards the weather/babies/drugs’ (viewpoint
adverbial). The same is true for tradenames such as Netwise or Webwise.

4.2 OE wı̄se ‘manner, fashion; cause’

The root OE wı̄se ‘manner, fashion; cause’ is attested in most North and West
Germanic languages (cf. OHG wîsa, MHG wîsa; OFris wîs, OS wîsa, ON vísa; cf.
OE witan) and has survived in several functions in German Weise, Dutch -wijs
and Norwegian, Swedish and Danish -vis (OED s.v. wise n.1 II; Paraschkewoff
1976: 169–170). In today’s Englishes, it occurs in four different forms (for the
following account and the examples, see OED s.v. wise n.1; Marchand
1969: 357–358).12
a. Lexical noun: wise has retained its status as an originally independent lexical
noun only in archaic prepositional phrases such as in no wise, in like wise, on
this wise or in gentle wise.
b. Lexicalized items (likewise, otherwise): originally transparent formations of
the noun modified by an adjective only survive as lexicalized items, in
166 Ursula Lenker

particular likewise and otherwise, and the nowadays less frequent anywise,
contrariwise, leastwise and nowise (cf. OED ‘survived as simple words’).13
In the older stages of English, adverbial expressions meaning ‘in such-
and-such a manner or way’ could be qualified by an adjective or by a noun,
both either with or without a governing preposition, usually on or of (cf.
below, ealde wisan / on ald wise ‘in the old way, manner’).
(20) Beowulf 1865 Ic Þa leode wat (…) fæste geworhte, æghwæs untæle ealde
wisan.
(21) a1300 Cursor M. 10948 Als lagh was Þan on ald wise.

On ōðre wı̄san therefore alternates with ōðre wı̄san ‘in another manner’ in
Old English. In the fourteenth century, the latter yields the later lexicalized
otherwise.
(22) c900 tr. Bede’s Hist. i. xxvii. (1890) 72 Ne meaht þu on oðre wisan biscop
halgian (…)
(23) 971 Blickl. Hom. 177 þe læs þe oðre wisan ænig man (…)

c1. Manner adjunct ‘manner, way’: apart from these archaic or lexicalized
usages, wise occurs in three distinct denominal functions (c1, c2, d).
Formations on noun + wise meaning ‘in the manner of ’ appear as manner
adjuncts, such as (go/walk) frogwise/crabwise, pilgrimwise. These nominal
patterns are attested in OE phrases such as on scipwisan ‘in the manner of
or like a ship’. The OED refers to them as ‘non-syntactical’ because they
lack a morphological marker of the genitive.
(24) a950 Guthlac ii. (Prose) 107 (…) seo yld com þæt hit sprecan mihte æfter
cnihtwisan.
(25) 1377 Langl. P. Pl. B. xix. 138 Kulleden hym on-crosse-wyse.

From the fourteenth century onwards, they appear without a preposition.

(26) 1398 Trevisa Barth. De P.R. ix. xxxi. (1495) 368 On holy Saterdaye newe
fyre is fette and thus [= incense] is putte therin crossewyse. (OED s.v.
crosswise adv.)
(27) 1591 Savile Tacitus, Hist. i. lv. 32 No man presumed to make any
solemne oration assembly-wise [L. in modum concionis].

Nowadays, these formations are, however, productive only to a limited

extent. In Present-day English, they are rivalled by derivations in -fashion
 Domain adverbials and the history of English -wise 167

(arrow-fashion, baby-fashion) and -style (schoolboy-style) (see Marchand

1969: 358; Quirk et al. 1985: 1557; Bauer 1993: 225; Dalton-Puffer & Plag,
2000: 236–241).
c2. Manner adjunct ‘in direction of’: alternatively, wise-adverbs used as manner
adjuncts can refer to the concept of ‘dimension’, meaning ‘in direction of ’.
They occur, often with the variant -ways, in particular in such highly
frequent words as (anti-/counter-)clockwise, edgewise/-ways, endwise/-ways,
lengthwise/-ways, sidewise/-ways, slantwise, widthwise/-ways (OED s.v. wise
n.1 4; Marchand 1969: 357; Quirk et al. 1985: 1557; Burchfield 1996: 851).
While Marchand maintains that the manner-adjunct type “is strong”
(1969: 358) in Modern English, Quirk et al. (1985: 438, 1557) refer to it as
a “less common” suffix and assert that it is only “limitedly productive”.
Dalton-Puffer & Plag (2000: 239–240) corroborate the latter view of only
limited productivity by an analysis of the tokens, types and hapaxes in the
BNC (see above, 2.2). In the 100 million-word BNC 509 tokens are count-
ed, but only 39 different types (21 hapaxes). This means that, in today’s
English, the formation of new manner adjuncts according to this pattern is
possible, but not very frequent.
d. Viewpoint adverbs: as has been shown above (2.2), the noun + wise pattern
is fairly productive with regard to our domain or viewpoint adverbs.
Although the OED unfortunately still lists this pattern in the entry for the
noun, it is obvious that we are dealing with genuinely new coinages here.
The OED avoids the topic by stating: “(i) Used in the same way but with the
sense: as regards, in respect of” (OED s.v. wise n.1 [Link]).
Yet adverbs such as jobwise or bedwise are indeed only formally identi-
cal to those attested before the 1940s in that they are nominal combinations
without a marker of the genitive and with ellipsis of the preposition. In its
new function, -wise has definitely ceased to be an independent noun, since
it can no longer be paraphrased by employment of the now archaic noun
wise such as ‘in the wise (way) of a clock = clockwise’. Paraphrases such as
*‘in the wise (way) of a job’ or *‘in the wise (way) of a bed’ are impossible;
-wise has — as was predicted by Marchand (1969: 358) — now reached the
status of a lexicalized suffix.
There is thus no direct semantic path between the different meanings of -wise
(cf. the meanings listed under a–c) and its employment as a viewpoint adverbial
(d). Since the origin of sentence adverbial -wise is therefore more complicated
than it may have seemed at first glance, the next sections will test different
168 Ursula Lenker

suggestions for its rise (cf. 5, 6) and will, in conclusion, offer a new (functional)
explanation for its emergence and diffusion (cf. 8, 9).

5. The origin of -wise (Foster 1968)

As I have noted above, literature on the origin of -wise as a viewpoint adverbial

is rather scarce. In one of the very few suggestions, Foster argues that
we see here an astonishing reversal of fate whereby an ‘archaism’ suddenly
becomes the latest fad. It may well be that the old usage had lingered in the
United States to a greater extent than in the English of Britain, but it can
scarcely be doubted that its blossoming forth in the twentieth century was due
to a more or less conscious transference of a German habit, whether on the
part of scholars or — more likely — of immigrants. (Foster 1968: 95)

Foster’s suggestion of the revival of an archaism is not convincing, however.

The diachronic analysis has shown that the new formations are both semanti-
cally and functionally different from the manner adjuncts attested from Old
English onwards. The new formations are so novel that the suggestion of a
“reversal of fate” has to be discarded. Foster obviously senses the problems
associated with his assumption and accordingly also — rather wordily —
suggests German loan influence.

6. German -weise

6.1 PDG Weise, -weise

This second suggestion will now be tested by a short account of the meanings
and nominal and suffixal uses and functions of German Weise and -weise. In
German, the noun Weise ‘manner, fashion; way’ has retained all the semantic
and syntactic possibilities documented for the history of its English cognate
(Paraschkewoff 1976; Kluge/Seebold 1995: 883).
Weise has survived as a common noun — sie (macht es) auf ihre (diese)
Weise ‘she does it her (this) way’ (see above 4.2a) — and can be freely employed
in all kinds of prepositional phrases with modifying adjectives — auf ange-
nehme/gemächliche Weise ‘in a pleasant/slow way’ (see above 4.2b).
 Domain adverbials and the history of English -wise 169

There are also denominal compound adverbs in -weise which function as

manner adjuncts (see above 4.2c), such as kreuzweise (‘crosswise’) or schrittweise
(‘stepwise’).

6.2 PDG adjective + -erweise

In contrast to English, however, a specific adjectival pattern — adjective +
-erweise — is widely productive in today’s German (cf. glücklicherweise ‘luckily’
or interessanterweise ‘interestingly’). This formation is particularly interesting
because German does commonly not mark adverbs morphologically but employs
the unmarked form of the adjective for the adverbial function. Yet in the case of
these content disjuncts (see above 3.2), the suffix -erweise is obligatory.
(28) He answered me cleverly.
GER Er antwortete mir klug. (manner adjunct)
(29) Cleverly, he answered me.
It was clever of him to answer me.
GER Er antwortete mir klugerweise. (content disjunct)

Cleverly, when used as a content disjunct (It was clever of him to answer me…),
can only be rendered by klugerweise.14 Etymologically cognate suffixes are also
productive in other Germanic languages, for example Dutch -erwijs, Swedish
and Danish -vis (Ramat & Ricca 1998: 203–206). These forms seem to constitute
one of the extremely few morphological devices marking sentence adverbials
(Ramat & Ricca 1998: 204) and are therefore related to English -wise. They do
not mark domain adverbials, however, but content disjuncts.
In sum, if adjective + -erweise is the German pattern Foster is thinking of,
it is definitely not a good choice. -erweise cannot be a model for the suggested
loan -wise, because the coinages are not at all similar, neither morphologically
nor semantically:
english -wise german -erweise
denominal deadjectival
domain adverb content disjunct
170 Ursula Lenker

7. German -mäßig

There is, however, a fairly new German suffix which is functionally equivalent
to the use of -wise in its functions as a domain or viewpoint adverb, namely
-mäßig (examples are taken from Inghult 1975).
(30) (…) eine altersmäßig schlechte Zusammensetzung (=‘agewise, an un-
favourable distribution’)
(31) Nordrhein-Westfalen — räumlich und bevölkerungsmäßig vergleichbar
mit der DDR (…) (=‘Nordrhein-Westfalen — sizewise and population-
wise comparable to the GDR’)

This function of -mäßig appeared in the middle of the nineteenth century. Just
like -wise, it spread considerably in the middle of the twentieth century and was
(and still is) harshly criticized; a German style critic even regards the use of
-mäßig as “sprachmäßig saumäßig” (‘linguistically like a pig’; cf. “Sprachmäßig
saumäßig”, Kölner Stadtanzeiger August 4/5, 1978, cited from Welte 1996: 232).
This fairly recent emergence and fast diffusion of new morphological
patterns marking the function ‘domain adverbials’ in both English and German
support my idea that functional considerations are essential for an analysis of
-wise, and lead to my last major point, the history of domain or viewpoint
adverbials in English and German.

8. The history of viewpoint adverbials

8.1 Viewpoint adverbials in the history of German

The most interesting fact about this subgroup of adverbials is that they are
comparatively young. In both English and German, the category ‘viewpoint
adverbial’ is not attested until the nineteenth century.
In his study of German -mäßig, Inghult shows that the first instances of
-mäßig in the limiting function appear as late as the nineteenth century. Still
more interesting, however, is the fact that other German suffixes marking this
function are not recorded before 1800 either:
EINSCH[ränkung]-Derivate auf -mässig sind also eine relativ neue Erschein-
ung in der deutschen Sprache (…) [Sie] kommen vereinzelt schon im 19. Jh.
vor, und es sieht aus, als ob der Gebrauch von Adjektivderivaten mit limitativer
Funktion im Deutschen überhaupt erst zu dieser Zeit entstanden sei. Jedenfalls
 Domain adverbials and the history of English -wise 171

sind keine Beispiele anderer Suffixtypen vor 1800 zu belegen. Aus dem 19. Jh.
dagegen sind u.a. folgenden Bildungen auf -isch, -lich, -al, -är und -iv zu
verzeichnen. (Inghult 1975: 152; emphasis supplied)
[Limiting derivations in -mässig are a comparatively recent phenomenon in
German (…) There are some instances in the nineteenth century, and it seems
as if the use of adjective-derivations with limiting function was indeed only
employed from that time onwards. There are (…) no examples of other
suffixes attested before 1800. In the nineteenth century, however, we find
formations in -isch, -lich, -al, -är and -iv.]

8.2 Viewpoint adverbials in the history of English

Inghult’s data for German match my findings for the history of domain
adverbials in English (for a more comprehensive account, see Lenker in
progress). An analysis of adverbs in -ly and their syntactic distribution under-
taken in the diachronic corpora on the ICAME CD-ROM (Helsinki, CEECS,
Lampeter, Innsbruck), which comprise texts up to the eighteenth century, yields
no unambiguous results for domain adverbials. This means that this adverbial
function must have indeed emerged after that period, in particular if we are not
talking about a few doubtful cases from religious discourse, but of a semantic
and syntactic pattern.15
For the corpus analysis, data were lifted from the best computer-readable
source for the nineteenth and early twentieth century, the OED (1989, 2nd ed.)
and the OED-online ([Link] Table 1 shows the results of the ‘Word’
search for -(c)ally (letters A to D). The first column lists the earliest attestation
of the adjective. This is essential because all of the adverbs in question go back
to Neo-Latin/Greek roots in -al or -ical (Latin -āl-em, -āl-is) and were mainly
borrowed in the EModE period (see OED s.v. -al suffix1; Marchand 1969:
238–244).16 The second column lists the earliest attestation of the adverb, the
third and essential column records the date for its first recorded use as a domain
adverbial.
The dates given in the third column clearly show that the domain usage
appears and is indeed already strong from the second half of the nineteenth
century onwards, irrespective of when the adjective or adverb had entered the
language.
The results are basically the same for the other letters. Among the earliest
adverbs used as domain adverbials, philosophically and politically are attested in
the first half of the nineteenth century (Table 2).
172 Ursula Lenker

Table 1. OED adverbs in -(c)ally (letters A–D)

Adjective Adverb Domain Adverbial
(-al, -ical)

academically 1610 1591 1879

analytically 1525/1601 1656 1879
architecturally 1762 1843 1876
arithmetically 1543 1571 1865
atmospherically 1664 1874 1871
botanically 1658 1757 1849, 1870
choreographically 1893 1911 1959
chronologically 1614 1691 1881
clinically 1780 1862 1876
crumenically – 1825 1825 (nonce-word)
culturally 1868 1889 1906
developmentally 1849 1849 1949
diagnostically 1625 1657 1891
diplomatically 1787 1836 1877

Table 2. Other OED adverbs in -(c)ally

Adjective Adverb Domain adverbial
(-al, -ical)

philosophically 1500 1580 1825

politically 1420 1638 1841

8.3 The origin of the viewpoint or domain use

This emergence of domain uses for these adverbs can be traced back to their
original employment as manner adjuncts. Before their independent use as
domain adverbials, the adverbs often surface in phrases such as to speak(e)
(more) ~, ~ speaking, ~ considered, ~ studied, ~ viewed (see above, 3.3). These
phrases, which must have served as the bases or models for the use of the
adverbs as domain adverbials, are attested in large numbers also only from the
end of the eighteenth century onwards.
For a full example lifted from the OED by means of a combined ‘Word’ and
‘Quotation’ search, see botanically:
(32) OED: botanically
a. OED s.v. botanically adv. (botanical adj., 1658; botany n., 1696)
‘In a botanical manner; in relation to botany; according to the prin-
ciples or technical language of botany.’
 Domain adverbials and the history of English -wise 173

1757 Da Costa in Phil. Trans. L. 229 note, Scheuchzer has arranged

the fossile plants botanically.
1793 W. Curtis Bot. Mag. VI. 215 In its improved, or to speak more
botanically, in its monstrous state.
1848 C. A. Johns Week at Lizard 291 It is botanically distinguished
from the other Heaths, by its anthers.
1870 Yeats Nat. Hist. Comm. 102 Botanically, this is the region of
palms.
b. OED s.v. myrtle n.
1849 Rural Cycl. III. 538/1 *Myrtle Bilberry, — botanically
Vaccinium Myrtillus.
c. OED s.v. herald n.
1894 H. Drummond Ascent Man 295 The Flower, botanically, is the
herald of the Fruit.

In the earliest quotation (1757), botanically is a genuine manner adjunct (to

arrange botanically).17 In 1793, it is attested in the phrase to speak more botani-
cally, which serves as the basis for its use as a viewpoint or domain adverbial
(see the quotations from 1870, 1849, 1894).18

8.4 Scientific language in the nineteenth century

This late emergence of the category of viewpoint adverbials in the nineteenth
century corresponds to certain socio-historical developments in that period
which may be (very roughly) summarized as follows. It may be suggested that
the need for a linguistic pattern of marking perspective only arose with the
diversification of perspectives after the Middle Ages or even after the Renais-
sance, at a time when the theological or religious perspective had ceased to
dominate (scientific) thinking.
The number of conceivable perspectives on a given subject increased
rapidly — and thus gained additional linguistic relevance19 — with the advent
of the new sciences, and especially with the increasing diversification of the
empirical sciences into discrete individual disciplines in that period. In a
contemporary nineteenth-century source, Trench criticizes the influx of an
“army of purely technical words” which occur in treatises on “chemistry or
electricity, or on some other of the sciences which hardly or not at all existed
half a century ago” (1860: 57–58). At that time, the many new suffixes formed
from Neo-Latin or Greek stems indicate that science in the nineteenth century
was particularly preoccupied with measuring and precise classification — cf.
174 Ursula Lenker

e.g. -graph, -meter, -ite, -ide or -ine (Bailey 1996: 145–151; Görlach
1999: 111–114). Similarly, the new domain adverbials in -(c)ally are also formed
from Neo-Latin/Greek roots (see above, 8.2) and a large number of them refer
to the domain of a particular science (architecturally, arithmetically, atmospheri-
cally, botanically) or its methods (chronologically, clinically, diagnostically).
Biber & Finegan (1997:260ff.) track overall changes in the patterns of register
variation over the past four centuries, and document a general trend towards a
less ‘involved’, more ‘informational’ style in specialist written registers such as
medical prose. The new domain adverbials in -(c)ally are an excellent vehicle for
that new informational pattern. They indicate the speaker perspective (most
often sentence-initially) without having to name the speaker directly and thus
help to avoid the use of a personal pronoun. More recently, Biber & Clark
(2002) have also shown that over the past 100 years medical prose has devel-
oped a much more compressed style of presentation, which is at the same time
less explicit in the expression of relational meaning. Domain adverbials
definitely satisfy these new stylic demands because they are an extremely
condensed and therefore quick and efficient means of stating the perspective
chosen for the proposition, a property that is indispensable for scientific texts,
but also quite convenient in other contexts.

9. Pattern transfer: -wise, likewise and otherwise

9.1 -cally and -wise: Morphology

I would therefore like to suggest that this late emergence of the function of
‘domain adverbial’ in the nineteenth century was the beginning of the creation
and spread of the coinages in German -mäßig and, later, English -wise. The
linguistic efficiency of viewpoint adverbials is obvious and we cannot really
imagine twentieth-century academic or scientific writing without them. Yet, the
English coinages in -(c)ally have one apparent inconvenience: they can only be
derived from Latin (or Greek) adjectival bases in -al or -ical. The suffix -wise, on
the other hand, can be combined with all nominal roots, irrespective of their
etymological origin, and is therefore an ideal candidate to fill this gap in word-
formation.20
 Domain adverbials and the history of English -wise 175

9.2 Earliest attestations

The earliest attestations of -wise support this view. Although the OED (s.v. wise
n.1 [Link]) maintains that -wise with the sense ‘as regards, in respect of ’ is
“colloq. (orig. U. S.)”, the quotations given for the first attestations show that
the decisive factors for its use are neither ‘medium’ (spoken vs. written) nor
‘attitude’ (formal, informal, etc.), but ‘field of discourse’ (for these types of
variation see Quirk et al. 1985: 15–26).21
(33) 1942 E. R. Allen in J. J. Mattiello Protective & Decorative Coatings II. viii.
252 It should be noted that there are two types of hydrogen atoms
positionwise.
1948 Sat. Rev. 6 Mar. 16/3 Plotwise, it offers little more or little less of
what-happens-next interest than may be found [etc.]

The co- and contexts of these examples refer to the fields of discourse of science
or arts. The same is true for most of my corpora-examples: they may be more
frequent in the spoken medium (see above, (3b)), but they almost exclusively
relate to these technical registers.
(34) John Guillebaud, (…) and author of The Pill (…) confirms that the
pattern of having 13 breaks a year (…) is “amazingly safe, healthwise”.
(The Guardian, June 17, 2000)
(35) (…) as far as crayfish are concerned are th (…) you know that the ideal
unit ummm sizewise (…) (ICAME CD-ROM, Wellington Corpus of
Spoken New Zealand English)

In German, this relation to the field of science and technology is even more
obvious, because in nonstandard German -mäßig is rivalled by the new suffix
-technisch ‘technically’ (cf. Arbeitsmäßig/arbeitstechnisch war das ein blöder Tag
‘Workwise, it was a horrible day’).
For English, this idea of a transfer of the functional pattern from the
scientific discourse is also corroborated by one of the American native speaker
informants in Rahn’s (1969) study who says about wise-creations that he has
“heard them usually from people who had studied engineering” (Rahn
1969: 234). All the early studies from 1968 and 1969 furthermore mention the
fact that -wise is often used in the languages of business, trade and industry, all
of which are technical varieties more or less modelled on the field of discourse
of scientific language.
176 Ursula Lenker

9.3 The morphological pattern

Once we accept a transfer from the field of scientific discourse, we have to
consider why the suffix -wise was chosen for filling the obvious gap in word-
formation. The most important factor must have been that -wise had become
archaic in its original usages as a lexical noun and infrequent or only limitedly
productive as a manner adverb (see above, 4.2 a, b, c). The form could therefore
be functionally reshifted.
The missing link between the still existent uses of -wise and the new domain
use must have been highly frequent lexicalized words in -wise, in particular
likewise and otherwise (see above, 4.2b).
Likewise, for instance, is functionally similar to domain adverbials in -wise
when it is used as a conjunct ‘similarly; also’ (OED s.v. likewise 2, 3). In this
sense, it can be paraphrased by ‘seen in the same perspective’.
The more likely model, however, is otherwise, which often serves as a domain
adverbial of a more or less vague kind. A sentence starting with otherwise allows
speakers to open up a new perspective without necessarily having to refer to
their change of perspective. The OED describes this meaning and function by
‘[i]n other respects; with regard to other points’ (OED s.v. otherwise adv. 3), a
definition which shows the tight link between otherwise and the new wise-ad-
verbials (OED definition: ‘as regards, in respect of ’; s.v. wise n.1 [Link]).
Otherwise is accordingly often found in one phrase together with a domain
adverb. Consider, as one of many examples from the OED, the phrase etymolog-
ically or otherwise in the definition for nylon (OED s.v. nylon).
(36) The word is a generic word coined by the du Pont Co. It is not a regis-
tered name or trademark. We wish to emphasize the following additional
points: First, that the letters n-y-l-o-n have absolutely no significance,
etymologically or otherwise.

This close connection with otherwise is actually also employed as a stylistic device
by such a language-conscious writer as Arundhati Roy (1997: 83) when she says
(37) It didn’t make any sense at all. Weatherwise or otherwise.

10. Conclusion

The new coinages in -wise marking sentence adverbials are thus not archaisms
which reappeared and became trendy in American usage as a “linguistic
<DEST "len-n*">

Domain adverbials and the history of English -wise 177

gimmick” or “fad”. Neither are they borrowings from German. They are
English/American innovations, and were coined for a functional reason: they
are employed to form derivations for the comparatively new category of domain
adverbials from non-Latin roots which do not allow the element -(c)ally.22 The
form -wise proved to be a good candidate for fulfilling this function because its
old meanings and functions had become archaic and because the lexicalized
likewise and otherwise provided the morphological pattern needed.

Notes

* I would like to thank the anonymous commentators, the editors of this volume, Christiane
Dalton-Puffer, Nicholas Jacob-Flynn (a devoted user of sentence adverbial -wise!), Andreas
Mahler and in particular Walter Hofstetter for many helpful comments on earlier versions
of this paper.
1. The examples in this paper are taken from either the ICAME CD-ROM or from a corpus
of more recent British newspaper publications, The Independent on CD-ROM for 1993 and
The Guardian and The Observer for 1999 and 2000 (see also (3) and (4)).
2. The data are lifted from the ICAME CD-ROM; for further information, see
[Link]
3. Only wise-coinages with the meaning ‘as regards, as far as x is concerned’, i.e. adverbs
functioning as sentence adverbials, are listed. For other uses of the suffix -wise (adjective
compounds such as streetwise, manner adverbs such as crosswise or lexicalized items such as
likewise and otherwise), see below 4.1 and 4.2.
4. The data are lifted from The GuardianUnlimited website ([Link]
5. This peculiar deadjectival form found in one source in the corpus of East African English
should probably be regarded as a true nonce-formation.
6. I owe this reference to Christiane Dalton-Puffer; see also Dalton-Puffer & Plag (2000).
7. There are 35 hapaxes among the 37 tokens in the about 11 million words of the syn-
chronic corpora of the ICAME CD-ROM which were searched for the present study (about
one million words per corpus). Unfortunately, there was no way to ascertain the total
number of words in the British newspaper corpora; for the question of productivity,
however, only the relation ‘large corpus — high number of hapaxes’ (81% BNC / 94%
ICAME / 77% total corpus) is important.
8. In the new Longman Grammar (Biber et al. 1999), adjuncts, disjuncts and conjuncts are
called, respectively, ‘circumstance’, ‘stance’ and ‘linking’ adverbials (Biber et al. 1999:762 and
elsewhere). The examples in this section are mainly taken from Bellert (1977), Quirk et al.
(1985) and Biber et al. (1999).
9. For the restriction of these coinages to (Neo-)Latin and Greek bases, see below 8.2.
178 Ursula Lenker

10. The terms ‘adjunct’ and ‘subjunct’ (instead of ‘disjunct’) seem to suggest that viewpoint
adverbs are not regarded as sentence adverbials by Quirk et al. (1985). Apart from this
source, however, they are generally described as one of the subclasses of sentence adverbials.
The whole category of ‘subjunct’ has been repeatedly criticized; for a recent assessment, see
Valera (1998: 267–270). Biber et al. (1999) return to a tripartite system (see fn. 8).
11. For a comprehensive comparative investigation of the distribution and productivity of
the different patterns, see Dalton-Puffer & Plag (2000).
12. West-Germanic wîsa was borrowed by the Romance languages and survives as French guise
or Italian guisa. In medieval Spanish, guisa competed with the adverbial formatives cosa and
mente, which in the fourteenth century ousted the other suffixes (cf. Paraschkewoff 1976:170).
13. All the examples in the diachronic sections are taken from the quotations in the OED.
14. Paraschkewoff therefore even ceases to regard -erweise as an inflected form: he analyses
-er- as a “Fugenelement” and accordingly wants to move -erweise to the category of modals
(1976: 193–196, 211). This transfer is not necessary, however, when we employ the term
‘disjunct’, which intrinsically includes a certain modality.
15. The few doubtful cases belong to the field of discourse of religion or theology (spiritually,
ghostly, etc.); for the relation of this field of discourse to scientific language, see below 8.4.
16. In Old French and hence in early English, -ālem became -el (cf. mortālem and mortel); in
English (and to some extent in French), this was refashioned after Latin -al. The number of these
adjectives increased immensely in medieval and Early Modern Latin and hence in Early Modern
English, where the pattern is finally also extended to Greek roots (see OED s.v. -al suffix1).
17. In a semantically more detailed analysis, Quirk et al. (1985: 563–564) refer to them as
adjuncts of respect, e.g. she’s advising them legally / with respect to law. These predicational
‘adjuncts of respect’ are the natural bridge between the different uses in question because
they fulfil the same role on the phrase level as viewpoint adverbials do on the sentence level.
18. The work of the lexicographers of the OED proved to be extremely reliable: the earliest use
as a domain adverbial in quotations is usually also listed in the entry for the respective adverb.
19. For the properties of scientific discourse in the EModE period, see e.g. Nate
(2001: 141–200).
20. The derivation of adverbs by an originally nominal suffix is not a problem; cf. the
comparable case of Romance -ment, -mente which goes back to Latin mens, mentis.
21. I will here follow the terminology of Quirk et al. (1985). For the differences and
similarities to other notational terms such as ‘register’ or ‘style’, see Lipka (1992: 9–26).
22. Whether this is a grammatical function and therefore a case of grammaticalization is
doubtful because of the awkward status of adverbs as a grammatical class (for a fuller
treatment of the question, see Lenker in progress).
 Domain adverbials and the history of English -wise 179

References

Baayen, Harald. 1993. “On Frequency, Transparency, and Productivity”. Yearbook of

Morphology 1992 ed. by Geert Booij & Jaap van Marle, 181–208. Dordrecht: Kluwer.
Baayen, Harald & Antoinette Renouf. 1996. “Chronicling The Times: Productive Lexical
Innovations in an English Newspaper”. Language 72.69–96.
Bailey, Richard W. 1996. Nineteenth-Century English. Ann Arbor, Mich.: The University of
Michigan Press.
Bartsch, Renate. 1972. Adverbialsemantik. Die Konstitution logisch-semantischer Repräsen-
tationen von Adverbialkonstruktionen. Frankfurt: Athenäum.
Bauer, Laurie. 1983. English Word-formation. Cambridge: Cambridge University Press.
Bellert, Irena. 1977. “On Semantic and Distributional Properties of Sentential Adverbs”.
Linguistic Inquiry 8.337–351.
Biber, Douglas & Edward Finegan. 1997. “Diachronic Relations among Speech-based and
Written Registers in English”. To Explain the Present: Studies in the Changing English
Language in Honour of Matti Rissanen ed. by Terttu Nevalainen & Leena Kahlas-Tarkka,
253–275. Helsinki: Société Néophilologique.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999.
Longman Grammar of Spoken and Written English. Harlow, Essex: Pearson Education
Limited.
Biber, Douglas & Victoria Clark. 2002. “Historical Shifts in Modification Patterns with
Complex Noun Phrase Structures: How Long Can You Go without a Verb?” English
Historical Syntax and Morphology. Selected Papers from 11 ICEHL, Santiago de Compos-
tela, 7–11 September 2000 ed. by Teresa Fanego, María José López-Couso & Javier Pérez-
Guerra, 43–65. Amsterdam & Philadelphia: John Benjamins.
Burchfield, Robert W. 1996 [3rd ed.]. Fowler’s Modern English Usage. Oxford: Clarendon
Press.
Dalton-Puffer, Christiane & Ingo Plag. 2000. “Categorywise, some Compound-type
Morphemes Seem to Be Rather Suffix-like: On the Status of -ful, -type and -wise in
Present Day English”. Folia Linguistica 34.225–244.
Follett, Wilson. 1966. Modern American Usage. A Guide ed. and completed by Jacques Brown
et al. London: Longman.
Foster, Brian. 1968. The Changing English Language. London: Macmillan.
Görlach, Manfred. 1999. English in Nineteenth-Century England. An Introduction. Cam-
bridge: Cambridge University Press.
Houghton, Donald E. 1968. “The Suffix -wise”. American Speech 43.209–215.
Howard, Godfrey. 1993. The Good English Guide. English Usage in the 1990s. London:
Macmillan.
Inghult, Göran. 1975. Die semantische Struktur desubstantivischer Bildungen auf -mässig. Eine
synchronisch-diachronische Studie. Stockholm: Almqvist & Wiksell.
Kluge, Friedrich. 1995 [23rd ed., rev. by Elmar Seebold]. Etymologisches Wörterbuch der
deutschen Sprache. Berlin & New York: Mouton de Gruyter.
Lenker, Ursula. In progress. Sentence Adverbials and (Speaker-)Perspective in the History of
English.
</TARGET "len">

180 Ursula Lenker

Lipka, Leonhard. 1992 [2nd ed.]. An Outline of English Lexicology. Tübingen: Niemeyer.
Marchand, Hans. 1969 [2nd rev. ed.]. The Categories and Types of Present-Day English Word-
formation. A Synchronic-Diachronic Approach. München: Beck.
Nate, Richard. 2001. Wissenschaft und Literatur in England der frühen Neuzeit. München:
Fink.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S. C. Weiner. Oxford: Oxford University Press.
OED Online. 2000– [3rd ed., in progress]. Ed. by John A. Simpson. Oxford: Oxford Universi-
ty Press. [Link]
Paraschkewoff, Boris. 1976. “Zur Entstehungs- und Entwicklungsgeschichte der Bildungen
auf -weise”. Beiträge zur Geschichte der deutschen Sprache und Literatur 97.165–211.
Pulgram, Ernst. 1968. “A Socio-linguistic View of Innovation: -ly and -wise”. Word
24.380–391.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1972. A Grammar of
Contemporary English. London: Longman.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive
Grammar of the English Language. London & New York: Longman.
Rahn, Walter. 1969. “Das Suffix -wise im heutigen amerikanischen und britischen Englisch”.
Literatur und Sprache der Vereinigten Staaten. Aufsätze zu Ehren von Hans Galinsky ed.
by Hans Helmcke, Klaus Lubbers & Renate Schmidt-v. Bardeleben, 228–241. Heidel-
berg: Winter.
Ramat, Paolo & Davide Ricca. 1998. “Sentence Adverbs in the Languages of Europe”.
Adverbial Constructions in the Languages of Europe ed. by Johan van der Auwera (with
Dónall P. Baoill), 187–275. Berlin & New York: Mouton de Gruyter.
Roy, Arundhati. 1997. God of Small Things. London: HarperCollins / Flamingo.
Trench, Richard Chevenix. 1860. On Some Deficiencies in our English Dictionaries (1857).
London: John W. Parker & Son.
Valera, Salvador. 1998. “On Subject-orientation in English ly-adverbs”. English Language and
Linguistics 2.263–282.
Welte, Werner. 1996 [2nd ed.]. Englische Morphologie und Wortbildung. Ein Arbeitsbuch mit
umfassender Bibliographie. Frankfurt: Lang.
<LINK "los-n*">

<TARGET "los" DOCINFO

AUTHOR "Bettelou Los"

TITLE "The loss of the indefinite pronoun man"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

The loss of the indefinite pronoun man

Syntactic change and information structure*

Bettelou Los
Free University of Amsterdam

1. Introduction

Rissanen (1997) presents a detailed discussion of the collapse of the Old English
system of indefinite marking with hwa ‘who’, hwilc ‘which’, hwæthwugu
‘something’, etc., and connects its loss to the generalization of the wh-pronouns
as expressions of the relative pronoun. Wh-pronouns already had two functions
— interrogative and indefinite — and an additional third use may have
compromised the communicative function of the system. The old system was
replaced by the some/any paradigm, with its clear contrast of assertive versus
non-assertive contexts, and by the use of one as a ‘proform’, the lexical expres-
sion of the head of the noun phrase — a development which Rissanen connects
to the levelling and loss of the nominal endings.
The loss of the indefinite pronoun man ‘one’ (cf. Swedish man, German
man, Dutch men) by the fifteenth century has proved more difficult to account
for, although various contributing factors have been identified. I will argue in
this paper that there are two important factors that appear to have been
overlooked: one is the competition between subjunctive that-clauses and
to-infinitives, which affected man in that it entailed competition between the
indefinite pronoun in such clauses and generic (or arbitrary) PRO. The result
was a decline in the occurrence of man in subclauses. There was also a decline
in main clauses due to the loss of verb-second in the course of the fifteenth
century, after which only subjects could be ‘unmarked themes’ in an informa-
tion-structural sense. The indefinite pronoun man/me(n) is unlikely to occur in
this position as it cannot provide an anaphoric link with previous material, and
its niche was increasingly taken over by the impersonal passive.
182 Bettelou Los

2. Previous proposals

The indefinite expression man develops from the noun man ‘man, human
being’ in the prehistory of Old English. In what follows I will refer to it as ‘the
indefinite pronoun’, although it formally wavers between a noun and a pro-
noun (but see van Bergen 2000a, b, c), as is typical of the ‘general’ nouns that
constitute an intermediate category of vague, generic, ‘light’ predications:
people, creature, thing, stuff (Halliday & Hasan 1976: 274). Man is first weakened
into men and me (although it is possible that these spellings reflect the loss of
the West-Saxon written standard rather than actual phonetic change in the
transition to Middle English) and is lost in the course of the fifteenth century.
Last mention in the OED of men is (1), and of me (2):
(1) Men muste putte hym self at the vpper syde of hym.
(1484, Fables of Æsop v. vii)
“One must put oneself at the upper side of him.”
(2) Thinges (…), Of whiche me may not be withoute.
(c. 1483, Caxton Dialogues 6/20)
“Things (…) which one cannot do without.”

The demise of man may have been hastened by the fact that the form may have
become too opaque because of the existence of similar forms with similar
meanings. Firstly, the plural of the full lexical form man, i.e. men, can only be
disambiguated from the ME form of indefinite man, i.e. men, by the singular or
plural ending of the finite verb — and number is not consistently marked on
ME verbs (this is especially true of the modal verbs, which are developing
invariant forms at this period). Secondly, the full lexical noun man ‘man’ was
acquiring some indefinite meanings, repeating to some extent the earlier
grammaticalization that had produced the OE indefinite pronoun in the first
place, although this new use of the full lexical noun man never became a proper
indefinite (Rissanen 1997: 521, reporting on Raumolin-Brunberg & Kahlas-
Tarkka 1997).1 However, interference from these other man/men forms is
unlikely to have been the primary cause of the loss of indefinite man. The only
form such interference could have taken would have been the rise of some other
form to take over the indefinite pronoun function. There is one likely candidate
for such a form, i.e. one — but it is clear that there was no straightforward
replacement of OE man with LME/EModE one, for two reasons.
First, there is the time-lag between the rise of indefinite one as an indefinite
pronoun and the loss of the older indefinite man, men, me, which suggests that
 The loss of the indefinite pronoun man 183

the items were not in competition at any one time (Rissanen 1967, 1997), so
that we cannot speak of a straightforward case of one replacing the other.
Second, such a one-to-one replacement is also unlikely in view of the
significant differences between the use of one as an indefinite pronoun in
Present-day English and the uses of surviving indefinites man, men in the other
Germanic languages. These differences concern register and inclusion or
exclusion of interlocutors (see Siewierska 1984: 238–54; Seoane-Posse 2000, and
references cited there). Impersonal constructions (with one, people, they, we or
you) are not viable alternatives for impersonal passives in Present-day English
(ibid.), unlike OE constructions with man, as we will argue below.
We will see in what follows that the linguistic niche of indefinite man was itself
largely lost, and that this may have been one of the primary causes of its demise.

3. Syntactic changes and information structure

3.1 Man in Ælfric

In order to establish the various functions of the indefinite pronoun man in Old
English we investigated its use in Ælfric. Ælfric’s works consist of, for the most part,
original, untranslated prose; though Ælfric draws heavily on various Latin and
Greek sources, he is a conscious stylist, as his own views on the translation process,
as set out in his preface to the translation of Genesis, testify, and his language
may be safely regarded as a good representative of authentic Old English.
Man occurs 430 times in Ælfric’s works. 72 instances occur in the second
conjunct of a pair of clauses coordinated by the conjunctions and ‘and’ or ac
‘but’, as in (3):
(3) and man þa deadan ne mihte, þe þær adydde wæron, (…)
and one the dead not could who there cast out were
nateshwon bebyrgean for heora mægenleaste, ·ÆHomM 4, 77Ò2
[Link] bury for their feebleness
“and because of their weakness people were completely unable to bury
the dead who had been cast out there”

There are so many problems surrounding such clauses that they for the mo-
ment have been excluded. One problem is that it is often impossible to decide
whether they are main or subclauses; the word order of OE and/ond or
ac-clauses varies and can be either subordinate (i.e. verb-late) or root (with
184 Bettelou Los

inversion of man after þa, ne and wh-constituents); see Mitchell (1985: §1685,
§1720–1729).3 We will therefore concentrate on the remaining 358 instances of
man. They can be broken down into the following syntactic contexts:

Table 1. The use of man in Ælfric

Clause type Subtype Numbers

main clause 1. man clause-initial 26

2. man not clause-initial 68
subclause 3. subjunctive that-clause 135
4. other 129

We will look at each of these subtypes in detail.

3.2 Man clause-initial

Subtype 1 is the one in which man occupies its most prominent position: the
first position of the sentence. The fact that this group is the smallest of all is not
surprising: this position is a topic position, the position of given, background
information, often reserved for anaphoric material that provides a link with the
previous sentence, to maximize textual cohesion (‘unmarked theme’). The topic
position in Present-day English is reserved for subjects. There is a secondary
focus position in front of the subject in Present-day English, reserved for
‘marked themes’ and hosting material that is prominent in that it is either new
or contrastive with what has gone before; for a discussion of marked and
unmarked themes see Downing & Locke (1995: 222–237) and Halliday (1994).
Man is too weak in content to be a marked theme; as unmarked theme it only
provides anaphoric reference in the loosest sense, and is therefore only expected
to occur when textual cohesion is relaxed, in other words, when it starts a new
discourse unit, or when other elements take over the burden of textual cohe-
sion. In (4), for instance, the two sentences are linked by the repetition of the
verb in the second sentence (smyrað ‘anoints’), a reminder that textual coher-
ence does not always have to be accomplished by the unmarked theme:
(4) Hu is he gesmyrod? Man smyrað cyning mid gehalgodum ele
how is he anointed one anoints a king with consecrated oil
þonne man hine to cyninge gehalgað. ·ÆCHom II, 1, 7.162Ò
when one him as king consecrates
 The loss of the indefinite pronoun man 185

“How is he anointed? One anoints a king with consecrated oil when he is

consecrated as king”

A study of the other 25 examples of man in clause-initial position bears out that
they all occur at, or close to, the beginning of a new episode, often (in six cases)
signalled by the presence of þa ‘then’, which has been argued to be an episode
boundary marker (Foster 1975; Enkvist 1986; Brinton 1993); but even when
there is no þa, it is clear that many of these examples of clause-initial man
introduce a new episode, e.g. (5):
(5) (ælc man sceal arisan þonne þe æfre on life wæs; wære he on wætere
adruncen, oððe hine wilde deor æton, oððe hine fyr forbærnde færlice to
duste, and ðæt dust wurde toworpen mid blædum, swaðeah se ælmihtiga
God mæg hine eft aræran, se ðe ealle þas woruld geworhte of nahte, and se
ðe thises ne gelyfð, ne bið his geleafa naht.)
Man bewint þone deadan gewunelice mid reafe (…) ·ÆHom 11, 332–339Ò
One shrouds the [Link] usually with garment
“(Every man will then rise again who was ever alive; if he was drowned in
water, or if wild animals ate him, or if fire burned him unexpectedly to
ash, and the ash was scattered by blasts, the Almighty God can neverthe-
less raise him up again, he who created this whole world out of nothing,
and he who does not believe this, has no faith.)
The dead person is usually shrouded in a garment (…)” [and the text
goes on to describe the spiritual significance of this shroud]

Note that the last clause is best translated by a passive, and not by an indefinite
— an indication that man is an emptier element than PDE one, people, they, we
or you and is hardly more than a place-holder for the agent role.
Eleven of the 26 clause-initial uses are precepts: Man ne mæg ‘You/One
cannot…’, Man mot ‘You/One are not allowed to…’ or Man sceal ‘You/One
must…’, often, but not invariably, at episode boundaries. This appears to be a
use in which man refers to the collectivity of human beings rather than an
indefinite set of human beings, and as such could be said to have more semantic
content than in the previous examples:
(6) Hit is awriten: Man sceal hine gebiddan to his drihtne
it is written one must himself pray to his lord
“It is written: One must pray to one’s lord” ·ÆCHom I, 11, 172.23Ò
186 Bettelou Los

The remaining four cases involve the verb cweþan ‘say’, also at episode bound-
aries; an example is (7), in which the textual cohesion is accomplished by
þissum ‘this’:
(7) Man cwæð on bocum gehu be þissum bearnteame,
one says in books [Link] about this progeny
þæt (…) ·ÆHomM 2 98Ò
that
“It is said in books in various ways about this progeny, that (…)”

The other variants of this expression are either active with indefinite we: We
rædað on bocum ‘we read in books’ (e.g. ·ÆCHom I, 18 244.15Ò), or passive, as
in On bocum is geredd ‘In books is read’ (e.g. ·ÆCHom I, 22 320.2Ò), or start
with an expletive, as in hit stent on bocum awriten ‘It stands written in books’
(e.g. ·ÆHom 21 397Ò). There are two things to note about such phrases. The
first one is that they confirm the link between OE man-clauses and the use of
passives which has been noted before (e.g. Fischer 1994); man does the same job
as the passive in that it obscures the agent of the action. The second one is that
all initial constituents here are unmarked themes that do not refer back to the
previous sentence — in (7), for instance, that function is delegated to þissum
‘this’ in the prepositional phrase be þissum bearnteame ‘about this progeny’; in
most of the other instances, there is no overt link at all, usually because these
sentences start a new episode.
Clause-initial man occupies a niche in Old English that is not affected by
the syntactic changes that affect the other uses of man, as we will discuss below,
but it is clear that it is a small niche, and could well be too small to save man
from its decline through the various factors outlined in Section 2 above.

3.3 Man not clause-initial

3.3.1 Verb-second
Rissanen notes that OE man seems to be used “when the focus of interest is
totally removed from the referent of the subject to some other element of the
clause” (Rissanen 1987: 417–418; 1997: 514) and this is an important pointer to
its loss in this subtype.
The position of man in subtype 2 is far less prominent than in subtype 1, and
is as such far better suited to such an empty element as man, which may explain
why subtype 2 is also far more numerous. Subtype 2 locates man in the specifier
of a functional position which has been identified as the MoodPhrase (MP) by
 The loss of the indefinite pronoun man 187

van Kemenade (2000), with SpecMP hosting clitics. In this position, man will
follow the finite verb when the first constituent, in SpecCP, is a wh-word, the
negator ne or a member of a restricted group of adverbs, most prominently þa
‘then’ — as do all other subjects, nominal and pronominal alike:
(8) ða gebrohte man him to, tomiddes þam folce, ænne dumne mann,
then brought one him to among the people a dumb man
& se wæs eac swilce deaf (…) ·ÆHom 18, 25Ò
and he was also likewise deaf
“then was brought to him/then people brought to him, among the peo-
ple, a man who was dumb, and also deaf”

The clitic nature of pronoun subjects emerges when the first constituent is not a
wh-word, the negator ne or an adverb like þa ‘then’, but a topicalized nominal or
prepositional object. When these constituents appear in the first position, subject
nominals invert in Old English, but pronouns do not, and neither does man:
(9) Be ðisum lytlan man mæg understandan (…) ·ÆGenPref 72Ò
by this little one may understand
“By means of this little thing can be understood (…)”

Van Bergen (2000a, b, c) identifies this as one of the diagnostics for pronominal
status, indicating that man in these instances at least behaves as a pronoun
rather than as the nominal expression it must have been originally. This special
behaviour of these pronominals can be translated structure-wise as a failure of
the verb to move all the way up to C in such cases; it sits lower, in the Head of
the MP (M in tree (11) below). The clitic behaviour of man is in itself not
important for our argument, but explains why we discuss examples like (8) and (9)
under one heading: man occupies the same position in each case (SpecMP); it is
the position of the verb that is different: C in (8), having moved there from V, via
T and M, as in (10), but M in (9), having moved there from V via T as in (11).
188 Bettelou Los

(10) CP

Spec C′
ða
C MP
gebrohte
Spec M′
man
M TP

Spec T′

T VP

Spec V′

(11) CP

Spec C′
Be ðisum lytlan
C MP

Spec M′
man
M TP
mæg
Spec T′

T VP

Spec V′

Because subject pronouns sit in SpecMP, it is only attestations of inversion of

the finite verb and a nominal (as opposed to pronominal) subject that consti-
tutes solid evidence that V2 is still in place. The decline of such attestations
indicates that the loss of V2 sets in at the end of the fourteenth century and
gains momentum during the fifteenth century (van Kemenade 1987: 219–223).
 The loss of the indefinite pronoun man 189

The evidence suggests that the loss of V2 should be translated in syntactic terms
as a failure of the finite verb to move to a higher structural position in the clause
(C after clause-initial ne, þa and wh-forms, M after clause-initial NPs or PPs);
after the demise of V2 the verb does not climb any higher than T (‘Tense’).
Subject and topic in this analysis are both in SpecCP in Old English; but when
V2 is lost, the subject remains in SpecTP and no longer appears in SpecCP
(based on van Kemenade 1997a, b, 2000, 2001).4

3.3.2 V2 and information structure

What exactly does V2 mean in terms of information structure? The first
position of a clause provides the starting point of the entire message, and
usually contains given information (the ‘theme’); the most favourable position
for new information is towards the end of the clause (‘end focus’). In Present-
day English, unmarked themes in ordinary declarative clauses are invariably
encoded by the nominative subject position in SpecTP, which can only be filled
by an NP. Other material may precede it, but such items constitute marked
themes which usually imply a contrast of some sort, and have a specific dis-
course function (see Section 3.2 above). In a V2 language like Dutch and Old
English, ‘unmarked’ themes are not restricted to a syntactic category, as just
about any XP may occupy the first position. Unmarked themes may be adverb-
ials, objects or prepositional complements, which means that all these constitu-
ents may be used for textual cohesion without evoking any of the special
contrastive discourse effects that are associated with marked themes.
When V2 is lost and SV is generalized in English, only subjects can be
unmarked themes. This diminishes the syntactic options open to the language
user for positioning old information, and it is probably no coincidence that the
early fifteenth century produces a number of innovative strategies to turn
constituents into subjects: prepositional passives as in PDE he was well thought
of, the doctor was sent for etc. (see Denison 1985: 190–191), and subject-to-
object raising (ECM). These innovations represent new mechanisms to front an
NP which is given information and place it in the unmarked theme position,
and are therefore connected to the loss of V2.
The case of the passive subject-to-object raising (or ECM-) construction is
instructive in this respect as it shows the advantage of an impersonal passive
construction over an indefinite construction.5 It has often been noted that
ECM-constructions occur mainly in the passive, from the time of their first
attestations to the present day (e.g. Warner 1982 and Fischer 1989, 1992 for
Middle English; Fanego 1992 for Shakespeare’s English; Mair 1990 for Present-day
190 Bettelou Los

English). A corpus example from Mair (1990) demonstrates that the answer lies
in the way such passive constructions enable old information to be fronted
without becoming marked themes (for a detailed discussion of passive ECM
and information structure, see also Noël 1998):
(12) Thanks to the ubiquitous television set, the best known Canadians in
Britain are, quite possibly, Bernard Braden, Hughie Green and Robert
McKenzie. Others more talented — Jon Vickers, Lynn Seymour, Morde-
cai Richler, Sir William Butlin, John Hemming, Oscar Petersen, Garfield
Weston, Paul Anka, Glenn Ford, Yvonne de Carlo, Raymond Burr, Don-
ald Sutherland and Christopher Plummer — are probably seldom iden-
tified as Canadians. Many of them are generally assumed to be Ameri-
cans, which raises the whole struggle to maintain a separate identity
from her giant neighbour. (Mair 1990: 180)

Textual coherence demands that the given information of that final sentence
(many of them, referring to the list of names) appears as subject, and Mair notes
that none of the available alternatives to a passive ECM (active ECM as in (13a),
indefinite pronoun and finite clause as in (13b), or expletive subject and finite
clause as in (13c)) will meet this requirement. Note that a fourth strategy,
fronting out of an active ECM, as in (13d), is also available, but equally undesir-
able, as it turns many of them into a marked theme:
(13) a. People generally assume many of them to be Americans.
b. People generally assume that many of them are Americans.
c. It is generally assumed that many of them are Americans.
d. Many of them people generally assume to be Americans.

A V2 language does not need such stratagems as there is no ban on non-subjects

to function as unmarked theme, as illustrated by the Dutch translations of the
relevant clause in (12); (14a) is active with indefinite subject men ‘one’ (akin to
OE man), and (14b) is passive. The two constructions are functionally equiva-
lent. The unmarked theme is in both cases a PP adjunct (in bold) — a strategy
that was no longer available to English after V2 was lost:
(14) a. Van deze mensen neemt men meestal aan dat ze Amerikanen
of these people takes one generally on that they Americans
zijn
are
 The loss of the indefinite pronoun man 191

b. Van deze mensen wordt meestal aangenomen dat ze Amerikanen

of these people is generally [Link] that they Americans
zijn.
are
“These people are generally assumed to be Americans”

Such contrasts between V2 and non-V2 languages are discussed in translation

studies (e.g. Hannay & Keizer 1993) and in translation manuals (Lemmens &
Parr 1995), as particular care has to be taken when translating from Dutch to
English to make sure that unmarked themes in Dutch remain unmarked themes
in English.
This example demonstrates the important role of subjects in Present-day
English: the burden of textual cohesion is mainly carried by them. This means
that they can rarely afford to be contentless.6

3.3.3 The content of Dutch men: evidence from translation studies

How contentless was OE man? There are two pointers. One is the evidence from
the role of a cognate indefinite in a language in which V2 is still in place. There
is a profound difference between the Dutch indefinite men and PDE indefinites
like people, you, we or they. Translation studies have shown that although Dutch
men generally provides an excellent equivalent for such PDE indefinites, the
reverse is not true: Dutch men rarely allows a translation by PDE indefinites.
Translation textbooks observe that men is often used as a deliberate strategy to
obscure the agent, and note that a good English translation generally needs a
more specific indication of the subject (Lemmens & Parr 1995: 275). The first
option is to go for a passive construction in translation. An example is (15), a
V2 construction with the adjunct daarom ‘for this reason’ as its unmarked
theme, and the finite verb voegt ‘adds’ in second position followed by the
subject men (Modern Dutch does not have a system of pronominal clitics as we
saw operating in the OE example of (9) above):
(15) Sap van druiven is maar kort houdbaar. Daarom voegt men
juice from grapes is only briefly non-perishable therefore adds one
wijngist aan het sap toe.
yeast to the juice to
“Grape juice can be stored only for a short period of time, and that is
why yeast is added to it.” (Recommended translation, ibid. 308)

Dutch men is apparently so devoid of content that it obscures the identity of the
agent as effectively as a passive construction, which explains the equivalence
192 Bettelou Los

we saw in (14a–b) above. In contrast to contentless men, PDE indefinite

pronouns like we, you, they or people always define interlocutors, either by
inclusion or exclusion. Even one, which is perhaps the most impersonal of all
indefinites, defines and includes interlocutors.7 It clearly conveys a stronger
sense of involvement on the part of both interlocutors than the passive, witness
examples like (16a–b):
(16) a. Physical pain is often inflicted upon children.
b. One often inflicts physical pain upon children.

These examples have been taken from Siewierska (1984: 243) who notes that
(16b) carries unfavourable overtones because even with one there is a sense that
the interlocutors are involved in some way. No such effects are conveyed by
(16a), the impersonal passive.
When the patient argument conveys new information, a passive is less
desirable because it would locate an informationally salient constituent in the
subject position, and the sentence, though perfectly grammatical, becomes
stylistically awkward, especially if the constituent is particularly heavy and
complex; an example is (17). In such cases, one may come to the rescue, as in
(18), from Siewierska (1984: 247; see also Seoane-Posse 2000: 105–106).
(17) The snorting contempt with which Truman would have regarded the
$1000 cowboy boots and the Adolfo gowns can be imagined.
(18) One imagines the snorting contempt with which Truman would have
regarded the $1000 cowboy boots and the Adolfo gowns. (“A good snob
is hard to find”, Time Sept. 19, 1983, quoted in Siewierska 1984)

One, then, is something of a last resort: if the passive, the most natural choice
when the agent is an indefinite entity, is unavailable for information-structural
reasons, one will rescue the structure.
There is evidence that OE man exhibits the same functional equivalence
with passives as Dutch men. Consider the following Old English translations
from Latin:
(19) Lat.: Eos qui ducuntur ad mortem eruere ne cesses
those who ledpass3pl to death freeinf not hesitate2sg
OE translation:
þa þe man læt to deaðe alys hi ut symble.
those whom people lead to death free them out always
“always set free those who are being led to death” ·Æls (Edmund) 214Ò
 The loss of the indefinite pronoun man 193

The Latin passive is translated as an active structure, with man as its subject, even
though a passive translation is perfectly possible. Another example is (20), which
translates a non-finite passive construction by a finite active construction, even
though a finite passive construction would again be perfectly acceptable:
(20) Lat.: Sicut Filius hominis non venit ministrari sed ministrare
so son [Link] not came servepassinf but serveactinf
OE translation:
Swa mannes sunu ne com þæt him man þenode ac þæt he þenode
so of-man son not came that him people served but that he served
·Mt (WSCp) 20.28Ò
“So did the Son of Man not come to be served but to serve”

Because subjects have such a prominent function in Present-day English (as

only they can constitute unmarked themes after V2 was lost), there is no niche
for contentless indefinites of the men/man type. Indefinite subjects almost
inevitably have to refer, witness the strong inclusion effect of PDE one in (16b)
above. A typical example of this referring function is the use of one as a circum-
spect, non-committal way of introducing or summarizing the views of the
speaker, especially in logical discursive or scholarly texts (see also OED s.v. One
21). Authors of fiction often play upon the formal, distancing, pompous effect
of one in such texts. In the following fragment, one is made into a protective
shield behind which the maligned Tuke cowers:8
(21) “It is of course utterly unrealistic to suppose that reputations in litera-
ture are made overnight”, said Tuke, who had been brooding on Oding-
sels’ hard words. “One despises egotism, of course, but one instances
oneself; one can give Giles a few years, and one is perhaps more engagé,
but one has certainly not been overwhelmed with recognition. As for
music being, au fond, more serious than letters, well one feels perhaps
that those who are committed to an art are the best judges of its limits.”
“Better judges than technicians, however capable”, said Miss Tooley, bri-
dling. Everybody knew that when Tuke began to refer to himself as
“one” Bridget would do battle for him. (Robertson Davies (1980 [1951])
The Salterton Trilogy, 653–654)

If the functional equivalence of OE man with Dutch men is valid, the conclu-
sion of this section must be that the loss of V2 played an important role in the
demise of man in main clauses. Subjects no longer occupied the same position
as topics (SpecCP), which meant that marked themes came to be also syntactically
194 Bettelou Los

marked (as topicalized constituents), whereas unmarked themes were restricted

to subjects in SpecTP. This means that subjects came to play a far greater role in
maintaining textual cohesion. This left little scope for the indefinite pronoun
man whose main role had been to provide a contentless subject, functionally
equivalent to the passive. The changes in information structure resulting from
the loss of V2 promoted the use of passives, which took over the function of
man in main clauses.

3.4 Man in subclauses

3.4.1 In subjunctive that-clauses
OE man in subjunctive subclauses (subtype 3 in Table 1) appears typically in
clauses of the type in (22) (evaluative predicates), (23) (verbs of commanding
and permitting) and (24) (verbs of persuading and urging):
(22) Nis na god þæt man nime his bearna hlaf. and
not-is not good that one takessubj his childrengen bread and
wurpe hundum ·ÆCHom II,8 69.88Ò
throwsubj dogsdat
“It is not right to take one’s children’s bread and throw it to the dogs”
(23) Godes æ byt eac þæt man arwurþige symble hys fæder &
God’s law commands also that one honourssubj always his father and
modor mid mycelre underþeodnysse ·ÆAbusMor 133Ò
mother with great obedience
“God’s law commands also that people always honour their father and
mother with great obedience”
(24) swa swa Crist lærde þæt man don sceolde, þa þa he on his godspelle
as Christ taught that one do should when he in his gospel
swutolice þus cwæð ·WPol 2.1.1 53Ò
clearly thus spoke
“as Christ taught that people should do, when he in his gospel clearly
spoke as follows”

Sentence (24) is unusual in that verbs of persuading and urging, when comple-
mented by a subjunctive clause, are almost invariably accompanied by an object
whose reference is identical to that of the subject of the subjunctive clause, as in
(25) (object and dependent subject in bold):
(25) & we lærað eac georne manna gehwylcne, þæt he Godes ege hæbbe
and we teach also gladly mengen eachacc that he God’s awe havesubj
 The loss of the indefinite pronoun man 195

symle on his gemynde ·LawICn 25Ò

always on his mind
“and we also gladly teach each man that he should have the fear of God
always in his mind”

This is a form of obligatory object control, in which the content of the subject
of the embedded clause is determined by the object in the higher clause. This
property of verbs of persuading and urging made it possible for the to-infinitive
to appear as complement, as the identity of the non-overt infinitival subject
could readily be inferred from the object of the higher clause (‘object-controlled
PRO’). An example of a to-infinitive after this same verb (læran) is given in (26)
(controlling object and controlled PRO subject in bold):
(26) he (…) heo lærde [PRO] to healdanne regollices liifes þeodscipe
he themacc taught to observe regulated life’s discipline
·Bede 3 16.226.26Ò
“he taught them to observe the discipline of a regulated life”

Verbs of commanding and permitting exhibit object control as well, but do not
always require the presence of an object. Objects (invariably in the dative) are
present in (27) (with subjunctive clause, and object control: the overt subject of
the embedded clause is identical in reference to the object of the higher verb)
and (28) (with to-infinitive, and object-controlled PRO). Controllers and
controlled subjects in bold:
(27) þam cildum ic bead þæt hi gehyrsume wæron fæder and
the childrendat I ordered that they obedient were father and
meder ·ÆCHom I, 26 378.23Ò
mother
“I ordered the children to be obedient to their father and mother”
(28) & ic (…) þe bead [PRO] mine bebodu to healdanne.
and I youdat order my commands to keep
“and I order you to keep my commands” ·Hom S 3, 53Ò

When there is no object in the higher clause that may serve as a controller of the
lower subject, the subject position is filled by man if the complement is a
subjunctive clause (which, as a finite clause, requires a subject), as in (23)
above, or by generic PRO when the complement is a to-infinitive, as in (29):
(29) ic æfre fram frymðe bebead [PROgen] þone drihtenlican dæg to
I ever from beginning ordered the lordly day to
196 Bettelou Los

healdenne ·HomU 46 134Ò

hold
“I have ever from the beginning ordered to keep the Lord’s day”

Evaluative predicates rarely have overt controllers; an example is (30) (controllers

and controlled subjects in bold). The majority of instances have generic PRO
(in to-infinitives) as in (31), or man (in subjunctive clauses), as in (22) above.
(30) þæt he (…) smeage (…), hwæt him sy betst [PRO] to
that he considersubj what himdat issubj best to
donne and hwæt [PRO] to forganne. ·WPol 6.2 146Ò
do and what to forgo
“that he may consider what he should do and what he should not do”
(31) acsiað georne hu betst sy [PROgen] to farenne ·WHom 11 216Ò
ask eagerly how best issubj to go
“ask eagerly how it is best to go”

It seems that the to-infinitive developed into a non-finite alternative for the
subjunctive clause in Old English. Quantitative studies show that such subjunc-
tive that-clauses (after desiderative or conative verbs meaning ‘try’ or ‘want’,
verbs of commanding and permitting, and verbs of persuading and urging)
were replaced by to-infinitives in the transition from Old English to Middle
English (Los 1998, 1999). This may seem strange at first sight, as one would not
expect the subjunctive clause to be functionally equivalent to an infinitive, as
the latter has no positions for independent tense (T) or overt subjects (AgrS).
The information in T and AgrS in subjunctive clauses is, however, redundant in
the majority of cases: subjunctive Tense is never independently meaningful but
copied from the Tense of the higher clause (as confirmed by crosslinguistic
studies, e.g. Noonan 1985: 53). The content of AgrS in the subjunctive clause is
similarly predetermined (‘controlled’) — by the object (or, in the case of
monotransitive verbs, the subject) of the higher clause. The absence or presence
of these structures does not materially affect or limit the range of expression of
the to-infinitive, compared to the subjunctive clause; the subjunctive clause
does not have an edge over the to-infinitive, in spite of its more extensive
structure, and it is this that allowed the to-infinitive to become functionally
equivalent to the subjunctive that-clause.
OE man, then, could be said to be in competition with generic PRO. Note
that OE man (and German man, and Dutch men) has all the hallmarks of an
overt manifestation of generic PRO, including the ban on appearing in object
 The loss of the indefinite pronoun man 197

position (of verb or preposition): man/men only has a nominative form. When
the to-infinitive ousts such subjunctive clauses (and note that most of the
that-clauses in this section translate as a to-infinitive in Present-day English),
this use of man disappears with it. A small niche remains in those cases in which
the selection of generic PRO is unavailable; after a verb of persuading and
urging like teach, for instance, PRO would be interpreted as controlled PRO,
and this is one of the uses in which we encounter a finite clause in Present-day
English with one as subject:
(32) The Spanish daily El Pais quoted Mr Cela as saying that he did not in-
tend to let the award change his habits, because the British side of his
character had taught him one should only appear in the press thrice in a
lifetime — “when born, when dead and when receiving the Victoria
Cross”. (Microconcord Corpus)
(π the British side of his character had taught himi PROi to appear in the
press thrice in a lifetime)

3.4.2 In other subclauses

Together with subtype 1 (man clause-initial), man in subclauses other than the
subjunctive that-clause (= subtype 4) appears to be the type that is least affected
by the changes discussed so far, although, like Dutch men, it does not invariably
allow a translation by one of the PDE indefinites. In example (33), for instance,
a PDE passive appears to be the best option.
(33) And þu wel wast, leof, þæt hit wile hearmian þinum cynerice heora
and you well know dear that it will harm your kingdom their
receleasnysse, gyf him man ne gestyrð heora stuntnysse.
neglect if them one not urges their folly
“and you know well, dear King, that these people will harm your king-
dom by their flouting of the rules if they are not urged from their folly”
·ÆHomM 14 152Ò

4. Conclusion

Many proposals have been made in the literature about the disappearance of the
OE indefinite pronoun man (Middle English men/me). We have suggested two
additional factors which almost completely destroyed the niche occupied by man
in Old English. The first one was the loss of V2, which affected the information
<DEST "los-n*">

198 Bettelou Los

structure of the clause, and promoted the use of various passive constructions
over the use of an active construction with a man subject. The second one was
the competition between to-infinitives and subjunctive clauses which resulted
in man in these contexts being largely ousted by generic (or arbitrary) PRO.

Notes

* The research reported in this paper was made possible by a grant from the Netherlands
Organization for Scientific Research (NWO), grant number 300–74–021. The author
gratefully acknowledges helpful hints and comments from Teresa Fanego and an anonymous
referee.
1. Rissanen (1997: 521) points out that this new use should not be regarded as some form of
regrammaticalization, as it does not represent a reflex of the original indefinite pronoun (OE
man, ME men, me), which had clearly been formally differentiated from the noun man well
before this period.
2. Throughout this article, the reference to an OE text is enclosed in ·Ò and follows the
system of short titles as employed in Healey & Venezky (1985) (in turn based on the system
of Mitchell, Ball & Cameron 1975, 1979).
3. Mitchell also notes the problems of editorial punctuation practice and scribal confusion
between the positura (signifying a full stop) and the Tironian sign for the Latin word et,
which means that some instances at least of 7 in OE manuscripts do not stand for and but
mark the beginning of a new sentence (ibid. §1724), which further complicates the analysis
of and/ac clauses.
4. The evidence for a ‘MoodPhrase’ remains as long as pronominal subjects continue to
show a positional difference compared to full nominal subjects. Such a difference is attested
until well into the seventeenth century with respect to the placement of the negative element
not (Fischer et al. 2000: 135; Rissanen 1994, 1999).
5. These labels refer to different analyses of the same construction, depending on how the
subject of such infinitival constructions (e.g. him in a PDE sentence like I consider him to be
a fool) is treated: as a deep-structure subject that raises to the object position of the higher
clause (as in the earlier Extended Standard Theory or, more recently, Minimalist Theory) or
as a constituent that is assigned Case by the higher verb without leaving its deep-structure
subject position (Government & Binding Theory). The question of which label or analysis
fits this construction best is irrelevant for our present purpose.
6. It could be argued that contentless subjects did survive in Present-day English, witness the
‘dummy’ element it. The raison d’être of this element is entirely structural: English clauses
must have overt subjects, even if the verb has no thematic roles to assign (weather-it) or
assigns only an internal role, which happens to be conferred on a constituent that cannot
raise to the subject position for structural reasons (expletive it). OE man, on the other hand,
does carry a thematic role (agent of the verb) and is therefore not as ‘dummy’ as dummy it.
 The loss of the indefinite pronoun man 199

7. There is an additional difference between PDE one and Dutch men in that the former is
associated with a formal register. The greater formality of one versus the indefinite use of you,
we, they appears to be a fairly recent development (cf. Seoane-Posse 2000 for the situation of
these indefinites in Early Modern English). Unlike PDE one, Dutch men is not restricted to
formal discourse and easily surfaces in casual conversation, usually in the less prominent,
inverted position, as in (i):
(i) maar dat realiseer je je normaliter niet dat het een eh zo’n groot gedeelte
but that realize you yourself normally not that it a uh such.a large part
van de elektriciteit wordt opgewekt door atoomcentrales hè,
of the electricity is generated by nuclear power stations [discourse prt]
niet in Nederland, in Nederland is men dat zich absoluut niet bewust
not in Netherland in Netherland is one that oneself absolutely not aware
“But you usually don’t realize that it is, uh, that such a large part of electricity is
generated by nuclear power stations, do you, not in the Netherlands, in the Netherlands
people are not aware of this at all” (Ernestus 2000).
8. The same effect could apparently be produced with other indefinite expressions, as with
a man in the following Dickens fragment, in which the pompous distancing effect of the
indefinite expression to refer to one’s own person is taken to ridiculous excess:
(i) One of Steerforth’s friends was named Grainger, and the other Markham. They were
both very gay and lively fellows; Grainger, something older than Steerforth; Markham,
youthful-looking, and I should say not more than twenty. I observed that the latter always
spoke of himself indefinitely, as “a man”, and seldom or never in the first person singular.
“A man might get on very well here, Mr Copperfield”, said Markham — meaning
himself.
“It’s not a bad situation”, said I, “and the rooms are really commodious.”
“I hope you have both brought your appetites with you?” said Steerforth.
“Upon my honour”, returned Markham, “town seems to sharpen a man’s appetite.
A man is hungry all day long. A man is perpetually eating.” (Dickens, 1849–1850, David
Copperfield, 419)

References

Bergen, Linda D. van. 2000a. “The Indefinite Pronoun man: ‘Nominal’ or ‘Pronominal’?”
Generative Theory and Corpus Studies: A Dialogue from 10 ICEHL ed. by Ricardo
Bermúdez-Otero, David Denison, Richard M. Hogg & Chris B. McCully, 103–122.
Berlin & New York: Mouton de Gruyter.
Bergen, Linda D. van. 2000b. “More on man”. Paper delivered at the Eleventh International
Conference on English Historical Linguistics (11 ICEHL), Santiago de Compostela,
September 2000.
Bergen, Linda D. van. 2000c. Pronouns and Word Order in Old English, with Particular
Reference to the Indefinite Pronoun man. Ph.D. dissertation, University of Manchester.
200 Bettelou Los

Brinton, Laurel J. 1993. “Episode Boundary Markers in Old English Discourse”. Historical
Linguistics 1989. Papers from the 9th International Conference on Historical Linguistics,
New Brunswick, 14–18 August 1989 ed. by Henk Aertsen & Robert J. Jeffers, 73–89.
Amsterdam & Philadelphia: John Benjamins.
Davies, Robertson. 1980 [1951]. The Salterton Trilogy. Harmondsworth: Penguin.
Denison, David. 1985. “Why Old English Had no Prepositional Passive”. English Studies
66.189–204.
Dickens, Charles. 1849–1850. David Copperfield. Harmondsworth: Penguin.
Downing, Angela & Philip Locke. 1995. A University Course in English Grammar. Hemel
Hempstead: Phoenix ELT.
Enkvist, Nils E. 1986. “More about the Textual Functions of the Old English Adverbial þa”.
Linguistics across Historical and Geographical Boundaries: In Honour of Jacek Fisiak on the
Occasion of his Fiftieth Birthday ed. by Dieter Kastovsky & Aleksander Szwedek, vol. I,
301–309. Berlin & New York: Mouton de Gruyter.
Ernestus, Mirjam. 2000. Voice Assimilation and Segment Reduction in Casual Dutch: A
Corpus-based Study of the Phonology-Phonetics Interface (= LOT Dissertation Series, 36).
Utrecht: LOT
Fanego, Teresa. 1992. Infinitive Complements in Shakespeare’s English. Santiago de Compos-
tela: Universidade de Santiago de Compostela (Servicio de Publicacións e Intercambio
Científico).
Fischer, Olga C. M. 1989. “The Origin and Spread of the Accusative and Infinitive Construc-
tion in English”. Folia Linguistica Historica 8.143–217.
Fischer, Olga C. M. 1992. “Syntactic Change and Borrowing: The Case of the Accusative-and-
Infinitive Construction in English”. Internal and External Factors in Syntactic Change ed.
by Marinel Gerritsen & Dieter Stein, 17–88. Berlin & New York: Mouton de Gruyter.
Fischer, Olga C. M. 1994. “The Fortunes of the Latin-type Accusative and Infinitive Con-
struction in Dutch and English Compared”. Language Change and Language Structure:
Older Germanic Languages in a Comparative Perspective ed. by Toril Swan, Endre Mørck
& Olaf Jansen Westvik, 91–133. Berlin & New York: Mouton de Gruyter.
Fischer, Olga C. M., Ans van Kemenade, Willem Koopman & Wim van der Wurff. 2000. The
Syntax of Early English. Cambridge: Cambridge University Press.
Foster, Robert. 1975. “The Use of þa in Old and Middle English”. Neuphilologische Mitteil-
ungen 76.404–414.
Halliday, M. A. K. 1994 [2nd rev. ed.]. An Introduction to Functional Grammar. London:
Edward Arnold.
Halliday, M. A. K., & Ruqaiya Hasan. 1976. Cohesion in English. London: Longman.
Hannay, Mike & Evelien Keizer. 1993. “Translation and Contrastive Grammar: The
Grammatical versus the Communicative Strategy”. Toegepaste Taalwetenschap in
Artikelen 45.65–88.
Healey, Antonette diPaolo & Richard L. Venezky, comps. 1985. A Microfiche Concordance to
Old English. Toronto: Pontifical Institute of Mediaeval Studies, University of Toronto.
Kemenade, Ans van. 1987. Syntactic Case and Morphological Case in the History of English.
Dordrecht: Foris.
 The loss of the indefinite pronoun man 201

Kemenade, Ans van. 1997a. “Topics in Old and Middle English Negative Sentences”.
Language History and Language Modelling ed. by Raymond Hickey & Stanislaw Puppel,
293–306. Berlin & New York: Mouton de Gruyter.
Kemenade, Ans van. 1997b. “Negative-initial Sentences in Old and Middle English”.
Festschrift for Roger Lass on his Sixtieth Birthday (= Studia Anglica Posnaniensia, 31) ed.
by Jacek Fisiak, 91–104. Poznan: Uniwersytet im. A. Mickiewicza.
Kemenade, Ans van. 2000. “The Syntax and Use of some Light Elements in OE”. Paper
delivered at the Eleventh International Conference on English Historical Linguistics (11
ICEHL), Santiago de Compostela, September 2000.
Kemenade, Ans van. 2001. “Jespersen’s Cycle Revisited: Formal Properties of Grammaticali-
zation”. Diachronic Syntax: Models and Mechanisms ed. by Susan Pintzuk, George
Tsoulas & Anthony Warner. Oxford: Oxford University Press.
Lemmens, Marcel & Tony Parr. 1995. Handboek voor de vertaler Nederlands-Engels. Gronin-
gen: Wolters-Noordhoff.
Los, Bettelou. 1998. “The Rise of the to-infinitive as Verb Complement”. English Language
and Linguistics 2.1–36.
Los, Bettelou. 1999. Infinitival Complementation in Old and Middle English (= LOT Disserta-
tion Series, 31). The Hague: Thesus.
Mair, Christian. 1990. Infinitival Complement Clauses in English: A Study of Syntax in
Discourse. Cambridge: Cambridge University Press.
Microconcord. 1993. [= A concordance program by Mike Scott & Tim Johns (Oxford English
Software)]. Oxford: Oxford University Press.
Mitchell, Bruce. 1985. Old English Syntax, 2 vols. Oxford: Clarendon.
Mitchell, Bruce, Christopher Ball & Angus Cameron. 1975. “Short Titles of Old English
Texts”. Anglo-Saxon England 4.207–221.
Mitchell, Bruce, Christopher Ball & Angus Cameron. 1979. “Addenda and Corrigenda”.
Anglo-Saxon England 8.331–333.
Noël, Dirk. 1998. “Infinitival Complement Clauses in Modern English: Explaining the
Predominance of Passive Matrix Verbs”. Linguistics 36.1045–1063.
Noonan, Michael. 1985. “Complementation”. Language Typology and Syntactic Description
ed. by Timothy Shopen. Vol. II: Complex Constructions, 42–140. Cambridge: Cambridge
University Press.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S. C. Weiner. Oxford: Oxford University Press.
Raumolin-Brunberg, Helena & Leena Kahlas-Tarkka. 1997. “Indefinite Pronouns with
Singular Human Reference”. Grammaticalization at Work: Studies of Long-term
Developments in English ed. by Matti Rissanen, Merja Kytö & Kirsi Heikkonen, 17–85.
Berlin & New York: Mouton de Gruyter.
Rissanen, Matti. 1967. The Uses of One in Old and Middle English (= Mémoires de la Société
Néophilologique de Helsinki, 31). Helsinki: Société Néophilologique.
Rissanen, Matti. 1987. “Old English Indefinite Pronouns Meaning ‘Some’ and ‘Any’, with
Special Reference to hw-forms”. Neophilologica Fennica (= Mémoires de la Société
Néophilologique de Helsinki, 45) ed. by Leena Kahlas-Tarkka, 411–428. Helsinki: Société
Néophilologique.
</TARGET "los">

202 Bettelou Los

Rissanen, Matti. 1994. “The Position of Not in Early Modern English Questions”. Studies in
Early Modern English ed. by Dieter Kastovsky, 339–348. Berlin & New York: Mouton de
Gruyter.
Rissanen, Matti. 1997. “Whatever Happened to the Middle English Indefinite Pronouns?”
Studies in Middle English Linguistics ed. by Jacek Fisiak, 513–529. Berlin & New York:
Mouton de Gruyter.
Rissanen, Matti. 1999. “Isn’t It? or Is It not? On the Order of Postverbal Subject and Negative
Particle in the History of English”. Negation in the History of English ed. by Ingrid
Tieken-Boon van Ostade, Gunnel Tottie & Wim van der Wurff, 189–205. Berlin & New
York: Mouton de Gruyter.
Seoane-Posse, Elena. 2000. “Impersonalising Strategies in Early Modern English”. English
Studies 81.102–116.
Siewierska, Anna. 1984. The Passive: A Comparative Linguistic Analysis. London: Croom
Helm.
Warner, Anthony R. 1982. Complementation in Middle English and the Methodology of
Historical Syntax. London: Croom Helm.
<LINK "meu-n*">

<TARGET "meu" DOCINFO

AUTHOR "Anneli Meurman-Solin"

TITLE "The progressive in Older Scots"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

The progressive in Older Scots*

Anneli Meurman-Solin
University of Helsinki

1. Introduction

This article offers a corpus-based description of the use and development of the
progressive in Scots from 1450 onwards. The discussion is structured as follows.
I will start by briefly summarizing patterns of variation and change as reflected
in the frequencies and distributions of the variant forms of the present partici-
ple; nominal forms in -ing will be referred to when these allow us to under-
stand patterns of idiolectal variation (Section 2). The aim of this analysis is to
see which variants can be given paradigmatic status, and to provide evidence as
to whether these have specialized distributions as regards their use in various
functions. In this context I will also raise the question of whether the use of
large electronic databases will allow us to reconsider the criteria we should
apply to distinguishing between orthographic and phonological variation on the
one hand and morphological variation on the other. I will try to justify the need
to re-examine ways of categorizing variants in terms of grammatical systems by
illustrating specialized distributions of the various -ing forms in idiolects,
including those variants traditionally grouped together as reduced.
Secondly (Section 3), I will analyse the different types of the progressive,
providing evidence of the low frequency of the type be + preposition + a verb
form in -ing in texts illustrating the early use of the progressive in Scots.
Information on the frequency and spread of this type in Scots might allow us to
consider the possible influence of Celtic languages on the development of the
progressive in English. The debate concerning Celtic influence is motivated by
formal similarities with the type ‘copula + preposition/aspectual marker +
gerund’ in the insular Celtic languages (for a summary of the various hypothe-
ses, see Mittendorf & Poppe 2000: 117–119). I will also examine the use of the
active form in a passive sense. Finally, in Section 4 I will relate the use of the
progressive in Older Scots to types of discourse.
204 Anneli Meurman-Solin

The study is chiefly based on the Helsinki Corpus of Older Scots,

1450–1700 (HCOS; cf. Meurman-Solin 1995a, b);1 in addition, a number of
examples have been extracted from an expanded sample of the Criminal Trials
of Scotland (1561–1591), and from seventeenth- and early eighteenth-century
letters in the Corpus of Scottish Correspondence (CSC; cf. Meurman-Solin
forthcoming), being compiled at present.

2. Variation in form

2.1 -and and -ing

What do we know about the distributions of the main variants of the present parti-
ciple from earlier research? Devitt has examined the diffusion in Scots of what
she calls the Anglo-English form in -ing, summarizing her findings as follows:
With this high initial frequency of 62% -ing, the movement to 99% -ing by
1659 is less dramatic than the movement of the other features [the relative
clause marker, the preterite inflection, the indefinite article, the negative
particle]. (…) After initial periods of relative stability (the apparent decline
between 1520 and 1559 being non-significant), use of -ing increases very
gradually until 1600, when it virtually levels off. This pattern for the
(PresPrtcpl) variable differs significantly from each of the other variables, at a
level of < 0.0001.
Although this change occurs more gradually than the changes in the other
variables, it still represents a substantial move toward the Anglo-English
standard. The 62% use of -ing in 1520–1539 allows considerable variability
among texts (…). But such variability within texts becomes less common after
1600, when categorical usage of -ing becomes the norm (…) the levelling of
anglicization overall from 1600 to 1659 masks a new change: from variable
-ing usage to categorical -ing (…)
Certainly the shift from -and to -ing must have been well on its way during
the fifteenth century. The sixteenth century sees the increasing and almost
complete dominance of -ing, while the first half of the seventeenth century is
largely a period of refinement for this advanced change, of moving to categori-
cal use of -ing in individual texts, of -ing becoming the only possible form.
(Devitt 1989: 28–30)

The following observations can be made. The period of co-variation is relatively

long.2 As Devitt’s interest is exclusively in the relative frequencies of the variant
forms, there is no information about what lexical items or syntactic functions
 The progressive in Older Scots 205

adopt the incoming form earlier than others. Moreover, her database is not
sufficiently representative to provide usage information, for example, in more
private documents. I will illustrate the importance of such documents later in
this study.
According to Dons & Moessner (1999: 24), the frequency of present
participles ending in -and in the HCOS is relatively low in the progressive. The
absolute number of occurrences is 112, approximately 9 per cent of all uses of
this present participle form in the corpus. My somewhat lower figures (see
Table 1) can perhaps be explained by the application of stricter criteria, listed in
Denison (1993: 372–380), in the identification of ‘true progressives’.3
In Older Scots, specialized distributions of the variant forms of the present
participle as well as those of the gerund and the verbal or the deverbal noun (for
these labels, cf. Quirk et al. 1985: §17.54) can usually be attested on the idio-
lectal level, but in some idiolects there is some degree of interchangeability (see
example 5).4
I would like to stress the following methodological point. Instead of
immediately restricting attention exclusively to the morpho-syntactic feature
concerned, a thorough analysis of the orthographic and phonological system in
each idiolect is necessary. If an analysis of this kind is not available, neither variants
of inflexions in each morphological paradigm nor the correlation patterns
between each form and function can be identified.5 The scope of the present
study has only permitted illustration of this methodological principle by a small
number of examples in Section 2.3. A further important point is that the set of
relevant variants in regional varieties cannot be defined in terms of the generally
available grammatical descriptions of standard varieties, even though a suffi-
cient degree of equivalence or similarity is usually claimed in the literature.
A closer analysis of my corpora of Older Scots has shown that, beside
important differences between the dialect areas of Scotland, variants in -ing
were first introduced into linguistic contexts of a very specific kind. In the
HCOS, the variant cuming occurs exclusively in fixed phrases such as in (all)
tyme(s) cum(m)ing/cum(m)yng in pre-1500 texts; other early occurrences of
present participles in -ing have been attested in stereotypical non-finite
adverbial clauses with predicate verbs such as concerning, consedryng and
prowydyng, often used prepositionally, and in nochtwithstandyng. Other
contexts are according replacing accordand, exseding as an intensifying adverb,
evirlesting (god) for evirlestand in formulae, and the fixed expression God
willing. In fact the earliest use of -ing in a context different from those above
206 Anneli Meurman-Solin

has been attested in Lady Home’s mid-sixteenth-century letter, in which the

present participle sertifying occurs as a predicate verb in a non-finite clause.6
In addition to the items listed above, present participles in -ing of verbs
such as beseech and pray occur in letters, where the variant is used with a first-
person subject left implicit. It is interesting that variants in -ing are earlier in
this stylistically marked context:7
(1) Madame pleis zour grace / It is writin to me how zour grace hes gottin ye
eschete of william / edmistoun” qlk Is ueray proffetable vnto zour grace
becaus / all enteress and clame yt william edmistoun” allegit to haue of /
zour grace Is now Indowtit zouris Maist hummelie beseking / zour grace
notwtstanding ony laubouris of our contray partiis yt / yair be na part
yarof disponit to yame bot at all ye samyn / Remane haill In zour grace
handis qlkis doand Is zour grace honour and proffett and als zour grace
may ye mair esely / help zour seruitouris my bruyeris wyff & bairnis
(CSC 1545? Lord Methven NAS SP2/2: 102, Correspondence of Mary of
Lorraine, CII: 138)

As illustrated by this example, a writer may vary between variants in -and (e.g.
is doand) and -ing, but the latter is preferred in the marked context. Similarly,
the choice of praying instead of prayand is particularly frequent in final formu-
lae of the type praying God to conserve/preserve you, as in example 2:8
(2) praying god to preserff zur grace (CSC 1547 Sir Adam Otterburn NAS
SP2/2:136, Correspondence of Mary of Lorraine, CXXXVI: 188)

In the post-1570 texts of the HCOS, the present participle form -and in the
progressive is exclusively used in only six texts. There are no examples of this
form in the progressive in the seventeenth century. Example 3 illustrates a
sporadic use of the variant in an otherwise highly anglicized sermon by Bruce:
(3) and the Lord is beginnand to abstract his mercie and grace from this
Countrey, for the contempt of this quikning worde, quhilk hes sa clearlie
sounded heare (HCOS 1590–1591 Bruce, 25)

Table 1a–d presents the statistical findings. The reduced variants discussed in
Section 2.2 have not been included as they have primarily been attested in
another database, the Corpus of Scottish Correspondence. The be + in/a + -ing
type (11 examples in the HCOS; see Section 3.1) has also been excluded.
Table 1 illustrates the replacement of the present participle in -and in the
progressive, which was completed by 1640, the gradual increase of the progres-
 The progressive in Older Scots 207

Table 1. The progressive in the HCOS. Types be + -and and be + -ing.

Absolute numbers and mean frequencies per 1,000 words. N = number of texts per
period, collections of private and official letters being counted as one text each.
Table 1a. Period 1450–1500
Text -and Mean -ing Mean Total Mean

Acts of Parliament 2 0.07 2 0.07

Gilbert Hay 3 0.10 3 0.10
Dicta Salomonis 4 0.62 4 0.62
Porteous of Noblenes 1 0.27 1 0.27
Total (N = 10) 10 0.10 10 0.10

No occurrences in 6 (60 per cent) out of the total of 10 texts.

Table 1b. Period 1500–1570

Text -and Mean -ing Mean Total Mean

Acts of Parliament 3 0.10 3 0.10

Stirling Records 4 0.36 4 0.36
Peebles Records 8 0.40 8 0.40
Fife Sheriff Court Bk 2 0.23 2 0.23
St. Andrews Kirk S. 1 0.18 1 0.18
Criminal Trials 9 1.17 6 0.78 15 1.95
Compl. of Scotland 1 0.12 1 0.12
Lamb, Resonyng 1 0.15 1 0.15 2 0.31
Boece, History 4 0.18 2 0.09 6 0.28
Official Letters 12 0.34 2 0.06 14 0.40
Bible 2 0.45 1 0.23 3 0.68
Total (N = 17) 47 0.23 12 0.06 59 0.29

No occurrences in 6 (35 per cent) out of the total of 17 texts.

sive over time, clearly accelerated in the latter half of the seventeenth century,
and a high degree of heterogeneity between texts. With respect to the growing
frequency of the progressive in Scots, my findings largely coincide with those
reported by Elsness (1994) in his analysis of the progressive in the EmodE
section of the Helsinki Corpus of English Texts: he notes that the number of
progressive examples rises from 33 (mean frequency per 1,000 words = 0.17) in
the period 1500–1570 to 52 (= 0.27) in the period 1570–1640 and to 100
(= 0.58) in the period 1640–1710.
As shown in more detail in Section 4, in the sixteenth century the highest
frequencies for the progressive have been attested in trial proceedings, but
208 Anneli Meurman-Solin

Table 1c. Period 1570–1640

Text -and Mean -ing Mean Total Mean

Acts of Parliament 1 0.02 2 0.04 3 0.06

Stirling Records 1 0.10 1 0.10
Huntar, Handbook 2 0.29 2 0.29
Skeyne, Pest 1 0.13 1 0.13 2 0.27
Basilicon Doron 1 0.05 1 0.05
Bruce, Sermon 1 0.16 1 0.16
Row, Sermon 4 1.63 4 1.63
St. Andrews Kirk S. 5 0.30 5 0.30
Criminal Trials 12 1.21 12 1.21
Roy Trial 1 0.26 1 0.26
Wishart Trial 1 0.26 1 0.26 2 0.51
Fowler 1 0.17 1 0.17
Lesley, History 4 0.38 4 0.38
Pitscottie, History 3 0.28 3 0.28
Moysie, History 3 0.28 1 0.09 4 0.38
Lithgow, Travelogue 4 0.33 4 0.33
Melville, Diary 10 0.68 10 0.68
Birrel, Diary 8 0.63 8 0.63
Johnston, Diary 28 3.02 28 3.02
Lesley, Diary 1 0.12 1 0.12
Private Letters 2 0.14 4 0.27 6 0.41
Official Letters 1 0.09 1 0.09
Total (N = 30) 31 0.10 73 0.24 104 0.34

No occurrences in 8 (27 per cent) out of the total of 30 texts.

attention can also be drawn to the mean frequencies in Sir Archibald Johnston
of Wariston’s diary, dating from 1632–1639, and James Row’s sermon, the so-
called Pockmanty Preaching, which has been described as “exceptionally
informal and vernacular in its style” (Aitken 1978: 102). Table 1(d) shows
consistently higher frequencies for texts representing the narrative text category
(i.e. histories, biographies, travelogues and diaries), and confirms the finding
that a generally greater number of occurrences of the progressive can be attested
in trials.9 In Sinclair’s scientific treatise, sections describing the so-called
‘experiments’ are primarily expository, those recording ‘observations’ being
narrative (Meurman-Solin 1993: 92–93).

2.2 ‘Reduced’ variants

 The progressive in Older Scots 209

Table 1d. Period 1640–1700

Text -and Mean -ing Mean Total Mean

Acts of Parliament 2 0.05 2 0.05

Stirling Records 3 0.31 3 0.31
Aberdeen Records 5 0.29 5 0.29
Skene, Handbook 1 0.62 1 0.62
Reid 1 0.09 1 0.09
Sinclair, Science 17 1.12 17 1.12
Sinclair, Narratives 14 1.78 14 1.78
Welsh, Sermon 4 0.90 4 0.90
Standsfield Trial 15 1.81 15 1.81
Presbyterian Eloquence 6 0.70 6 0.70
Apology for Clergy 2 0.24 2 0.24
Spalding, History 19 1.23 19 1.23
Prince of Tartaria 2 0.60 2 0.60
Lauder, Journals 16 1.63 16 1.63
Lamont, Diary 9 1.29 9 1.29
A. Brodie, Diary 6 0.49 6 0.49
J. Brodie, Diary 16 2.42 16 2.42
A. Hay, Diary 11 1.10 11 1.10
Somerville, Biography 10 1.36 10 1.36
J. Turner, Biography 3 0.39 3 0.39
Private Letters 17 0.93 17 0.93
Official Letters 3 0.41 3 0.41
Total (N = 23) 182 0.74 182 0.74

No occurrences in 1 (4 per cent) out of the total of 23 texts.

There are texts in the HCOS where the variants -in(e), -yn(e), -en(e) occur in
verbal nouns as well as present participles, as illustrated below. The variant of
the present participle failing in the fixed phrase failing thereof is failyene in
example 4:
(4) The inqueist ordanis the baillies gang vesy the briggis and calsayis thairof
gif thai be sufficient vphalding (‘be held up’) be the dichtaris as thai
promist, and to caus thame mend all failyeis sufficient, quhilkis beand
mendit (‘being mended’) to se gif thai will vphald the samin, and
failyene (‘failing’) thairof to discharge thame and cheis vtheris in thair
places.
And als ordanis Andro Wychtman to caus fill vp the holis with erd or
gravall quhair he tuk the clay till his kill bigging (‘kiln building’), incon-
tinent, becaus the fludis ar cummand (‘are coming’) on hand. (HCOS
1555 Peebles B. Rec., 218)
210 Anneli Meurman-Solin

In A Dictionary of the Older Scottish Tongue (s.v. failzeand / failzeing of), the
variant failyene is grouped together with variants of forms ending in -ing, and
an example (dating from 1572) is quoted from the same source as example 4.
However, there is not yet sufficient evidence to justify a grouping of this kind
since further corpus-based research may show that the forms -in(e) or -en(e)
have an independent status as variants in a pattern of complementary distribu-
tions, and that they are systematically preferred in specific functions in certain
idiolectal and/or geographical varieties of Scots (in this connection, cf. Sec-
tion 2.3, examples 7 and 8). Their occurrence may be conditioned by features
of individual lexical items, spelling conventions, function, genre-specific
practices, or pronunciation. Moreover, as also illustrated by example 4, a careful
reconstruction of relationships between a number of other variants (the past
participle vphalding, instead of vphaldin, the use of beand mendit in an absolute
clause, the verbal noun in the compound kill bigging, the Present Continuous ar
cummand) should be included in the descriptive account.
For a closer examination of formal variation of this kind, a larger corpus of
data extracted from diplomatic editions or manuscripts will have to be analysed;
the HCOS has allowed us to identify local records and letters as particularly
important sources for such a study.

2.3 Distributions of variants in idiolects

As stressed above, a statistical analysis of the so-called reduced variants is not
presented because of their relative infrequency in the HCOS. However, it is
noteworthy that in the data analysed so far, the prevailing trend is that the
variants of the present participle are not used interchangeably. Yet it is possible
to find idiolects such as the one illustrated in example 5:
(5) My dearest Sister I am impatient to hear how dear Mary is. I wou’d have
sent last week but was expecting every day to hear from yow, becaus yow
told me yow wou’d send. I am now very feard she is worce that I have
not heard from yow. I am wearie with wreten before I am well begune to
yow, for I have been wreten to Leslie. My mother tells me that her dear-
est cusen, Montrose, is dying; (…) Your bairne is very well and going
alone. (HCOS c. 1697 Margaret, Countess of Northesk, to her sister,
Countess of Leven, Memorials of the Family of Wemyss of Wemyss,
Vol. 3: 158)
 The progressive in Older Scots 211

This writer has a grammar of the following kind: -ing in the Present and Past
Continuous (was expecting, is dying, is going), -en in the Present Perfect Contin-
uous (have been wreten ‘have been writing’) and as the gerund form of an
irregular verb (wearie with wreten ‘weary with writing’).
Another pattern of variation can be illustrated by a passage from Sir James
Melville’s Memoirs:
(6) sche (…) askit how I cam ther. I said, as I was walken with my L. of
Hundsden, as we past by the chamber dur, I hard sic melodie, quhilk
rauyst and drew me within the chamber I wist not how; excusing my falt
of hamelynes, as being brocht vp in the court of France, and was now
willing to suffer what kynd of punissement wald pleise hir lay vpon me
for my offence. (HCOS 1610 Sir James Melville of Halhill, Memoirs of his
own Life, 1549–1593, 124)

The prevailing present participle form is in -ing, but in this passage reporting a
dialogue (cf. Section 4), the form walken (‘walking’) is chosen in the Past
Continuous.
Letters are a rich source in the study of the progressive. The number of
variants in -in or -ine, or -en, -ene seems to be on the increase as a larger corpus
of letters becomes available. Patterns of variation may reflect a relatively high
degree of systematicity. This is illustrated in examples 7 and 8. Example 7 is a
letter by Alexander, sixth Earl of Eglinton, to his spouse:
(7) My hert my most louing deutie remem/beritt to zou and all zour
coumpanie vithe / my seruis to my lady our mother, I haue / bein this
day seine my lord our brother and / sister, vho hes bein bothe very euille
atteise / and is mendin very vekly, I did neuer sie / any of them in the
fasoun, I houp the vorst / of bothe ther seiknes is past; I haue meid zour /
excus to them and thy did exsep it most gledly / for thy say, thay do knau
that ze ar not ebeill / to trauille as ze ar for the present. I am beissie /
doine my tourins that I may deispache my / self homuard to zou so soun
as I can (…) as for neuis I haue non bot that my lady / sempeill is
deleyuer of ane docheter: sue wissing / zou all helthe and happines I rest.
Zour most louing / housbande till dathe / Ponoun 21 iune / Eglintoun
(CSC 1613? Alexander, sixth Earl of Eglinton, to Anna, Countess of
Eglinton, NAS GD 3/5/52, Memorials of the Montgomeries, 30:187–188)

It is relevant that the only instances of the variant in -ing occur in marked
contexts in formulae, namely the optative verb wissing ‘wishing’, in a non-finite
clause with an implicit first-person subject, and the participial adjective louing
212 Anneli Meurman-Solin

‘loving’. Variants in the progressive end in -in or -ine (haue bein seine ‘have
been seeing’, is mendin ‘is mending’); there is a gerund form in am beissie doine
‘am busy doing’, and the deverbal noun tourins ‘tourings’. The Earl seems to
have a richer paradigm of forms, since, beside the Present and Past Continuous
forms prevailing in contemporary texts, he uses a Present Perfect Continuous
I haue bein this day seine ‘today I have been seeing’.
One of his younger sons had a military career. In this son’s letter, written in
Paris, the present participles and verbal nouns end in -in, but the form in the
progressive is -ing. His spelling system differs from that of more trained writers,
and some of the variants have therefore been explained by giving a Present-day
English equivalent (cf. Meurman-Solin 1999):
(8) I am in peries (‘Paris’) and hes be goun (‘begun’) my exercies / to fenes
(‘fence’) and danes (‘dance’) with your lo (lordship’s) oled (‘old’) mester
angles / and that my fencin and dansin extendes monthli / to 25 lib~ 10
soues and my mathamatikes monthli to 8 lib~ 10 s~ / bot con cernin
(‘concerning’) my reyeddin (‘riding’) my bririn (‘brethren’) uil (‘will’)
not let / me begin it uil (‘while’ in the sense of ‘before’) I hef dereksion
fre (‘from’) your lo / uich (‘which’) I expec day bifor and that aer (‘are’)
all the / girrer (‘together’) in pencion and peyes (‘pays’) fortin crouenes /
in moneth for bay (‘buy’) mani extreordineres and / at my ariuel in
perris I touk of a sout (‘suit’) of cloth / of sil uher (‘where’) of ther aer
four elles for clothes and / kassak uith four elles of pax to leyen (‘line’)
my kassak / uith uhich med (‘made’) en end of the monnies uhich I / had
restin (…) bot I am expecing dayli for comanscien[?] (‘comission’) /
{verso} fre your lo bot to tel yowr lo truli uer (‘we are’) all / louking errer
(‘either’) for monnies or elis (‘else’) for a letter / uher your lo thinkes to
fornis (‘furnish’) ous (‘us’)… praying to god to continou your lo in all
hellth and hapines (CSC 1633 Alexander Montgomery NAS GD 3/5/192,
Memorials of the Montgomeries, 230)

It seems relevant that the writer is consistent in the choice of variants; the
variant ending in -ing is exclusively used in the progressive and in the final
formula, whereas forms such as fencin, dansin, concernin, reyeddin, restin
(‘fencing, dancing, concerning, riding, resting’) are used in other functions.
It seems a corpus of approximately one million words is not sufficiently
large to provide conclusive evidence of factors conditioning the choice of
variants; moreover, the generally low frequency of the progressive forms in pre-
eighteenth-century texts does not allow a statistically significant account of
 The progressive in Older Scots 213

formal variation (cf. also the conditioning of discourse type discussed in

Section 4). In my view, the examples of forms ending in -in(e) or -en(e) given
above cannot be exclusively explained by referring to the system of phonetic
spellings often adopted by these writers. Instead, they may consciously distin-
guish between, for instance, -in(e) or -en(e) and -ing, using the latter in particu-
lar contexts. That the Earl of Eglinton should choose -ing exclusively in formu-
lae in his early letters may suggest change from above, i.e. a practice consciously
adopted from prestigious language use (cf. Meurman-Solin 2000b).10 As illustrated
above, in present participles in the progressive he uses the variant -in(e).
I would thus like to suggest that the form in -ing was perhaps consciously
selected by the above informants as a more prestigious incoming variant, and
was typically introduced as part of a stylistically marked discursive practice (cf.
Meurman-Solin 2000a, b). It is evident that the co-occurrence patterns of the
different variants are worth studying in great detail. The study of correlation
patterns between form and function in varieties of English can shed light on
whether variants previously marginalized as irregular or sporadic may be
viewed as candidates for paradigmatic status in the grammar of a specific
regional or local variety. At the same time, however, we need not reject the
hypothesis of there also being variation resulting from the use of phonetic
spellings and hypercorrect spellings (cf. Meurman-Solin 2001b). Comparisons
between idiolects are important; example 9 is particularly interesting as it
exhibits an explicit intertextual link between a and b:
(9) a. Caus buy ane pund of razines to give hir by the way. Also caus buy
ane pair of rucht mittenes (CSC 1642 Alexander, sixth Earl of
Eglinton, to Henry Seton, Memorials of the Montgomeries,
Vol. 1: 256–257)
b. Item, for ane pond of resings 00 lib 6 s
Item, for ane payr off rucht metting and twa to the bairne 00 lib 6 s
(ibid., account by Henry Seton)

The Earl of Eglinton asks his merchant to send him, among other things, raisins
(razines) and a pair of mittens (mittenes), but Henry Seton, his Edinburgh
merchant, writes resings and metting in the account annexed to the letter.

3. Variation in type

3.1 be + preposition + -ing

214 Anneli Meurman-Solin

The constructions with a preposition are formally similar to the insular Celtic
types involving the copula + preposition/aspectual marker + gerund (for a list
of examples, see Mittendorf & Poppe 2000:117), which has produced the hypo-
thesis of Celtic influence. In the literature, it is also generally hypothesized that the
collocate be + preposition + gerund influenced the development of the English
progressive (cf. Mustanoja 1960:587ff.). However, before anglicization develop-
ments in Scots may have been different since the present participle had a distinc-
tive form in -and, which was only very rarely used in nominal functions.11
The findings indicate that both the be + in + -ing and the be + a + -ing
types are infrequent in the HCOS and several contemporary family correspon-
dences, including a selection of early eighteenth-century letters. Variants in
-in(e) or -en(e) have not been attested in these types in the present data. There
are altogether 10 examples of the be + in + -ing type in the HCOS. While this
corpus yields only one example of the be + a + -ing type, five instances have
been attested in the still incomplete CSC. The dates of the examples are very
interesting, especially as, in addition to one example in Sir John Lauder’s
Journals, dating from 1665–1676, the be + a + -ing type has only been attested
in post-1670 letters.
Examples 10–15 illustrate the be + in + -ing type:12
(10) and spak sum quh[ ]t” with dene john~ pencher for apair of Endenturis
yat wos in makyng betuex ye said priour & ser alexander of hvim~ of xv
yher tak of alcambos (ECOS 1442 Durham, Dean and Chapter
Muniments: Locellus, XXV.6)
(11) Item, sancte Ierom~e sais that till here mes~ with clen hart and gud
dewocioun~e garris the saulys that he prays for feil na payne in purgato-
ry quhil that mes~ is in doinge. (HCOS c. 1460 Vertewis of the Mess, 192)
(12) Quhill thir actis war in doing, Ethod, king of Scottis, had trubill in his
realme. (HCOS c. 1533 The Mar Lodge Translation of Boece’s History of
Scotland, 283)
(13) Quhen thir things was in doing in Scotlande, Edwarde the sixt King of
Ingland (…) departit out of this lyf (HCOS 1570 Lesley, History of Scot-
land, 248)
(14) and farder salang as it salhappin the said wode to be in cutting and
quhilk salhappin to be transportet and carrit be thame throw this burgh
to the schoir thairof, for payment of 100 merks. (HCOS 1603 Stirling B.
Rec., 107)
 The progressive in Older Scots 215

(15) The actioun anent the discussing of the richtis of the chantorie is in
doing now, and I hawe wryttin (HCOS 1616 Sir Alexander Gordon of
Navidale, Sutherland Book, 126)

The collocate is usually be + in making or be + in doing, but to be in cutting ‘be

being cut’, occurs in example 14. In examples such as these the structure has
been interpreted in the literature as having a passive sense. There are seven
examples of this type in the HCOS.
The following is an early example of be + in + -ing with an active sense:
(16) this Turisday my lord Borthik cuming frome the cardenall till his awn
hows I was in huntyng (CSC 1544 Sir George Douglas (non-autograph)
NAS SP2/1: 87, Correspondence of Mary of Lorraine, LXXXVII: 104–105)

In addition, three examples in the active have been attested in George Sinclair’s
scientific treatise. These can be illustrated by example 17:
(17) But while the Coalhewers were in digging down, and had come the
deepness of 13 or 14 fathom, they were stopped from working by
Damps, or ill Air (…) (HCOS 1672/1683 Sinclair, Hydrostaticks, 197)

In the HCOS and CSC data, the be + a + -ing type is even less frequent than the
type with the preposition in. Moreover, while the type with in has been attested
as early as 1450–1500, the first subperiod in the HCOS, the first occurrence of
the be + a + -ing type is as late as the mid-seventeenth century.
(18) for I doe confes or (‘before’) now ye micht have doin quhat ye ar now
adoeing. and quhat ye have now doin is with expressioun of kyndnes and
schew of tendirnes. (CSC 1671 Margaret Kennedy, Lady Kers, Memoirs
of the Maxwells of Pollok, Vol. 2: 319)
(19) And or (‘before’) my sones fitt uas could or his br[e]ath out, shoe uas a
working hir beasse ends and calumniatting me wher euer shoe went
amongest hir oune creue, lick ane ungriatt false doughter. (CSC 1671
David, second Earl of Wemyss, Memorials of the Family of Wemyss of
Wemyss, Vol. 3: 113)

Examples 20–22, which have been attested in early eighteenth-century women’s

letters, suggest that the be + a + -ing type may indeed be a relatively late
development in Scots:
(20) I dare say noe more for feare of loosing the post, for they are just come
to tell me that it is just a going, and I have my nephew’s letter to writ, so
216 Anneli Meurman-Solin

must end writing (CSC 1717 Winifred, Countess of Nithsdale, Book of

Carlaverock, 258)
(21) My husband was the other day a saying that that part that William
A[l]ves had was now free (CSC 1720 Winifred, Countess of Nithsdale,
Book of Carlaverock, 315)
(22) as for the peadmony we are a doing, sister Lucy and I, we have made two
pursses, one for my Lord, another for your Ladyship (CSC 1714 Lady
Anne Stuart, Book of Carlaverock, 187)

The occurrence of the type be + a + -ing in letters may perhaps be seen to

reflect discourse features shared with oral narratives (cf. “Others say that” in
example 23).
The only example of this type in the HCOS has a passive sense ‘was being
built’:
(23) Others say that when the Church was a bigging, the Devill appeared to
one of the maisons (HCOS 1665–1676 Lauder, Journals, 56)

Unfortunately, it is not possible to comment on the hypothesis of Celtic

influence on the basis of just 16 examples. For this to be possible, we would
need an even larger database enabling us to verify the chronology of the be +
preposition + -ing type in Scots. Since such an aim is notoriously difficult to
achieve in the early periods, we may have to resort to the concept of ‘a notional
corpus’ used by Mittendorf & Poppe (2000: 127) to refer to an awareness that
whatever data has been looked at so far is a fraction of what we would have to
examine to say anything conclusive about Celtic influence. Their comments on
problems besetting the study of Middle Welsh are revealing. They point out that
no digitized database of Middle Welsh texts is available, and that “[m]any
questions concerning the dates of Middle Welsh texts and the diachronic and
geographic stratification of Middle Welsh are still unresolved”. In this situation,
they find it appropriate “to treat every single text initially as a separate sub-corpus
of the notional corpus of Middle Welsh texts, in the hope to capture idiosyncra-
cies in the use of the periphrastic construction relating, for example, to chronol-
ogy, genre, and register and to detect the degree of its grammaticalisation”.

3.2 be + -and in pre-1570 texts

According to Devitt (1989: 29), “the shift from -and to -ing must have been
well on its way during the fifteenth century”. Yet as shown in Table 1(a) above,
 The progressive in Older Scots 217

it is significant that in pre-1500 texts of the HCOS only the variant in -and is
used in the formation of the continuous tenses. All in all, these verb forms also
remain infrequent in the sixteenth century except for a few genres. If they
occur, the forms are chiefly used in the Present or Past Continuous; there is
only one instance of a Present Perfect Continuous in the Acts of Parliament:13
(24) And yt merchand & vyerris of ye Realme / has bene sekand Justice at ye
king of france & at his consale / (HCOS 1482 Acts of Parliament, 144.C2)

The type being + -ing (e.g. being lying) had some currency in English from the
mid-sixteenth century to the early nineteenth (Denison 1993:394). The first
occurrence of its equivalent in Scots is from the early sixteenth-century Stirling
records:
(25) Robart Spettal allegit and offerit to prefe (…) that [he] com to the said
Gilbert, he beand liand on hus (‘his’) deid bed (HCOS 1527 Stirling B.
Rec., 29)

The verbs in the progressive represent a wide range of semantic categories, but
as there are only sporadic examples of each category, a detailed quantitative
presentation of the findings is not useful. Even though there may be a higher
frequency of intransitive verbs, and verbs expressing state (such as lie) or motion
(such as come and go), there are also transitive verbs of action.14 The following
examples illustrate the occurrence of the various semantic categories of verbs:
intransitive: motion
(26) Madame pleis zour grace george Dowglas was not heyr bot ye lard of
blakater / is cumand to zour grace (CSC 1545? Thomas, Commendator
of Dryburgh NAS SP2/2: 105, Correspondence of Mary of Lorraine,
CV: 142)
(27) Mj lord I commend mj seruice to zour l (‘lordship’) / for surte ye armye
off ingland is cumand be sey (CSC 1546 James Cockburn of Langtown
NAS SP2/2: 122, Correspondence of Mary of Lorraine, CXXII: 170)
(28) And this nycht quhen I wes passand to my supper I ressauit ane writting
fra the Lard of Eistir Wemys (HCOS 1573 David Wemyss of Wemyss,
Memorials of the Family of Wemyss of Wemyss, Vol. 3: 70–71)
intransitive: state
(29) he Is liffand as ane deid man (HCOS 1490 Porteous of Noblenes, 180)
intransitive: action
(30) the metis yat ar etyn jn the dyner’ ressauis the hete of the day jn mannis
corps . quhen he is wakand and trauailand / and bathe the membris of
218 Anneli Meurman-Solin

mannis body / and his witt is than vext and trauailit / (HCOS 1456 Hay,
Prose MS, 95)
(31) Thir thingis war done in Bethany beyond Jordan, quhare Johnne was
baptysand. (HCOS c. 1520 St John I: 28)
And Johnne was baptizand in Ennon, beside Salem, for mony watris war
thare; and thai com, and ware baptizit. (ibid. III: 23)
(32) Item he sais that al the vismanis wyt is in his mouth, and thinkis that he
has neuir yneuch of It and euir is techand and lerand… (HCOS ?1460
Dicta Salomonis, 185)
(33) yaj ar / aduertess yat zour partye is gadderand and purposis to be in
styrlinge on frydaye (CSC 1543 Richard Kincaid NAS SP2/1: 12, Corre-
spondence of Mary of Lorraine, XII: 15)
(34) madame I belewe zur graice has hard how all yis cuntreth wes brekand /
hed not bein my haiste haym~ cumin … (CSC 1544 George, Earl of
Huntly NAS SP2/2: 92, Correspondence of Mary of Lorraine, XCII: 118)
transitive: relational
(35) Item he sais, gret lordis ar quhilum hafand powar our mony pupile (…)
(HCOS ?1460 Dicta Salomonis, 188)15
transitive: action
(36) Mademe, eftere humle commendatione of my service to your grace, I
haif bene doand part of besynes and service consernyng your grace (CSC
1545 Elizabeth, Countess of Moray (non-autograph), Correspondence of
Mary of Lorraine, 149–150)
(37) Ples zour grace sen~ my cumyng in yir partis I haiff ben” doand dili-
gence as I myt best do to haiff / had my lord to zour g (‘grace’) effect &
plesur … (CSC 1559 Archbishop Hamilton NAS SP2/4: 280, Correspon-
dence of Mary of Lorraine, CCLXXX: 424)

The examples listed above suggest that a relatively high degree of interactive-
ness, or addressee-orientedness, triggers the use of the progressive; this is
illustrated in reporting news to the addressee of a letter, giving advice in an
instructive text such as Hay’s prose works (example 30), and in reported speech
(note “Item he sais” in examples 32 and 35).
Progressives of verbs other than those of motion or state represent 44 per
cent of the occurrences of the be + -and type, and these become more frequent
in the latter half of the sixteenth century. The sixteenth-century burgh records
of Peebles are a good source for tracing their spread. It is interesting that all the
instances in this source should occur in modifier or dependent clauses, a
 The progressive in Older Scots 219

finding that converges with the results obtained by Strang (1982: 441) in her
analysis of the use of the progressive in narrative prose of the first half of the
eighteenth century; she notes that “the construction is truly at home only in
certain types of subordinate clause, especially temporal, relative or local”. In
contrast, Wright (1994) has found that the percentage of main clause progres-
sive occurrences in a corpus of sixteen plays (prose comedy) dating from
1670–1710 ranges from 55.5 per cent, in Congreve, to 78 per cent, in
Wycherley. This suggests that the frequency of the progressive in main and
subordinate clauses may vary in different genres.

3.3 be + -ing in 1500–1700

The continuous tenses with a present participle ending in -ing are infrequent
before 1570 in Scots, both in general and particularly as compared with the occur-
rence of continuous tenses with present participles ending in -and. It is relevant
that among 15 texts, some of which are compilations, and letters by 18 writers, only
three texts and one letter writer use both variants of the present participle in the
progressive in the period 1500–1570. Moreover, of these three texts, the scotticized
Bible and William Lamb’s Resonyng of ane Scottis and Inglis Merchand can be
expected to have a mixed system (cf. fn. 6). Example 31, also extracted from the
New Testament, can be compared with the following example:
(38) And I knew him nocht, bot that he be schewit in Jsrael, tharfor I am
baptizing in watire. (HCOS c. 1520 St John I: 31)

Adam Otterburn’s idiolect could be taken as an example of the high degree of

variation in the mid-sixteenth century:
(39) Forder yair Is ane flemyng schip yt wes tane be franchemen Reddy to /
departe in flandris and ye merchandis of Dunde and yis tovne hes thot /
Ryt hevy yt yair nytbouris ar lyand in presoun~ in flandris & ar abill to
periss & de in presoun~ wtout help (CSC 1544 Adam Otterburn NAS
SP2/1: 72, Correspondence of Mary of Lorraine, LXXII: 92)
(40) as for tydingis heir I see / na thing bot as ye ald kyng wer levyng & / Ilk
day I heir~ of our infelicite (CSC 1547 Adam Otterburn NAS SP2/2: 130,
Correspondence of Mary of Lorraine, CXXX: 180)

In the HCOS there are two instances of be + -ing in the extract from Bellen-
den’s translation of Boece’s history (a sample of c. 22,000 words), six in
Criminal Trials (c. 7,700 words) and two in the Correspondence of Mary of
220 Anneli Meurman-Solin

Lorraine (c. 31,000 words). Compilations such as the Criminal Trials are often
linguistically heterogeneous: the two variants co-occur but the forms in -ing are
still less frequent.
(41) The keyis of the lugeing wes partlie standing in the durris, and pairtlie
deliverit to this deponir be Robert Balfour, awnir, all except the key of
that dur, quhilk passit throuch the sellare and the town wall (HCOS
1567–8 Criminal Trials, 501)

Nurmi (1996:164–165) has analysed the period 1590–1620 in the Corpus of Early
English Correspondence, and be + -ing also remains infrequent in her data.
Considering her findings and the distributions in the Scottish texts, it seems
even more significant that in the database of Scots speech-based texts such as
recordings of trials should contain a relatively higher frequency of this feature.
Examples of the be + -ing type in a passive sense do not occur in pre-1570
texts. This could simply be a consequence of the paucity of examples before that
date; it deserves notice, however, that examples with a passive meaning occur
from very early (see examples 11–13 above) in the case of the be + preposition
+ -ing construction, despite the fact that this, too, is uncommon (only 11
examples in the HCOS; see Section 3.1).
After 1570, be + -ing is however found in a passive sense,16 as in, for
instance, Jean Ross’s letter dating from 1640:
(42) a. I haue noe tym to wrytt to your la (‘ladyship’) of / thos things that ar
doing heer
b. bot latt me kno particullarlie / what is doing in the lords cause and
how your / la is in your health
(CSC 1640 Jeane Ross, Lady Innes (wife of Sir Robert Innes of that
Ilk), to Margaret Lady Ross, her mother, NAS GD 3/5/244, Memori-
als of the Montgomeries 114: 242)17

The use in the passive has also been attested in a number of other texts in the
HCOS:
(43) Quhen thir thingis war preparing, the Erle of Huntlie caused take
Williame M’Kinlocke (HCOS 1570 Lesley, History of Scotland, 235)
(44) And further Depones, that when the Defunct’s Body was bringing up to
the House, the Deponent would have had him brought to his own
Chamber (HCOS 1688 Standsfield Trial, 18)
(45) quhairby he had daylie intelligens what wes doing in court or abroad
(HCOS c. 1650 Spalding, History, 59)
 The progressive in Older Scots 221

To conclude this section on variation in type, let me point out that, despite the
variation between -and and -ing in other functions of the present participle in
the pre-1570 texts (see Section 2), its prevailing form in continuous tenses in
that period is in -and (80 per cent between 1500–1570). This could suggest that
instead of the merger of be + present participle (be doing) and be + preposition
+ gerund (be on/a doing) that has at times been suggested in the literature for
the English progressive, two parallel, separate developments can be traced in the
North: one involving the verbal type, with -and being gradually encroached
upon by the Anglo-English participial form in -ing, and one involving the
nominal type with an intervening preposition. It is also possible, though further
research is needed to confirm this, that the nominal type with the preposition
in (see Section 3.1) may have provided a structure with a passive sense (i.e. be
in doing for ‘be being done’) at a time when the verbal type did not yet have a
passive transform.

4. Narratives and speech-based texts

Examples 46–52 highlight two specific environments in which the use of the
progressive can be related to genre-specific features. In proceedings of trials a
diversification of the discoursal properties of the recordings took place. The
sixteenth-century Scottish trials range from the recording of decisions made, as
in the Sheriff Court Book of Fife, to detailed recording of the process that took
place before decisions could be made, as for instance in the Criminal Trials of
Scotland and St Andrews Kirk Sessions. This change is reflected in various
degrees of syntactic complexity; on the one hand, a high frequency of passive
constructions and nominalizations of verbal processes; on the other, a high
frequency of finite clauses with predicate verbs in the main clauses in the
Simple Past and those in the dependent and modifier clauses in the Past
Continuous. On this continuum less syntactic complexity correlates with the
frequency of direct and reported speech in the recordings, and with a higher
frequency of the progressive.
Let me illustrate this evolution by discussing extracts from the Criminal
Trials of Scotland. In the following prototypical example, the Past Continuous
occurs in an adverbial clause of time in a narrative where a sequence of events
is related in the Simple Past:
222 Anneli Meurman-Solin

(46) Ansuerit, Remembring hir, quhen sche was lyand in chyld-bed-lair, with
ane of hir laiddis, that ane stout woman com in to hir, and sat doun on
the forme besyde hir, and askit ane drink at hir, and sche geif hir; (HCOS
1576 Criminal Trials, 56–57)

In example 47 detailed recording combines free indirect speech with some

direct quotes, and with features such as the use of the progressive:
(47) (3.) ITEM, Being interrogat, how and in quhat maner of place the said
Thome Reid come to hir? Ansuerit, as sche was gangand betuix hir awin
hous and the yard of Monkcastell, dryvand hir ky to the pasture, and
makand hevye fair dule with hir self, gretand verrie fast for hir kow that
was deid, hir husband and chyld, that wer lyand seik in the land ill, and
sche new rissine out of gissane. The foirsaid Thom mett hir be the way,
healsit hir, and said, ‘Gude day, Bessie;’ and sche said, ‘God speid yow,
gudeman.’ ‘Sancta Marie’, said he, ‘Bessie, quhy makis thow sa grit dule
and sair greting for ony wardlie thing?’ Sche ansuerit, ‘Allace! haif I
nocht grit caus to mak grit dule? For our geir is trakit; and my husband is
on the point of deid, and ane babie of my awin will nocht leve; and my-
self at ane waik point; haif I nocht gude caus thane to haif ane sair hart?’
(HCOS 1576 Criminal Trials, 51–52)

Given the hypothesis of the spoken origin of the progressive in the literature
(see, for instance, Dennis 1940), it is noteworthy that in the present data the
progressive form occurs relatively early in direct speech. Examples 48–50
illustrate direct speech recorded in trials:
(48) After the quhilk I never spake to the said Erle of it quhill the day he gate
his Assyse; quhaire the said Erle standing at the barr, luiking doun and
sad lyke, I plukit upon him and said, ‘Fye! my lord, what Divill is this yee
are doand! Your face shawes what ye are! Hald up your face, for Godis
sake, and luik swa and ye were gangand to the deid!’ (1567–1568 Crimi-
nal Trials, *512)
(49) Item. Deponis, Quhan the deponar and Pat Willson come to the Frier Zet
with the last convoy, and laid the same down, Robert Ormestoune come
furth, and said thir words. ‘This is not gude like! I trow this purpois will
not come to this nycht! I will in and se quhat yai are doing.’ (HCOS
1567–1568 Criminal Trials, 494–495)
 The progressive in Older Scots 223

(50) yis deponar said to Pat Willson, at the conveying of the last carriage, thir
words, ‘JESU, Patt, quhattin ane gaitt is yis we are gangand? I trow it be
not gude!’ (HCOS 1567–1568 Criminal Trials, 494)
(51) And my lord thocht lang, and speirit ‘Gyf yair was ony part of the house
yat they mycht se the lunt, gyff it was burnand anouch?’ (HCOS
1567–1568 Criminal Trials, 499)

There is only one example of a progressive form in the extract from St Andrews
Kirk Sessions dating from the mid-sixteenth century (c. 5,400 words), but it is
interesting that this should occur where the spoken deposition of a witness is
being recorded.
(52) Thomas Myretoun (…) depones that he knawis nathing in the causs of
adultery, nor of the crymes conteynet in the clame, except that in De-
cember twa yeir bypast, or therby, the deponar remembyris in the said
Williames awin howss, quhare James Rutherfurde dwelles instantlie, that
he and his wyfe wes wrasland togiddir, and in the meyntyme that sche
bait him in the arme, quhilk he belevis wes nocht of malice, Margaret
Lawsone beand present in the chalmer, plus nescit, etc. (HCOS 1559 St
Andrews Kirk Sessions, 21)

Mustanoja (1960: 584–585) points out that the progressive gives a graphic
quality to a text. Beside a durative colour, “not infrequently emphasized by the
addition of an adverb meaning ‘always’”, he suggests that “[i]n numerous
instances” the progressive form may be preferred “because it has a greater
descriptive force, i.e. it makes the narrative more graphic (…) or simply because
it is more emphatic”. The concept of a graphic presentation appropriately
depicts variation between the progressive and other finite and non-finite verb
phrases in the recordings of trials.18 Further study will have to be based on a
large database of trial proceedings, but even in this limited sample the effect of
the recording style on the frequency and type of progressive verb forms suggests
some systematic patterning. The more graphic the story, the higher the frequen-
cy of such structures as finite adverbial clauses of time and nominal
that-clauses. The more detailed the recording, the higher the frequency of the
progressive. Although the general frequency of the progressive forms is quite
low in the period, it has thus been possible to identify discourse properties of a
genre that trigger the choice of this specific linguistic feature (cf. Meurman-
Solin 2000a, b, 2001a). In earlier research on the evolution of genres, there is no
statistically significant evidence yet of co-occurrence patterns that would
suggest a straightforward correlation between syntactic complexity in general,
224 Anneli Meurman-Solin

or specific clause types in particular, and the progressive (cf. Biber

1988: 229–236).
To conclude, that a number of sixteenth-century instances of the progres-
sive in Scots have been attested in proceedings of trials can be interpreted as
related to the written text being based on either speech or script, and probably
both. As the testimony is not included verbatim in the proceedings, it may not
be appropriate to refer to the genre of court testimony as a conditioning factor;
we should rather talk about the way of recording a process in which both
written documents and spoken testimonies are presented and become part of a
new written document representing a mixture of discourses. The study of
intertextuality is thus highly relevant to interpreting corpus-based data.

5. Concluding remarks

An approach through which all variant forms of the present participle in Scots
are considered in order to identify specialized distributions of varying gram-
matical functions has allowed us to trace patterns reflecting a relatively high
degree of idiolectal systematicity. Further descriptive analysis in transitional
periods will provide information on the direction of change. Consistently
applied rules in idiolectal grammars of letter-writers, for instance, can provide
relevant evidence not available elsewhere. Developments in the idiolectal use of
the progressive form have been illustrated by systems in which the present
participle form in -in(e) is consistently preferred in the progressive, for in-
stance, whereas the form used in contexts such as formulaic expressions ends in
-ing (e.g. example 7). By examining the discourse-specific use of the progressive
and the specialized distributions of all the variants occurring as equivalents of
Present-day English -ing, we may be able to suggest relationships between these
idiolectal systems; ultimately, comparative work of this kind may provide
evidence for taxonomies that are perhaps somewhat different from those
presented in earlier work as well as evidence depicting the direction of change
in the various varieties of English. To study contact phenomena such as the
influence of Celtic languages, we must first focus on the compilation of repre-
sentative text corpora.
<DEST "meu-n*">

The progressive in Older Scots 225

Notes

* I am grateful to Teresa Fanego and an anonymous referee for valuable comments on

earlier versions of this paper.
1. The size of the Helsinki Corpus of Older Scots, 1450–1700, is approximately 850,000
words of running text representing 15 different, chiefly non-literary genres (acts of Parlia-
ment, burgh records, trial proceedings, histories, biographies, travelogues, diaries, pamphlets,
educational treatises, scientific treatises, handbooks, private letters, official letters, sermons
and the Bible). Bibliographical information is available in Meurman-Solin (1995b) and
[Link]
2. In the data examined by Devitt (1989: 29) “-and is the dominant form in only eleven of
the 121 texts, and all before 1580. More revealing is what happens in the texts after 1600.
Although the general pattern showed no significant increase in the proportion of -ing over
-and after 1600, individual texts are moving toward greater consistency. By 1659, only two
of twenty texts are using both -and and -ing. Thus the anglicization of the present participle
appears to continue even after -ing has become the norm; -ing is increasingly becoming the
only possible form”. She points out that the spread of -ing may have taken place earlier in
literary texts than in her data, which chiefly consists of non-literary ones.
3. According to Dons & Moessner (1999: 24), the mean frequencies (per 1,000 words) of
forms in -and in the progressive in the HCOS are as follows: 0.1 (10 instances) in pre-1500
texts, 0.28 (58 instances) in texts dating from the period 1500–1570 and 0.14 (44 instances)
in the period 1570–1640; the form does not occur in 1640–1700, the last subperiod of the
corpus. Whether the figures only comprise ‘true progressives’ (cf. Denison 1993: 372–380) is
not clearly indicated as the grammatical categories applied have only been defined relatively
briefly, chiefly in reference to Quirk & al. (1985: §17.54).
4. It is noteworthy that, in some sixteenth- and seventeenth-century texts, there is also
variation between past participles in -en(e), -in(e)/-yn(e) and -ing(e)/-yng(e), for instance
halding and speking for the prevailing halden and speken. Interestingly, in a letter dating from
1546, James Cockburn of Langton (NAS SP2/2: 122) first chooses the past participle form
gettyng, but then cancels the final -g. Variation of this kind also occurs in non-morphemic
word-final elements. Thus the variants samen and samyng ‘same’, and befor wretyng, instead
of befor wretyn, co-occur in the following example: “Alssua, that ilk day, John Smayl has
mayd the samen condission that he sal kyp the samyng ordenans befor wretyng and be the
samyng cwndission” (HCOS 1457 Peebles B. Rec., 120). Similarly, “yis / mornyng george
dowglass is ryddyng (‘has ridden’) by my place” (CSC 1543–4? Adam Otterburn NAS
SP2/1:44, Correspondence of Mary of Lorraine, XLIV:53–55).
5. A study of this kind must be based on diplomatic editions or manuscripts (see fn. 7).
6. I also looked at distributions of high-frequency items such as accordand and according: the
variant in -and occurs only in the 1450s. The variant pertenand resists longer, but, interest-
ingly, the variant pertening is more frequent in the Acts of Parliament in the post-1460 texts.
The variant considering occurs only in the Acts in pre-1500 texts. The formulaic expression
God willing is almost invariably used: there are only two examples of God willand (in Peebles
and Aberdeen records) in the HCOS. In analysing the distributions of forms in -and and
226 Anneli Meurman-Solin

-ing in the HCOS, I have excluded the New Testament and William Lamb’s Ane Resonyng of
ane Scottis and Inglis Merchand betuix Rowand and Lionis because of their special nature; the
former is a scotticization of Purvey’s revision of Wycliff’s New Testament, the latter a
dialogue between speakers of two varieties.
7. In the digitized texts, y is used for a thorn, z for a yogh. Expansions in contractions are in
italics. The texts of the HCOS have usually been quoted here as rendered in the editions as
the process of checking the texts against manuscript has not yet been completed. The letters
in the CSC corpus have been checked against manuscript if not stated otherwise. As regards
the Correspondence of Mary of Lorraine, 1542–1560, chiefly autographs have been included;
however, as indicated by the reference, two examples have been extracted from non-
autograph letters.
8. In the Correspondence of Mary of Lorraine, two of the four Edinburgh informants use only
variants in -and, the third the variant in -ing only in tweching and nochtwithstanding. Only
Adam Otterburn, a representative of the merchant class, has a fuller set of uses of present
participles realized by both variants. However, it may be significant that in his idiolect
variants in -ing are also particularly frequent in non-finite clauses with an implicit first-
person subject in formulae at the end of a letter. This pattern of distributions is repeated in
all mid-sixteenth-century letters in my corpora.
9. In Table 1(d), the mean frequency of the progressive in narrative texts is 1.2/1,000.
Interestingly, there is a significant difference between Alexander Brodie’s entries in his diary,
which date from 1652–1680 (0.49/1,000), and those by his son James in the period
1680–1685 (2.42/1,000).
10. It is interesting that Jane Drummond, who was a governess to the children of James VI
until 1617, does not use the progressive, whereas her contemporary Anna Livingston, who
was one of the maids of honour to Queen Anne (of James VI) before her marriage to the
sixth Earl of Eglinton, uses the progressive with verbs of motion and also with the verb cause
(for further information on women writers, see Meurman-Solin 1999 and 2001b).
11. According to Dons & Moessner (1999: 24), only three instances of variants in -and are
used in nominal functions out of the total of 1,295 occurrences in the HCOS.
12. Example 10 is from the Edinburgh Corpus of Older Scots (ECOS); I would like to thank
Keith Williamson (Institute for Historical Dialectology, University of Edinburgh), the
compiler of the corpus, for kindly pointing it out to me. Further study is necessary to
examine whether types such as be + in + -ing + of and be + long + in + -ing (both be lang in
cuming and be long a comeing have been attested in the data) can be significantly related to
developments in the spread of the progressive verb forms. There is an interesting example of
the former in James Melville’s Autobiography and Diary (1600–1610, f242v): “Bot as I was
thus about to win the king / as in me lay to the kirk / sa was he in winning of me to the /
Court”. The latter can be illustrated by an example from an early eighteenth-century letter by
John Tarbat in the HCOS: “Your ordors anent the elections of the shyre of Ross uas so long
a coming that (…) the Uhige partie would, uithout dout, have caried ther design” (Earls of
Cromartie, Vol. 1: 276).
 The progressive in Older Scots 227

13. The paradigm of the progressive in the history of English is summarized in Denison
(1993: 383–384); according to Visser (1973: §2148), the first Present Perfect Continuous
attested in English dates from the fourteenth century.
14. In the earliest autograph letters in Scots, addressed to Mary of Lorraine (of Guise) dating
from the mid-sixteenth century, the progressive is chiefly used with verbs of motion (ride,
pass, gang, come) or with verbs referring to state (lie, remain), but transitive verbs such as do,
gather, make and break also occur in the progressive, as illustrated by the following example:
“Madame pleis zowr grace witt ye man hes failzeit trist and we / ar makand new prowision to
se ye maneir & to affix ane new / trist /” (CSC 1548 Alexander Gordon NAS SP2/2: 161,
Correspondence of Mary of Lorraine, CLXI: 227).
15. This example is particularly interesting since the progressive of have has not been attested
between Late Middle English and its reappearance in the late eighteenth century (Visser
1973: §1841; Warner 1997: 167–168).
16. In one example in the passive sense the present participle is in -and: “and in the meyne
tyme quhen diligent inquisition was makand to have done justice in the said mater” (HCOS
1565 Criminal Trials, *464)
17. In another example in the same source there is an ambiguous character in the manuscript
between be and -ing which is probably y with an omitted flourish, perhaps suggesting a
variant of there: “your la shall kno that this beerer is sentt / from the genttrie of murray to
kno what / is y[ar] doing and to returne als soone as he / cane” (CSC 1640 Jeane Ross NAS
GD 3/5/244, Memorials of the Montgomeries, 114: 242).
18. Scheffer (1975: 196) draws the same conclusion in his study of Wærferth’s translation of
Gregory’s Dialogues. He notes that where the style is lively narrative, picturesque and detailed,
the frequency of the progressive is higher, where it is more reflective and moralizing, with the
narrative, if any, kept sober and factual, the frequency is lower. See also Rydén (1997: 421),
who refers in this way to the core meaning of the progressive in English: “non-progressive
predications (…) are factual, informative, presentative rather than graphic, analytical or
evaluative”.

References

Aitken, Adam Jack. 1978. “Oral Narrative Style in Middle Scots”. Actes du 2e Colloque de
Langue et de Litterature Ecossaises (Moyen Age et Renaissance) ed. by Jean-Jacques
Blanchot & Claude Graf, 98–112. Strasbourg: Université de Strasbourg.
Biber, Douglas. 1988. Variation across Speech and Writing. Cambridge: Cambridge University
Press.
Denison, David. 1993. English Historical Syntax: Verbal Constructions. London & New York:
Longman.
Dennis, Leah. 1940. “The Progressive Tense: Frequency of its Use in English”. Publications of
the Modern Language Association of America 55.855–865.
Devitt, Amy J. 1989. Standardizing Written English. Diffusion in the Case of Scotland
1520–1659. Cambridge: Cambridge University Press.
228 Anneli Meurman-Solin

A Dictionary of the Older Scottish Tongue. 1937– (part XLIV S(c)hake to S(c)hot 1996). Ed.
by Sir William A. Craigie et al. Oxford: Oxford University Press.
Dons, Ute & Lilo Moessner. 1999. “The Present Participle in Middle Scots”. Scottish
Language 18.17–33.
Elsness, Johan. 1994. “On the Progression of the Progressive in Early Modern English”.
ICAME Journal 18.5–25.
Meurman-Solin, Anneli. 1993. Variation and Change in Early Scottish Prose. Studies Based on
the Helsinki Corpus of Older Scots (= Annales Academiae Scientiarum Fennicae, Diss.
Humanarum Litterarum, 65). Helsinki: Academia Scientiarum Fennica.
Meurman-Solin, Anneli, ed. 1995a. The Helsinki Corpus of Older Scots, 1450–1700.
Distributors: Oxford Text Archive (Oxford) & Norwegian Computing Centre for the
Humanities (Bergen).
Meurman-Solin, Anneli. 1995b. “A New Tool: The Helsinki Corpus of Older Scots
(1450–1700)”. ICAME Journal 19.49–62.
Meurman-Solin, Anneli. 1997. “Differentiation and Standardisation in Early Scots”. The
Edinburgh History of the Scots Language ed. by Charles Jones, 3–23. Edinburgh: Edin-
burgh University Press.
Meurman-Solin, Anneli. 1999. “Letters as a Source of Data for Reconstructing Early Spoken
Scots”. Writing in Nonstandard English ed. by Irma Taavitsainen, Gunnel Melchers &
Päivi Pahta, 305–322. Amsterdam & Philadelphia: John Benjamins.
Meurman-Solin, Anneli. 2000a. “Geographical, Socio-spatial and Systemic Distance in the
Spread of the Relative Who in Scots”. Generative Theory and Corpus Studies: A Dialogue
from 10 ICEHL ed. by Ricardo Bermúdez-Otero, David Denison, Richard M. Hogg &
Christopher B. McCully, 417–438. Berlin & New York: Mouton de Gruyter.
Meurman-Solin, Anneli. 2000b. “Change from Above or from Below? Mapping the Loci of
Linguistic Change in the History of Scottish English”. The Development of Standard
English, 1300–1800: Theories, Descriptions, Conflicts ed. by Laura Wright, 155–170.
Cambridge: Cambridge University Press.
Meurman-Solin, Anneli. 2001a. “Genre as a Variable in Sociohistorical Linguistics”. Early
Modern English Text Types ed. by Lilo Moessner, 245–256. [Special issue of the European
Journal of English Studies 5:2].
Meurman-Solin, Anneli. 2001b. “Women as Informants in the Reconstruction of Geographi-
cally and Socioculturally Conditioned Language Variation and Change in 16th and 17th
Century Scots”. Scottish Language 20.20–46.
Meurman-Solin, Anneli, ed. Forthcoming. The Corpus of Scottish Correspondence,
1400–1800.
Mittendorf, Ingo & Erich Poppe. 2000. “Celtic Contacts of the English Progressive?” The
Celtic Englishes II ed. by Hildegard L. C. Tristram, 117–145. Heidelberg: Carl Winter.
Mustanoja, Tauno F.1960. A Middle English Syntax. Part I: Parts of Speech. Helsinki: Société
Néophilologique.
Nurmi, Arja. 1996. “Periphrastic DO and be + ing: Interconnected Developments?”
Sociolinguistics and Language History. Studies Based on the Corpus of Early English
Correspondence ed. by Terttu Nevalainen & Helena Raumolin-Brunberg, 151–165.
Amsterdam & Atlanta, GA: Rodopi.
</TARGET "meu">

The progressive in Older Scots 229

Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive
Grammar of the English Language. London & New York: Longman.
Rydén, Mats. 1997. “On the Panchronic Core Meaning of the Progressive”. Studies in the
Changing English Language in Honour of Matti Rissanen ed. by Terttu Nevalainen &
Leena Kahlas-Tarkka, 419–429. Helsinki: Société Néophilologique.
Scheffer, Johannes. 1975. The Progressive in English. Amsterdam: North-Holland.
Strang, Barbara M. H. 1982. “Some Aspects of the History of the Be+ing Construction”.
Language Form and Linguistic Variation: Papers Dedicated to Angus McIntosh ed. by John
Anderson, 427–474. Amsterdam & Philadelphia: John Benjamins.
Visser, Frederikus Th. 1973. An Historical Syntax of the English Language. Part III, Second
Half: Syntactical Units with Two and with More Verbs. Leiden: E. J. Brill.
Warner, Anthony R. 1997. “Extending the Paradigm: An Interpretation of the Historical
Development of Auxiliary Sequences in English”. English Studies 78.162–189.
Wright, Susan. 1994. “The Mystery of the Modal Progressive”. Studies in Early Modern
English ed. by Dieter Kastovsky, 467–485. Berlin & New York: Mouton de Gruyter.
<LINK "moh-n*">

<TARGET "moh" DOCINFO

AUTHOR "Ruth Möhlig and Monika Klages"

TITLE "Detransitivization in the history of English from a semantic perspective"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Detransitivization in the history of English

from a semantic perspective*

Ruth Möhlig and Monika Klages

University of Cologne

1. Introduction

Verbs which were originally transitive and developed an additional intransitive

use at some point in their history have primarily been studied from a syntactic
perspective. The focus of this paper will, by contrast, be on their semantic
dimension — following the assumption that “the properties of clause structure
are predictable from the semantics of predicates” (Faber & Mairal Usón 1999:37).
We look into four different types of uses of selected transitive verbs which result
in non-transitive constructions and analyse the semantic processes involved in
the development of these non-transitive presentations of transitive events.
Traditionally, transitive verbs have been defined on a purely morpho-
syntactic basis as those governing an object in the accusative case. For dealing
with languages like Old English or German, however, this definition has proved
to be too narrow, since there are verbs which govern a direct object in the dative
or genitive case.1 Apparently, semantic differences are expressed in these
distinctive case markings which are more or less systematic: thus, in Old English,
dative case marking on direct objects is often correlated with participants which
are less affected by the process denoted by the verb, compared to direct objects
in the accusative case (Traugott 1992: 204); the OE genitive case is typically
assigned to the direct object, e.g. when contact between the two participants is
only partial (cf. Mitchell 1985: 232; Langacker 1991; Traugott 1992: 205).
In contrast to the traditional definition, we define transitive verbs on a
functional-semantic basis as those verbs which typically take a Goal-object (Dik
1989, 1997)2 as their second argument, i.e. verbs which express an action or
process which affects a second participant, regardless of morphologic case. An
activity is conceived as some kind of transfer taking place between an Agent and
232 Ruth Möhlig and Monika Klages

another participant (Hopper & Thompson 1980: 251). Following Hopper &
Thompson (1980), we take transitivity as a graded concept which involves
semantic and syntactic features: highly transitive clause structures are character-
ized as being dynamic,3 telic4 (or perfective), punctual and in the affirmative; at
least two participants — a subject and an object — are present and the object is
totally affected by the event. The fewer of these features apply, the lower is the
transitivity of the clause. Thus, transitivity is not a property of verbs alone, but
of a combination of elements present in the clause (Hopper & Thompson
1980: 251). On this scale of transitivity verbs which govern direct objects in the
dative or genitive in Old English must be viewed as less transitive, since direct
objects in the dative or genitive usually represent less affected or only partially
involved participants. In this paper we have focused on verbs which — if they
were not introduced into English later than Old English — predominantly show
accusative case marking on their direct objects in Old English,5 i.e. verbs which
may originally be regarded as highly transitive.
The term intransitive use is here primarily understood as syntactic, that is,
as referring to transitive verbs which in certain contexts do not appear with a
Goal-object (e.g. generic use, see Section 3.2). We argue that not all verbs used
non-transitively arrive at the full status of intransitive verbs, the latter being
taken as denoting an action or process which does never extend beyond the first
participant (e.g. ModE talk, sleep; cf. Halliday 1994: 109–110).
The theoretical framework adopted by us largely follows that of functional
grammar as proposed by Dik (1989, 1997), Halliday (1994) and Givón (1984/
1990). We have preferred this framework to, for instance, transformational
approaches (e.g. Radford 1997), since it allows us to focus on the functional-
semantic aspects which may be explanatory of syntactic usage, rather than on
the description of the formal rules by which sentences are generated. However,
it is conceded that both functional and generative approaches may with some
benefit, but also with constraints, be applied to the diachronic analysis of
language and language change.
Our diachronic data were drawn from the Helsinki Corpus of English Texts,
Diachronic Part (1991, HC), the Oxford English Dictionary (1989, OED), the
Middle English Dictionary (1952–2001, MED), and the Microfiche Concordance
to Old English (1985, MC). It has to be borne in mind that these sources, which
are in themselves based on historical texts that may only sketchily reflect the full
range of historical usage, are selective. Thus, non-occurrence of a pattern at a
particular time does not necessarily rule out that this pattern was used at that
 Detransitivization in the history of English 233

time. The sources are, however, suggestive as to the degree to which a pattern
was established in the written medium at a given time.

2. Transitive–intransitive verbs in the history of English

Before we look at the different patterns according to which transitive verbs have
expanded to intransitive uses, it should be noted that transitive–intransitive
verbs in English go back to various sources and that their development involves
processes of transitivization as well as of detransitivization. Thus, several
morpho-phonological developments in Old English led to the merging of
previously differentiated verb classes, which resulted in homophonous transi-
tive–intransitive verbs of the alternating causative–inchoative type.
We can distinguish between the merger of primary intransitive inchoative
verbs with their transitive causative derivatives, and the morphological con-
flation of OE weak verb classes (cf. Görlach 1997: 83–85): Old English still
preserved the distinction inherited from Proto-Germanic between a primary
strong stative or inchoative verb and a secondary weak causative derived from
it; thus, Old English had sincan ‘sink’ : sencean ‘make to sink’, meltan ‘melt’ :
mieltan ‘make sth. melt’, and brinnan/beornan ‘burn (up)’ : bærnan ‘make sth.
burn (up)’. The pair meltan : mieltan was probably already confused in OE
times, whereas sincan : sencean (> ME sench, not later than 14C) and brinnan/
beornan : bærnan became conflated in sink and burn respectively in the ME
period. However, other OE stative/inchoative–causative pairs like licgean :
lecgean, feallan : fellan were kept phonologically distinct (lie : lay and fall : fell);
others like drincan : drencan were retained in drink : drench, although the latter
went out of use in the causative meaning ‘make to drink’ at the beginning of the
twentieth century and was replaced by the verb to water, which competed with
drencan/drench already in (late) Old English.6
Within the classes of weak verbs, class III of intransitive inchoative Gmc.
*-ē-verbs had fallen in with class II of the mostly transitive causative/agentive
Gmc. *-ō-verbs in the OE weak class II, which was marked by the infinitive
suffix -ian (Krahe & Meid 1969: 249; Schaefer 1984; Heidermanns p.c.). Thus,
a Gmc. *-ō-verb like OE openian ‘open’ shows both transitive causative and
intransitive inchoative uses already in Old English. Whether the inchoative or
causative use was primary is difficult to establish, the Gmc. evidence failing to
yield a clear pattern.7 However, be it a Gmc. or an early OE development, the
verb probably acquired either an additional inchoative or causative interpretation
234 Ruth Möhlig and Monika Klages

simply by virtue of its membership in a verb class which incorporated both

semantic types, the analysable adjectival base Gmc. *upana-/OE open showing
both readings — stative/inchoative ‘be/become open’ and causative ‘make
open’ — and this led to a functional expansion of the verb (cf. Görlach
1997: 84–85).
Such a functional expansion on the basis of an analysable adjectival base
may — possibly by analogy with cases like OE openian — also have been the
case with originally causative English verbs, which developed an inchoative use
somewhat later than Old English, e.g. ME drye(n) (< OE drygan : dry adj.,
inchoat. c1200) and close(n) (< OE clysan/ME closen : close < OFr. clos- adj.,
inchoat. 14C), EModE fill (< OE fyllan : full, inchoat. beg17C) and ModE clean
(derived from clean adj., inchoat. 18C).
The development of verbs showing alternative transitive and intransitive
uses in English may have been further strengthened by the loss of the OE verbal
prefix ge- (< Gmc. *ga-; cf. Krahe & Meid 1969: 37) in the course of Middle
English. For Old English, this prefix has frequently been seen as having a
transitivizing effect, which resulted in causatives — e.g. feallan intrans. ‘fall’ : ge-
feallan trans. ‘overthrow sb.’ (=‘make sb. fall’), hwı̄tian ‘be/become white’ : ge-
hwı̄tian ‘make white’, minsian ‘become small’ : ge-minsian ‘make small’ — or in
verbs which were additionally resultative/perfective — e.g. flōwan ‘flow’ : ge-
flōwan ‘overflow sth.’, ærnan ‘run’ : ge-ærnan ‘reach, gain by running’ (Visser
1963–1973: §134; cf. Kastovsky 1992: 380). With other OE verbs, however, there
is no difference in meaning between the simplex and the prefixed verb, e.g.
(ge)æmtian intrans. ‘be at leisure, unoccupied’, likewise (ge)openian trans./
intrans. ‘be/become open, make open/apparent’ and (ge)brecan trans./intrans.
‘break’. Thus, the OE data are inconclusive as to whether preverbal ge- had
transitivizing function, something which could have led to a wide-scale con-
flation of an intransitive simplex and a transitive prefixed verb after OE ge- had
been lost (via ME y-/i-). However, in individual cases, such a conflation seems
to have occurred.
It is evident from these considerations that from Old English on, a variety
of homophonous or homonymous verbs from various origins existed which
incorporated both transitive and intransitive uses; it is a special feature of
English that this group resulting from processes of detransitivization as well as
transitivization increased so considerably over the following centuries. Thus,
various verbs, e.g. cook (derived from cook n. 14C), which shows both transitive
and intransitive uses from the seventeenth century on, were originally intransitive
and have undergone processes of transitivization, thereby further strengthening
 Detransitivization in the history of English 235

the group of transitive– intransitive verbs. Rissanen (1999: 256) points out that
several instances of originally intransitive verbs of motion in the imperative
may be found used with a co-referential object in the sixteenth/seventeenth
centuries (isolated instances found for hasten as late as the eighteenth century)
— something which he associates with involvement and emphasis, e.g. “Good
Margaret runne thee to the parlour” (Shakespeare Much Ado about Nothing III.i;
see also Görlach 1994: 84).8 Such instances show that there have been alterna-
tives in degrees of transitiveness in the history of English, facilitated by the
overall loss of distinctiveness of syntactic-semantic verbal categories, which has
resulted in the transitivization or detransitivization of verbs and in the emer-
gence of various transitive–intransitive patterns (see also Görlach 1994: 83–84).9
Further support for the group of transitive – intransitive verbs may have
come via borrowing, e.g. ME moue and dresse: both verbs were apparently
borrowed in their transitive, reflexive and intransitive uses from OFr. (se) mover
(13C) and (se) drescer (14C), respectively. Whereas move has not undergone any
fundamental semantic changes during its history in English, the development
of dress (ModE ‘clothe sb./oneself ’) was different, the verb being already
polysemous in Middle English: the basic meaning of OFr. (se) drescer was ‘direct
sth./sb./oneself ’;10 the first known meaning of the verb in Middle English is
‘array, prepare, make straight’, used transitively and with a co-referential
reflexive pronoun. The meaning of ModE dress ‘clothe sb./oneself’ developed in
Late Middle English via ‘get armed, attired’, by metonymy from ‘prepare
oneself ’. The verb dress shows co-referential reflexive uses for all shades of
meaning already in Middle English, something probably borrowed from Old
French. First instances of the intransitive use (as in He dressed) are found in the
fourteenth century; it is not clear whether this use was also borrowed from
French — as is the case, for example, with move (see Section 3.1).
It may be argued that the morpho-phonological processes starting in Old
English triggered the emergence of the large group of English verbs which have
developed additional transitive or intransitive uses. The picture, however, is
complex and involves the interaction of phonological, morphological and
syntactic processes in addition to the possible effects of language contact. These
developments were at least partially semantically motivated and it is the
semantic processes that underlie four of the most important detransitivizing
syntactic constructions of English that we will discuss in the following.
236 Ruth Möhlig and Monika Klages

3. Diachronic patterns of detransitivized use

We will here focus on the diachronic development of four patterns of detransi-

tivization, namely co-referential intransitive use, ergative use, generic use and
middle use.11 These four patterns have frequently been linked with each other,
since they all show various detransitivizing effects which result in a change in
the predicate frame12 of individual transitive verbs. We will illustrate the nature
and extent of these links on semantic grounds. In doing this, we intend to show
that the features in which the four constructions differ are more prominent
than those which they share.
The passive voice, which also has a detransitivizing effect (Siewierska
1984: 44; Givón 1990: 564–574), has largely been left out of our present discus-
sion, since — as an increasingly structural part of the voice–aspect system of
English from late OE/ME on — it is not bound to isolated lexemes and does not
change the predicate frame of transitive verbs, as do the four patterns discussed
here (on the interaction between the progressive passive and middle use see
Section 3.4).
It is important to note that the assignment of individual verbs to one of the
four patterns is not exclusive; i.e. verbs may fall under more than one pattern,
although certain semantic restrictions may block multiple occurrence.

3.1 Co-referential intransitive use

Much emphasis has traditionally been placed (e.g. Jespersen 1927: 325–329;
Deutschbein 1931: 98–103) on those detransitivized verbs which developed via
a stage where they showed co-referential reflexive use;13 this usually led to
polysemous transitive–intransitive verb pairs, e.g.
(1) John washed the clothes. (transitive)
(2) John washed himself. (co-referential reflexive)
(3) John washed. (co-referential intransitive)

Synchronically, the semantic characteristics of this pattern are the following:

a. The Agent of the transitive verb typically is not only animate but human.
b. The transitive Goal-object is first assigned the semantic feature of co-
referentiality with the human Agent; this results in a co-referential (pro-
nominal) Goal-object; the feature of co-referentiality is in a second step
incorporated into the meaning of the predicate. Since the entity controlling
 Detransitivization in the history of English 237

the action (the Agent) and the entity which is affected (the Goal) are
identical, the subject of the co-referential intransitive use strictly speaking
is ambivalent to an interpretation as Agent or Goal. However, the Agent is
more salient,14 so that no change of perspective is involved.
c. The states of affairs described by the transitive, the reflexive and the
intransitive clauses are typically dynamic, controlled and telic; this means
that the semantic characteristics are the same in all three constructions.
The emergence of this type of pattern of intransitive use is not limited to one
specific historical period of English;15 it occurs, for instance, with OE baþian,
ME washe(n) (< OE wascan, intrans. c1175), clothe(n) (< OE clāþian, intrans.
1390s), dresse(n) (trans., reflex. < OFr. intrans. c1400), moue(n) (trans., reflex.,
intrans. < OFr. c1250), ModE shave (< OE sceafan, intrans. a1715)16 and
straighten (16C, intrans. 1890s),17 e.g.
(4) Bede 4 21.318.15 Ond seldon in hatum baðum heo baðian wolde, (MC)
“And rarely did she want to bathe in a hot bath,”
(5) c1175 Lamb. Hom. 159 þos fure kunnes teres boð þe fuwer wateres þa þe
beoð ihaten us on to weschen. (OED s.v. wash v. 3.j)
“These four kinds of tears are the four waters in which we are command-
ed to wash”
(6) c1320 Sir Tristr. 541 þe king no seyd no more, But wesche and Šede to
mete. (OED s.v. wash v. 3.j)
(7) 1393 Gower Conf. I. 14 The tresor…Wherof the pouer shulden clothe
And ete and drinke and house bothe. (OED s.v. clothe v. 2)
(8) c1440(?a1400) Morte Arth.(1) 2052 He drissede in a derfe schelde
endenttyd with sable. (MED)
(9) c1250 Kent. Serm. in O. E. Misc. 29 þo seide ure lord to þo serganz.
Moveth to-gidere and bereth to Architriclin. (OED s.v. move v. 16.a)
(10) a1715 Burnet Own Time (1766) II. 219 He was quickly dressed, but
would lose no time in shaving. (OED s.v. shave v. 6.b)
(11) 1891 Kipling Light that Failed xiii. 256 Dick’s shoulders straightened
again, for the words lashed like a whip. (OED s.v. straighten v. 4.a)

The development may be interpreted as a semantic process of incorporation of

the co-referential Goal into the semantic structure of the verb; this leads to a
polysemous verb pair of, for example, transitive wash ‘make sth./sb. clean by
means of water’ and co-referential intransitive wash ‘make oneself clean by
238 Ruth Möhlig and Monika Klages

means of water’; the parallel co-referential reflexive use is gradually understood

as the emphatic, more intensive variant of intransitive wash.18 However, this
incorporation of the feature of co-referentiality does not extend to all verbs
which occur in co-referential reflexive use — compare, e.g. ask oneself, offer
oneself (OE ascian (hine sylfne), (ge)offrian (hine sylfne); cf. Ogura 1989: 74,
96),19 etc. This implies that various semantic constraints are connected with the
emergence of this pattern.
The most important constraint appears to be related to the lexical domains
to which the verbs of this syntactic pattern belong: thus, all verbs investigated by
us belong to the lexical domains of Body-care or directed Body Movement
and are typically found in combination with animate Agents. By virtue of this
typical collocation pattern, it was possible for the notion of a co-referential
Goal-object to become incorporated into the semantic structure of the verb.
The picture is only slightly relativized when one looks at various synonyms of
e.g. wascan/wash, dresse/dress and moue/move: OE (ge)lafian ‘wash, bathe’,
which is a partial synonym of wascan ‘wash’, is documented only four times in
transitive use (cf. MC); the far more common partial synonym OE (ge)þwēan
‘wash, cleanse’ did develop a reflexive use, e.g. Exod. 2.5 “þa eode Pharaones
dohtor & wolde hi þwean æt þam wætere” (MC). But apparently it did not
survive long enough to develop an intransitive use as well. As a partial ModE
synonym of wash, the verb cleanse from OE clǣnsian ‘make clean’ (derived from
the OE adj. clǣne) acquired a co-referential reflexive use for the metaphorical
meaning ‘purge of sin or accusation’, e.g. c897 K. Ælfred Past. liv. 419 “Hi selfe
to clænsianne mid ðy wope”, but no co-referential intransitive use. However,
other verbs denoting directed Body Movement or the act of preparing similar
to ME dresse(n) ‘prepare, direct’, and ME moue(n) — e.g. direct (14C derived
from Lat. direct-us adj.) and prepare (15C < Fr. préparer) reveal fairly clear
patterns in that they both occur in all three syntactic uses by the sixteenth
century (prepare) and seventeenth century (direct), respectively.
The importance of influence from borrowing especially with regard to the
co-referential intransitive pattern was mentioned in Section 2 with respect to
the French borrowings in Middle English moue and dresse. However, since
various native verbs had developed a co-referential intransitive use by Old
English or early Middle English, such as baþian and wasshe(n), the pattern as a
whole cannot have been borrowed from Old French, although borrowings may
have supported this type of construction in English verbs.
 Detransitivization in the history of English 239

3.2 Generic use

Generic use of transitive verbs indicates the habitual and denotes characteristic
properties or given truths.20 It should be noted that — contrary to older
approaches (e.g. Jespersen 1927; Deutschbein 1931; Visser 1963–1973) —
generic use is here distinguished from absolute uses of transitive verbs: absolute
uses occur when the object is implied either by the textual or the situational
context (e.g. by being anaphorically given or visible in the communication
situation), whereas the objects of generic uses are typically implied by the
cultural context (e.g. world knowledge); thus, in absolute uses the object is
syntactically elided, whereas generics do not even subcategorize for an object.
Absolute uses can occur with any transitive verb in English without changing the
verb’s transitiveness; they lack the habitual notion which is typical of generics.
The ModE examples (12), (13) and (14) for read illustrate the differences:
(12) Mary read a book. (transitive)
(13) He is wise who reads. (generic)
(14) He picked up the paper and read. (absolute)

Since with absolute uses no change in the verb’s transitiveness is involved, only
generic uses are of relevance to our present discussion. The synchronic semantic
characteristics of the generic pattern are the following:
a. The state of affairs encoded is typically non-dynamic (or stative). This lack of
dynamism is what lends generic constructions their habitual aspect, so that
they are interpreted as referring to characteristic properties or given truths.
b. In generic use, the Goal is excluded from the predicate frame: the action
denoted by the verb no longer extends to an (implied) second participant.
Thus, generic use does not result in a shift of focus; rather, a distinct non-
dynamic state of affairs is presented. The first argument (the subject) is
interpreted as Zero, i.e. as an unchanging and non-acting entity which is
involved in a state (Dik 1997: 118; cf. Langacker 1991: 288, 556).
c. The Zero-subject is usually animate.
Generic use is already established for OE rǣdan ‘look at and understand
writing’, EModE drink and wash (trans.), and ModE drive, e.g.
(15) c950 Lindisf. Gosp. Mark xiii. 14 Seðe redes oncnauað [c1000 OnŠyte se
þe ræt]. (OED s.v. read v. 15.a)
“[This] understands he who reads”
240 Ruth Möhlig and Monika Klages

(16) 1640 Fletcher & Shirley Nightwalker iv. sig. H2v, Give me the bottle, I
can drink like a Fish now, like an Elephant. (OED s.v. drink v.1 11.a)
(17) 1697 Tryon Way to Health vi. (ed. 3) 101 It [rain-water] Brews and
Washes to greater advantage than others. (OED s.v. wash v. 2.h)
(18) 1717 Prior Alma iii. 140 The man within the coach that sits … Is safer
much … than he that drives. (OED s.v. drive v. 5.c)

However, generic use is missing in the historical data for most verbs of the co-
referential intransitive, ergative and middle use patterns investigated here (e.g.
dress, move; open, break, fill; tell, load). On semantic grounds, typical transitive
verbs of Experience such as feel, taste and see lack generic use, too.
Various semantic factors play a role in the exclusion of these verbs from
generic use: verbs which can occur in the generic non-transitive frame, such as
read, drive and drink, mostly belong to the less prototypical transitive verbs in
that they encode events in which the object does not register a change of state,
such as from close to open, intact to broken or empty to full. Thus, their objects
are not prototypical Goals in that they are not affected by the event in the sense
of the predicate (cf. Hopper & Thompson 1980: 251).21 On the other hand, it is
evident that in the events encoded by these transitive verbs an activity is
involved; i.e. the first participant in the transitive clause is an Agent — thus
excluding verbs of Experience or Emotion, which typically have a Processed-
Experiencer or ZeroExperiencer as first participant (cf. Dik 1997: 118–120).22
The first participant within the original transitive frame of verbs subject to
generic use must therefore be an Agent (cf. Hopper & Thompson 1980: 252).
Potentially three-place Transfer verbs such as tell and load are usually
excluded from intransitive generic use, although on the grounds of the above
considerations they do not belong to the most prototypical (di)transitive verbs
either. Here, a blocking factor probably comes in with the potential third
participant, i.e. Recipient, which is typically implied with Transfer verbs, even
if it is not realized (Goldberg 1992: 49–50). One might also argue that the
monotransitive use of tell and load (e.g. in She tells a story, He loads the van) is
already a case of generic use which reduces the verb from a three-place to a two-
place predicate. Nevertheless, the Transfer verbs buy (< OE bycgean) and sell
(< OE sellan ‘to give (for money)’) developed an intransitive generic use as early
as Middle English, which seems to be mostly restricted to the idiomatic combi-
nation of the two in e.g.
(19) 1340 Hampole Pr. Consc. 4399 Nan sal bye with þam ne selle. (OED s.v.
buy v. 1.b)
 Detransitivization in the history of English 241

but cf.:
(20) 1611 Bible Gen. xlii. 6 And hee it was that sold to all the people of the
land. (OED s.v. sell v. 4.a)

The most prominent semantic factor with verbs showing intransitive generic
use (e.g. read, drink, drive, transitive wash, buy and sell) seems to be that they
exhibit a closer collocational binding with their potential Goals: it is evident
that actions such as reading, drinking, driving, washing, buying and selling may
only be carried out with a limited number of objects (e.g. some sort of writing,
some liquid, vehicles, clothes, anything subsumed under the general concept of
wares),23 whereas, for example, dressing, moving, opening, breaking or filling
may be done with a larger number of diverse objects. Thus, specifying the Goal-
object when describing a characteristic property or given truth is of lesser
pragmatic importance with verbs like read, drive, drink, or wash (trans.) than
with, for example, dress, move, open, break or fill (Givón 1984: 108–109).24

3.3 Ergative use

Ergative use, i.e. the alternating use of verbs in transitive – causative and
intransitive –inchoative constructions, is probably the most interesting pattern
with respect to the interrelationship between verb semantics, morphological
developments and syntactic behaviour. Here, the directionality of the develop-
ment from either transitive to intransitive or vice versa is more or less impossi-
ble to establish on the grounds of diachronic data.
From a synchronic point of view, ergative use is semantically marked by a
single non-Agentive participant; the verb is usually interpreted as inchoative.
Verbs showing this type of construction have frequently been associated with
reflexive verbs (Siewierska 1984: 77–79; Palmer 1994: 142–145), since in both
uses the subject shares some properties of a Goal,25 cf.
(21) Mary opened the door. (transitive/causative)
(22) The door opened. (ergative/inchoative)

However, a closer semantic analysis reveals that the two uses differ. The
semantic characteristics of ergative uses are the following:
a. The subject of the ergative construction is assigned the participant role of
Processed, which denotes an entity that is undergoing a dynamic, but
uncontrolled Process.
242 Ruth Möhlig and Monika Klages

b. The Processed-subject of the ergative construction is typically inanimate. It

is identical with a prototypical Goal-object of the transitive frame; i.e. a
former Goal-object appears in subject position; the Agent is excluded from
the predicate frame.
c. The state of affairs in the ergative construction is dynamic and telic, but
uncontrolled. The states of affairs coded by the transitive and the ergative
constructions differ accordingly: the transitive clause denotes an Action,
whereas the ergative use denotes a Process.
Ergative constructions appear to be a means to exclude the Agent, making it
possible to view a causative event from the perspective of the entity which is
undergoing the process denoted by the predicate. In this they resemble the
passive, which has the pragmatic functions of promoting the Goal and demot-
ing the Agent. However, the effect achieved by ergatives goes beyond the
perspectivising function: ergatives exclude the element of cause and control (i.e.
the Agent) and thus place the focus on the process itself. Whether we perceive
the transitive or the intransitive use as primary, i.e. whether we synchronically
interpret the alternation as a result of detransitivization rather than of transiti-
vization, seems to be closely connected with the respective verb meaning: the
action of opening encodes an inherently causative event more than the action
of, for instance, boiling, since, as a dynamic verb, open implies the participation
of an Agent or Force/Instrument (Dik 1989, 1997). The transitive frame in open
may therefore be regarded as cognitively primary.26 The intransitive use of boil
(e.g. The potatoes were boiling), on the other hand, bears a stative notion, which
is why it has to be combined with, for instance, begin, if an inchoative reading
is wished for (e.g. The potatoes began to boil; note that the simple past in **The
potatoes boiled is ungrammatical, apparently for the same reason). Here, the
intransitive frame is perceived to be primary. We will, therefore, focus on
ergative verbs like open for which the transitive frame is perceived as primary
and which may accordingly be considered the result of detransitivization.
Similar to verbs that follow the co-referential intransitive pattern, transitive
verbs exhibiting an ergative use can be found for all stages of English, e.g. OE
openian and brecan, ME dryŠen (< OE drygean, erg. c1200) and close(n) (~ OE
clysan, erg. 14C), EModE tear (< OE teran, erg. 16C) and fill (< OE fyllan, erg.
beg17C), ModE move (< ME moue(n), erg. 18C)27 and clean (15C, erg. 18C); e.g.
(23) a900(?) Lch I (Herb) 4.12.4 Sona hyt sceal openian & syððan hyt
geopenud beo þonne nim ðu ða ylcan wyrte ungesodene & cnuca mid
hunige, lege to þære wunde oðþæt heo hal sy. (MC)
 Detransitivization in the history of English 243

“Soon it shall open, and when it is opened, take the same herb unboiled
and pound [it] with honey, lay [it] on the wound until it is healed”
(24) a1100 LS 17.2 (MartinPeter) 148 & efne swa swa se wind sloh on þone
lig, swa bræc se lig swiðor on ðam winde, efne in þam gelicnesse swa ða
gesceafta twa him betweonan feohtan sceoldon. (MC)
“and the more the wind hit at the flame, the more did the flame break in
the wind, just as in the parable, so were the two elements to fight with
each other”
(25) c1200 Trin. Coll. Hom. 155 Sum of þe sed ful uppe þe ston and dride
þere. (OED s.v. dry v. 2.a)
(26) 1526 Pilgr. Perf. (W. de W. 1531) 260 b, His handes & fete dyd rent &
teare for the weyght of his blessed body. (OED s.v. tear v.1 7)
(27) 1685 Cotton tr. Montaigne I. 211 A soul stretches and dilates itself pro-
portionably as it fills. (OED s.v. fill v. 3.a.)
(28) a1774 Goldsm. Surv. Exp. Philos. (1776) I. 299 A globe moving through a
fluid, such as air, that closes behind the body as it moves. (OED s.v. move
v. 16.a)
(29) 1708 Lond. Gaz. No. 4431/15 The same Day came in…Her Majesty’s
Ships…to clean. (OED s.v. clean v. 2.a)

The syntactic development from the transitive to the ergative use occurs
apparently without an intermediate syntactic stage.28 Verbs belonging to this
pattern typically form causative–inchoative verb pairs; i.e. in the uncontrolled
process, the transitive causative verb receives an inchoative reading, such as can
be seen in examples (23) to (29) above. Inchoativity may thus be said to be
semantically characterized by being uncontrolled and telic, which means that
some process denoted by the verb begins spontaneously; it is not set in motion
by some controlled action (cf. Lightfoot 1979: 254; Haspelmath 1993). The
possibility of being used ergatively seems to be largely restricted to prototypical
transitive verbs, i.e. verbs which usually occur with an animate or even human
Agent and an inanimate or at least concrete Goal which is totally affected (cf.
Hopper & Thompson 1980: 252; Givón 1990: 565–566).
From a diachronic perspective, several verbs following the ergative pattern
originate from morphological developments in Germanic and Old English (see
Section 2) — the pattern being either triggered or at least strengthened by the
morpho-phonological conflation of stative/inchoative strong verbs with their
causative weak derivatives, as happened with, for instance, OE sincan ‘sink’
(inchoative) and sencean ‘make sth. sink’ (causative) in the course of Middle
244 Ruth Möhlig and Monika Klages

English. Furthermore, various deadjectival verbs with causative function may

have acquired a secondary inchoative counterpart by functional expansion on
the basis of their adjectival base, as is possible, for instance, for ME drye(n) :dry
adj. (erg. c1200), EModE fill : full adj. (erg. beg17C) and others for which an
adjective existed that could be analysed as their derivational base at the time of
the first occurrence of ergative use; analogical influence from verbs like, for
instance, OE openian : open adj. is possible here (see Section 2). The develop-
ment of ergative use with a verb like EModE tear (< OE teran, erg. 16C), which
has no paradigmatic relationship to an adjective, makes it evident that by this
time the ergative inchoative use of homonymous transitive causative verbs was
well established as a means of viewing a causative event from the perspective of
an entity which was undergoing a process, i.e. excluding the Agent and thus
placing the focus on the process itself.

3.4 Middle use

Various labels have been given to this type of construction: Visser
(1963–1973: §163) refers to it as activo-passive and Denison (1993: 392)
identifies it as medio-passive. The term ‘middle use’ adopted by us should not
be mistaken for what is frequently known as middle voice, e.g. in Givón
(1984: 620), who applies the term ‘middle voice’ to what we call ergative
constructions (as in The glass broke).
Middle use has frequently been regarded as a special form of ergative use
(cf. Denison 1993: 392; Iwata 1999: 547–548). However, although middle uses
are at first sight similar to ergative constructions in that the Goal-object of the
transitive frame is in subject position, middle uses are characteristically marked
by the occurrence of an adverb of manner which is almost always obligatory in
the predicate frame.
(30) Mary read a book. (transitive)
(31) This book reads well. (middle)

Middle use has the following semantic characteristics (cf. Iwata 1999: 528–529):
a. The subject is usually inanimate. It is identical with a prototypical Goal-object
of the transitive frame. Thus, similar to passive and ergative constructions,
middle uses are characterized by promotion of the Goal to subject.
b. Middles occur with a manner adverbial, such as well, easily, better, smoothly,
heavily, sooner or not, which is usually obligatory.29 This adverbial syntactically
 Detransitivization in the history of English 245

introduces the notion of an implicit Agent who is passing judgement on the

event (cf. Iwata 1999). However, the adverbial may be omitted if the
context of the utterance sufficiently specifies that a judgement is passed on
the event, as in examples (32) and (33) for sell and (34) for wash:
(32) 1616 B. Jonson Epigr. iii, To my Book-seller. Thou, that … Call’st a
booke good, or bad, as it doth sell, Vse mine so, too. (OED s.v. sell v. 6.a)
(33) 1833 H. Martineau Brooke Farm v. 63 They sell at about a shilling a doz-
en. (OED s.v. sell v. 6.a)
(34) 1765 Franklin Lett. Wks. III. 402 Mrs. Stevenson bids me tell Sally, that
the striped gown I sent her will wash. (OED s.v. wash v. 2.i)

c. The state of affairs encoded in middle constructions is non-dynamic and

uncontrolled; i.e. it is a State;30 accordingly, the subject is assigned Zero-
function. The focus is on the outcome of the action rather than on the
action itself (cf. Iwata 1999: 546), i.e. the state of affairs is telic.
In PDE it seems as if most transitive verbs may be used in a middle construc-
tion.31 However, the diachronic evidence does not show a clear semantic
pattern as to what verbs were used in middle constructions first. The earliest
instances are for the Speech verb ME telle(n) ‘recount, narrate’ (c1450?), the
Cognition verb ModE read (17C?), as well as for the more concrete Change-
of-state verbs EModE wash (16C) and fill (16C) and the Transfer/
Transaction verb EModE sell (17C), e.g.
(35) ?c1450 [Link]. 6644 How many Šere in certayn I fand na boke þat tellis
playn. (MED)
(36) c1590 Marlowe Jew of Malta i. ii. 451 Who …Thinke me to be a
senselesse lumpe of clay That will with euery water wash to dirt. (OED
s.v. wash v. 15.d)
(37) 1592 Shakes. Ven. & Ad. 548 Glutton-like she feeds, yet never filleth.
(OED s.v. fill v. 10.b)
(38) 1668 Shadwell Sullen Lovers iii, ‘Tis a play that shall read and act with any
play that ever was born. (OED s.v. read v. 18.a)
(39) 1656 Earl of Monmouth tr. Boccalini’s Advts fr. Parnass. i. i. 4 There is no
Merchandize in this Ware-House which sels better, then certain Fans.
(OED s.v. sell v. 6.a)

Middle use of otherwise transitive verbs is distinct from the other three patterns
described in that it developed fairly late in the history of English; few instances
246 Ruth Möhlig and Monika Klages

are found in EModE, but as Denison (1993: 392–393) points out, the usage
increases from the eighteenth century onwards and most instances to be found
are from the nineteenth and twentieth centuries (see also Visser
1963–1973: §§168–169).
Middles have frequently been discussed within the context of passive
constructions; both middles and passives share the properties of Agent demo-
tion and promotion of the second participant (cf. Givón 1984/1990; Denison
1993; Palmer 1994; Iwata 1999). Denison (1993: 392–393) discusses middle
constructions particularly within the context of the progressive passive, which
developed around 1800 (Rissanen 1999: 218), and within the context of the
older ‘passival’ use;32 the patterns are illustrated in examples (40) and (41)
(Visser 1963–1973: §§2004–2018; Denison 1993: 148–158; Görlach 1999: 82;
Hundt 2001).
(40) The house is being built. (progressive passive)
(41) The house is building. (passival)

In contrast to middle constructions, progressive passives, like passivals, typically

encode uncontrolled, atelic (or imperfective), dynamic processes; the Goal-
object is promoted to subject; i.e. progressive passives and passivals select a
Processed (see fn. 2) as their first argument. Thus, what primarily distinguishes
middle constructions from the progressive passive and passivals are the proper-
ties of dynamism and telicity.
As Denison (1993: 439) observes, Mossé (Histoire de la forme périphrastique
‘être + participe présent’ en germanique (1938)) already made “the important point
that the existence of the progressive passive remedies two separate ambiguities,
since it admits an explicit contrast both with the active progressive (he is shaving —
he is being shaved) and with the simple passive (the house is built — the house is
being built)”. We cautiously assume that the occurrence of middle uses with more
and more transitive verbs may be indicative of a development towards a grammat-
icalized construction adding to the voice–aspect system of English which allows
the coding of a non-dynamic and uncontrolled state besides the dynamic
process variant encoded in the progressive passive (cf. Fischer 1992: 251–252).
The overall increase of middle use noted in the literature (Visser 1963–1973:
§§168–169; Denison 1993: 392; Görlach 1999: 78) may indicate a historically
ongoing process, whereby the middle construction has been gradually extended
to more and more groups of verbs, until it may potentially be applied to all the
transitive verbs of English. However, the actual productivity of the pattern may
only be measured through extensive corpus-based analysis in ModE.
 Detransitivization in the history of English 247

4. Conclusion

Our analysis of selected verbs following one or more of the four diachronic
patterns of detransitivization shows the decisive role which semantic properties
and developments play in syntactic change. In this light, syntactic detransiti-
vization must be viewed as a very complex process. It does not only involve
various semantically distinct lines of development, but it also leads to different
results concerning the structure of English.
The limited set of data investigated here suggest the following: the occur-
rence of individual patterns may be restricted to verbs which share certain
semantic properties; thus, verbs following the co-referential intransitive pattern
typically belong to the semantic fields of Body-care (e.g. wash, bathe, dress) or
directed Body Movement (e.g. move, ME dresse ‘direct oneself ’); verbs occur-
ring in the ergative pattern are usually such transitive verbs with a causative
function which denote an action that totally affects their objects — i.e. proto-
typical transitive verbs (e.g. open, break, fill). Verbs which develop generic uses
typically exhibit a close semantic, collocational binding with a limited number
of objects (e.g. read, drive, drink). Middles differ from these three patterns in
that they have become productive with verbs from most semantic classes in the
course of ModE.
Compared to the respective transitive uses, co-referential intransitive,
middle, generic and ergative uses in English affect the agentivity of the state of
affairs: in co-referential intransitive use (John washed) the semantic function of
the subject is ambivalent to an interpretation as Agent or Goal. In middle use
(This book reads well) the Zero-subject corresponds to the Goal-object in the
respective transitive use and the notion of an implicit Agent is retained.
Generics (He is wise who reads), like middles, denote States; however, the Zero-
subject here corresponds to the Agent-subject of the respective transitive use.
Ergatives (The door opened) totally exclude the Agent and select the Processed-
function instead. The types of states of affairs encoded in each of these four con-
structions differ accordingly along the lines of dynamism, telicity and control.
With respect to the historical effects of these diverse processes on the
linguistic system of English, it seems evident that the co-referential intransitive
and ergative patterns change the predicate frame of individual verbs: they result
in homonymous or polysemous transitive–intransitive verb pairs which differ
in their semantic features and syntactic behaviour, accordingly. The generic
pattern and that of middle use do not change the verbs’ predicate frames;
rather, the representation of the state of affairs as a whole is affected by these
<DEST "moh-n*">

248 Ruth Möhlig and Monika Klages

processes of detransitivization. This results in a difference in perspective and

aspect. Since the ergative pattern and that of middle use lead to different
structural results, we would like to cast some doubt on the frequently stated
opinion that middle use is a special case of ergative use (e.g. Denison 1993: 392;
Iwata 1999: 547–548).
Our investigation has shown that semantic properties and processes play a
decisive role in syntactic change. For a fuller account of the semantic and
pragmatic links which hold between the various transitive and non-transitive
constructions within the course of their development, detailed, corpus-based
studies of the various patterns in the different periods of English are necessary.
Given a larger database it would be possible to account in greater detail for the
various factors and structural dependencies which influence the occurrence and
functions of certain syntactic patterns (e.g. phonological and morphological
processes, influence from other syntactic constructions or even languages).
Future investigations should also focus on processes of transitivization, as both
transitivization and detransitivization play together in the evolution of transi-
tivity alternations in English.

Notes

* We would like to thank Jon Erickson, Manfred Görlach, Frank Heidermanns, Marianne
Hundt and, last but not least, the editors of this volume and an anonymous reviewer, whose
valuable comments and suggestions greatly contributed to this paper.
1. E.g. OE helpan ‘help’ + dat., Germ. helfen ‘help’ + dat., OE derian ‘harm’ + dat., Germ.
schaden ‘harm’ + dat.; OE ābirgan ‘taste’ + gen. (Germ. schmecken ‘taste’ + nach + dat., cf.
ModE to taste of sth.), OE fēlan ‘feel’ + gen., etc. Some verbs occur sometimes with an
accusative and sometimes with a dative, e.g. ābelgan ‘anger’ (Mitchell 1985: 232).
2. Dik (1989, 1997) introduces Goal to describe the affected participant and default function
of the second argument of trivalent predicates. It thus corresponds to the notion of Patient
as introduced by Bloomfield and used, for example, in Givón (1984/1990). We follow Dik for
his distinction between the functions Goal and Processed (the non-agentive function of the
first argument in predications denoting dynamic states of affairs; e.g. The door opened).
3. Hopper & Thompson (1980: 251) use the label [action] instead of [dynamism], the latter
being taken over by us from Dik (1997: 105–126). Please note that the functional and
semantic typology of participants and states of affairs used by us is primarily based on Dik
(1989, 1997).
4. Cf. Ikegami (1988: 389–390), who defines transitive verbs as those placing focus on the
achievement of the intended goal of the action rather than on the action itself.
5. For a discussion of the gradual loss of inflectional case marking see Allen (1995:158–220).
 Detransitivization in the history of English 249

6. Of course, drench is still used in the meaning ‘soak’, e.g. I have to change, I’m drenched; I
got drenched in the rain.
7. The cognates of OE openian in older Germanic already show a conflation of the inchoative
and causative readings, i.e. Old Icelandic opna ‘make/become open’, Old Frisian epenia
‘make/become open’, Old Saxon opanon ‘make/become open’, Old High German offanōn
‘make open; make known’ (Heidermanns 1993:640). Schaefer (1984:357) considers both the
inchoative and causative frame of deadjectival Gmc. -ō-verbs as old.
8. Henry (1995: 50–80) describes imperatives “involving a postverbal subject” (ibid.: 50) for
contemporary Belfast English, e.g. “Run youse to the telephone” (ibid.:51). This would imply
that Belfast English has retained an archaic feature which has been lost in Standard English.
9. This loss of distinctiveness of syntactic-semantic verbal categories which in Old English
were still morphologically marked is paralleled by the general loss of distinctive marking of
word categories, e.g. of verb vs. adjective (open v. : open adj.), noun vs. verb (love n. : love v.)
and noun vs. adjective (light n. : light adj.) in the history of English. Few Gmc. languages have
gone as far in the loss of inflexional and derivational affixes as English, which has led to the
‘homonymic clash’ of different lexemes belonging to the same or different word-classes.
10. Gamillscheg (1928) has OFr. drecier v. ‘direct’ (11C) < VLat. *directiare, which is derived
from Lat. directus. Rothwell et al. (1992: 197) gloss Anglo-Norman drescer v. as ‘direct’, and
reflexive Issi se dresce tut a leyser (Mar 491) as ‘make oneself ready’, suggesting that both
meanings of ME dresse(n) were borrowed from French.
11. Following general usage (e.g. Denison 1993), we apply the term ‘ergative’ to this type of
construction which is sometimes called ‘anticausative’ (e.g. Siewierska 1984) or ‘non-
ergative’ (e.g. Halliday 1994). For discussions of the diachronic development of the four
syntactic types discussed in this paper cf. Jespersen (1927); Deutschbein (1931); Visser
(1963–1973); Denison (1993).
12. Dik’s (1997: 59, 78–82) concept of predicate frames refers to the property of verbal
predicates as part of structures which form “a kind of ‘blueprint’ for the predications in
which they can be used (…). Each predicate frame specifies the form (…), the type (…), and
the valency or argument structure of the predicate” (1997: 59; see also Fillmore’s (1968)
concept of case frames).
13. By reflexivity we understand the syntactic marking of clauses by means of a reflexive
pronoun. Semantically, we can distinguish emphatic use, which anaphorically stresses the
role of Agent (He opened the door himself), and co-referential use, which indicates that the
Goal of the state of affairs is identical with the Agent (He washed himself; cf. Givón 1993: 90).
Only co-referential reflexivity is part of our present investigation. We have chosen the term
‘co-referential intransitive’ to indicate that verbs omitting the co-referential reflexive marker
in their intransitive use are still perceived as co-referring to the Agent-subject.
14. Conceptually, the Agent will always be selected as a first choice for the subject position,
since it ranks highest in the cognitive scale of participants (cf. Givón 1984: 139–140).
15. But as Görlach (1999: 78) points out, “[m]any grammarians have noted an increase” in
this pattern in the nineteenth century, as well as in other patterns related to processes of
transitivization or detransitivization.
250 Ruth Möhlig and Monika Klages

16. The verb shave < OE sceafan ‘scrape away the surface of sth.’ (cf. Germ. schaben)
developed a reflexive use only with the specialized meaning of ‘scrape away the hair of some
person (or animal)’ (14C), and even then it took some time until the first co-referential
intransitive uses appeared (18C).
17. Deutschbein (1931: 100–101) further explains along these lines OE transitive–intransitive
verbs such as wendan (trans., reflex., intrans.) ‘make sth. turn/direct sth. > make oneself
turn/direct oneself to > go’, fysan (trans., reflex., intrans.) ‘make/urge sb. go on his way/in a
direction > make/urge oneself (go) on the way/in a direction > haste’, and ætiewan ‘reveal/
show sth. > show oneself > appear’. Furthermore, he cites openian ‘open sth. > open oneself
> become open’ among this group, which does not, however, show an intermediate co-
referential reflexive stage in the OE data, but only the emphatic reflexive use (cf. Ogura
1989: 99–100; cf. also Section 2).
18. As is typically the case with intransitive verbs, the focus of intransitive wash is more on
the action itself than on the dimension of extension typically encoded in transitive verbs
(Halliday 1994: 110).
19. In Old English, co-referential reflexivity is expressed by means of the personal pronoun,
which may be used to refer back to the subject. Sometimes it is emphasized by the pronoun
self; the combination of the personal pronoun with self forms the reflexive pronoun we know
from Middle English onwards in emphatic as well as co-referential use. Purely emphatic,
non-coreferential reflexivity is expressed by means of the emphatic pronoun self without a
personal pronoun in Old English (cf. Mitchell 1985: 112; Ogura 1989: 1–2). It is interesting
to note that many, if not most, co-referential reflexive verbs which did not develop a co-
referential intransitive use are found with the personal pronoun + emphatic self in Old
English, whereas verbs of the co-referential intransitive pattern are usually found with the
bare personal pronoun when used reflexively in Old English.
20. Characteristic properties such as expressed in John is tall ~ John drinks (= John is an
alcoholic); given truths = given by virtue of world knowledge.
21. Unfortunately, Dik (1989, 1997) does not distinguish between Goal and, for instance,
Theme, which is in some typologies used for an entity that is moved by an action or whose
location is described, but which does not undergo a change of state (Saeed 1997: 140; for a
slightly different use of the term cf. Langacker 1991: 287–288). With respect to this, Dik’s
typology should probably be modified. Dik himself does, however, distinguish between
prototypical and less prototypical Goals. Also cf. Kastovsky (1973: 266), who calls verbs like
read, eat, etc., which are not susceptible to a causative interpretation, “basically” or “inher-
ently transitive”.
22. The choice of the first participant here depends on whether the event is experienced as
a dynamic process (= ProcessedExperiencer) or a non-dynamic state (= ZeroExperiencer).
23. Note that generic use of to read (lectures) has led to the polysemous meaning ‘teach’ for
read (e.g. He reads at Oxford).
24. Generic use bears some resemblance to the use of cognate objects with intransitive verbs
(e.g. She thought a terrible thought), in which the syntactic encoding of an (otherwise
redundant) Goal-object is important for pragmatic purposes, e.g. for presenting a qualitative
modification or specification of the event referred to (Klages 1996).
 Detransitivization in the history of English 251

25. The frequent association of ergative and middle constructions with syntactic reflexivity
(e.g. Givón 1984; Palmer 1994) is usually based on a comparison of different languages, e.g.
German and English. (Note that an ergative clause such as The door opened would in German
be rendered as a co-referential reflexive clause Die Tür öffnete sich (reflex. pron.)). In our
opinion, semantic or functional relationships between syntactic constructions may only be
postulated language-internally (cf. Wierzbicka 1996: 425–426). The diachronic syntactic-
semantic analysis of selected English verbs executed here does not yield much evidence for
a functional and semantic relationship between co-referential reflexive/intransitive and
ergative or middle uses. However, they may be linked by the concept of ‘middle voice’
(Kemmer 1993).
26. But see the contrary assumption in, for instance, Fillmore (1968:26), who regards the tran-
sitive use of open as secondary, in that only the Object case (e.g. the door in the door opened)
is obligatory in the frame, the Agent and Instrument being optional: open + [O (I) (A)].
27. The verb move developed an ergative use probably through metaphorical co-referential
intransitive uses of the verb with subjects denoting a Mass or Body, e.g. c1400 Destr. Troy
1601 “Thurgh myddis þe mekill toune meuyt a water, And disseuert þe Cite” (=‘through the
middle of the great town flows a river, and divides the city’; cf. OED s.v. move 16.a) and 1605
Shakes. Macb. v. v. 35 “Me thought The Wood began to moue” (OED s.v. move 16.a).
It may be argued that once a co-referential intransitive polysemous move had been
established in English, the combination with an inanimate subject after the pattern of the
many already existing ergative uses was just a logical consequence for such a prototypical
Motion verb as move.
28. Note that Deutschbein (1931: 100–101) interprets the development of OE intransitive
openian as one following the co-referential reflexive pattern. Our data, however, do not give
evidence for an intermediate co-referential reflexive use for Old English (cf. note 17).
Furthermore, the ·animateÒ Agent is lacking in OE instances of intransitive constructions
with openian, thus ruling out a possible development via the co-referential reflexive stage.
29. This adverb of manner modifies the state of affairs as a whole and not only the predicate,
such as in The door opened slowly (i.e. it is a predication manner satellite according to Dik’s
1989, 1997 typology).
30. According to Iwata (1999: 530–532), middle constructions typically occur in the present
tense with a generic (stative) meaning, but may also be used with other tenses to denote
specific (dynamic) events.
31. Exceptions are, for example, most Contact verbs, such as hit: ?The ball hits easily; beat:
?The drum beats well.

32. Before the progressive passive existed, the older passival (were bringing down) and the
non-progressive passive (were brought down) were used (Denison 1993: 439). The expanded
form — the predecessor of the ModE progressive aspect — has existed in English since OE
times, but in Old and Middle English it differed in function from the progressive, which
became grammaticalized by about 1700 (Fischer 1992: 254). The use of the progressive has
steadily increased from early in the ModE period onwards (Denison 1993: 143). For the
relationship between middles and the progressive passive see Denison (1993: 392–393).
252 Ruth Möhlig and Monika Klages

Passivals probably developed in Late Middle English (cf. Visser 1963–1973: §§168–169;
Fischer 1992: 256; Denison 1993: 390). They went out of use with the establishment of the
progressive passive (cf. Denison 1993:150). Visser calls passivals obsolete in ModE. However,
as M. Hundt points out, passivals have probably been retained at least in American English,
e.g. in “Casablanca is showing this weekend at the Alliance Theatre” (Miami Herald, 17 July
1992), and “CORONATION STREET goes onto its new timeslot this week, and is screening
alongside a blast from Rover’s past” (Evening Post, 27 December 1994), both quoted from
Hundt (1998: 116, 118). For a detailed discussion of the two patterns of passival use and
progressive passive and their respective developments see Hundt (2001).

References

Allen, Cynthia L. 1995. Case Marking and Reanalysis: Grammatical Relations from Old to
Early Modern English. Oxford: Clarendon.
Denison, David. 1993. English Historical Syntax: Verbal Constructions. London & New York:
Longman.
Deutschbein, Max. 1931 [4th ed.]. System der neuenglischen Syntax. Leipzig: Quelle & Meyer.
Dik, Simon. 1989. The Theory of Functional Grammar. Part I: The Structure of the Clause.
Dordrecht: Foris.
Dik, Simon. 1997. The Theory of Functional Grammar. Part I: The Structure of the Clause /
Part II: Complex and Derived Constructions. 2nd rev. ed. by Kees Hengeveld. Berlin &
New York: Mouton de Gruyter.
Faber, Pamela & Ricardo Mairal Usón. 1999. Constructing a Lexicon of English Verbs. Berlin
& New York: Mouton de Gruyter.
Fillmore, Charles J. 1968. “The Case for Case”. Repr. 1987 in Fillmore’s Case Grammar: A
Reader ed. by René Dirven & Günter Radden, 21–33 (= Studies in Descriptive Linguistics,
16). Heidelberg: Groos.
Fischer, Olga. 1992. “Syntax”. The Cambridge History of the English Language. Vol. II:
1066–1476 ed. by Norman F. Blake, 207–408. Cambridge: Cambridge University Press.
Gamillscheg, Ernst. 1928. Etymologisches Wörterbuch der französischen Sprache. Heidelberg:
Carl Winter Universitätsbuchhandlung.
Givón, Talmy. 1984/1990. Syntax: A Functional-Typological Introduction. 2 vols. Amsterdam
& Philadelphia: John Benjamins.
Givón, Talmy. 1993. English Grammar: A Function-Based Introduction. 2 vols. Amsterdam &
Philadelphia: John Benjamins.
Goldberg, Adele E. 1992. “The Inherent Semantics of Argument Structure: The Case of the
English Ditransitive Construction”. Cognitive Linguistics 3:1.37–74.
Görlach, Manfred. 1994 [2nd rev. ed.]. Einführung ins Frühneuenglische. Heidelberg: Winter.
Görlach, Manfred. 1997. The Linguistic History of English: An Introduction. London:
Macmillan.
Görlach, Manfred. 1999. English in Nineteenth-Century England: An Introduction. Cam-
bridge: Cambridge University Press.
 Detransitivization in the history of English 253

Halliday, M.A.K. 1994 [2nd ed.]. An Introduction to Functional Grammar. London: Edward
Arnold.
Haspelmath, Martin. 1993. “More on the Typology of Inchoative/Causative Verb Alterna-
tions”. Causatives and Transitivity ed. by Bernard Comrie & Maria Polinsky, 87–120.
Amsterdam & Philadelphia: John Benjamins.
Heidermanns, Frank. 1993. Etymologisches Wörterbuch der germanischen Primäradjektive.
Berlin & New York: Mouton de Gruyter.
The Helsinki Corpus of English Texts, Diachronic and Dialectal. 1991. Comp. by Matti
Rissanen, Ossi Ihalainen & Merja Kytö. Helsinki: Department of English, University of
Helsinki.
Henry, Alison. 1995. Belfast English and Standard English: Dialect Variation and Parameter
Setting. Oxford: Oxford University Press.
Hopper, Paul J. & Sandra A. Thompson. 1980. “Transitivity in Grammar and Discourse”.
Language 56.251–299.
Hundt, Marianne. 1998. New Zealand English Grammar: Fact or Fiction?. Amsterdam &
Philadelphia: John Benjamins.
Hundt, Marianne. 2001. “The Passival and the Progressive Passive — A Case Study of
Layering in the English Aspect and Voice Systems”. Paper presented at the Corgie
Symposium, Växjö, April 2001.
Ikegami, Yoshihiko. 1988. “Transitivity: Intransitivization vs Causativization: Some
Typological Considerations Concerning Verbs of Action”. On Language: Rhetorica,
Phonologica, Syntactica: Festschrift for Robert P. Stockwell from his Friends and Colleagues
ed. by Caroline Duncan-Rose & Theo Vennemann, 389–401. London: Routledge.
Iwata, Seizi. 1999. “On the Status of an Implicit Argument in Middles”. Journal of Linguistics
35.527–553.
Jespersen, Otto. 1927. A Modern English Grammar on Historical Principles. Part III: Syntax
(Second Volume). Copenhagen: Ejnar Munksgaard / Northampton: John Dickens & Co.
Ltd.
Kastovsky, Dieter. 1973. “Causatives”. Foundations of Language 10.255–315.
Kastovsky, Dieter. 1992. “Semantics and Vocabulary”. The Cambridge History of the English
Language. Vol. I: The Beginnings to 1066 ed. by Richard M. Hogg, 290–408. Cambridge:
Cambridge University Press.
Kemmer, Suzanne. 1993. The Middle Voice. Amsterdam & Philadelphia: John Benjamins.
Klages, Monika. 1996. “Act-effecting ‘Affectedness’: A Semantic Parameter for Argument
Selection and Semantic Function Assignment”. Paper presented at the VIIth Interna-
tional Conference on Functional Grammar, Córdoba, September 1996.
Krahe, Hans & Wolfgang Meid. 1969 [7th rev. ed.]. Germanische Sprachwissenschaft: III,
Wortbildungslehre. (= Sammlung Göschen, 2234). Berlin: Mouton de Gruyter.
Langacker, Ronald W. 1991 [3rd ed.]. Foundations of Cognitive Grammar. Vol. II: Descriptive
Applications. Stanford, Calif.: Stanford University Press.
Lightfoot, David W. 1979. Principles of Diachronic Syntax. Cambridge: Cambridge University
Press.
A Microfiche Concordance to Old English. 1985. Comp. by Antonette diPaolo Healey &
Richard L. Venezky. Toronto: Pontifical Institute of Mediaeval Studies, University of
Toronto.
</TARGET "moh">

254 Ruth Möhlig and Monika Klages

Middle English Dictionary. 1952–2001. Ed. by Hans Kurath, Sherman M. Kuhn & Robert
Lewis. Ann Arbor, Mich.: University of Michigan Press.
Mitchell, Bruce. 1985. Old English Syntax. 2 vols. Oxford: Clarendon.
Ogura, Michiko. 1989. Verbs with the Reflexive Pronoun and Constructions with Self in Old
and Early Middle English. Cambridge: Brewer.
The Oxford English Dictionary on CD-ROM. 1989 [2nd ed.]. Ed. by John A. Simpson &
Edmund S.C. Weiner. Oxford: Oxford University Press.
Palmer, Frank R. 1994. Grammatical Roles and Relations. Cambridge: Cambridge University
Press.
Radford, Andrew. 1997. Syntactic Theory and the Structure of English: A Minimalist Approach.
Cambridge: Cambridge University Press.
Rissanen, Matti. 1999. “Syntax”. The Cambridge History of the English Language. Vol. III:
1476–1776 ed. by Roger Lass, 187–331. Cambridge: Cambridge University Press.
Rothwell, William, Louise W. Stone & Thomas B.W. Reid, eds. 1977–1992. Anglo-Norman
Dictionary. London: The Modern Humanities Research Association.
Saeed, John I. 1997. Semantics. Oxford: Blackwell.
Schaefer, Christiane. 1984. “Zur semantischen Klassifizierung germanischer denominaler ōn-
Verben”. Sprachwissenschaft 9.356–383.
Siewierska, Anna. 1984. The Passive: A Comparative Linguistic Analysis. London: Croom
Helm.
Traugott, Elizabeth Closs. 1992. “Syntax”. The Cambridge History of the English Language.
Vol. I: The Beginnings to 1066 ed. by Richard M. Hogg, 200–289. Cambridge: Cambridge
University Press.
Visser, Frederik Th. 1963–1973. An Historical Syntax of the English Language. 3 parts in 4
vols. Leiden: E.J. Brill.
Wartburg, Walther von. 1967. Französisches etymologisches Wörterbuch: Eine Darstellung des
galloromanischen Sprachschatzes. 24 vols. Basel: Zbinden Druck und Verlag AG.
Wierzbicka, Anna. 1996. Semantics: Primes and Universals. Oxford: Oxford University Press.
<LINK "sch-n*">

<TARGET "sch" DOCINFO

AUTHOR "Julia Schlüter"

TITLE "Morphology recycled"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Morphology recycled
The Principle of Rhythmic Alternation
at work in Early and Late Modern English
grammatical variation*

Julia Schlüter
University of Paderborn

1. Introduction

The present contribution is part of a research programme that explores linguis-

tic variability as a source of new insights into the workings of functional factors
in language. It is assumed that grammatical variation and change are deter-
mined by a network of interacting factors which may reinforce or counteract
each other. Some factors that deserve mention here are, among others, semantic
tendencies, stylistic biases, cognitive complexity, avoidance strategies, frequency
and requirements of information structure. The focus of this paper is on
phonological factors, which have been frequently neglected or even ruled out as
determinants of grammatical variation. In particular, I will concentrate on the
Principle of Rhythmic Alternation, whose effects will be illustrated with
reference to a number of variation phenomena.
For the Principle of Rhythmic Alternation, Selkirk (1984: 37) proposes the
following definition: “[t]here is arguably a universal rhythmic ideal, one that
favors a strict alternation of strong and weak beats”. It follows from this that an
ideal rhythm consists of a stressed syllable followed by an unstressed syllable,
followed by a stressed syllable again, etc. Put differently, both sequences of
stressed syllables (‘stress clashes’) and sequences of (more than two) unstressed
syllables (‘stress lapses’)1 tend to be avoided.
Making reference to the Principle of Rhythmic Alternation, the data
reviewed in this paper support the claim that in areas involving rhythmically
different grammatical variants, those variants which give rise to a regular
256 Julia Schlüter

succession of stressed and unstressed syllables will be preferred over those

leading to stress clashes or lapses. Relevant effects of this type have been pointed
out by Fijn van Draat (1910; 1912a; 1912b) and Bolinger (e.g. 1965) for English
prose as well as by Stroheker (1913), Bihl (1916) and Franz (1939) for versified
language. I aim to buttress these authors’ findings by providing quantifiable
empirical evidence for the effectiveness of this principle in explaining the
determination of grammatical choices and linguistic change.
All of the analyses to be presented in the following subsections illustrate
cases in which grammatical morphemes and markers are available as rhythmic
buffers that can be optionally inserted in accordance with the Principle of
Rhythmic Alternation. Section 2.1 takes a look at the rhythmic conditioning of
the gradual loss of the formerly obligatory participial suffix -en, showing that
the past participle drunken was preserved in certain rhythmically defined
environments. Section 2.2 focuses on the establishment of the adverbial suffix
-ly and the facilitating or inhibiting role of rhythm in this respect. The adverb
scarcely, variably replaced by its suffixless counterpart scarce, is chosen to
exemplify this process of linguistic change. Section 2.3 deals with the variable
marking of infinitives dependent on the verb make in the passive. It will be
demonstrated that, as long as the infinitive marker is not obligatory in this
construction, the influence of rhythm manifests itself by slowing down the rate
of change in favourable contexts. Finally, Section 2.4 analyses the variable
presence of the a-prefix in -ing forms associated with the verb set. While English
usage exhibits an up-and-down evolution in this respect, the effect of rhythmic
alternation is noticeable at every historical stage. All of these studies are
concerned with intermediate phases of language change in which morphemes
and markers are no longer or not yet quite obligatory in terms of grammatical
motivations. It is precisely in such phases of grammatical indeterminacy that
the Principle of Rhythmic Alternation assumes the role of an influential
determinant.

2. Empirical studies

Methodologically, the present work is based on a large collection of historical

corpora (details are given in the references). The database covers the sixteenth
to nineteenth centuries, i.e. the Early and Late Modern English periods.
However, where appropriate, the data are placed in a wider context, ranging
from Middle English to late twentieth-century English.
 Morphology recycled 257

2.1 The distribution of mono- and disyllabic past participle variants

Even today, variation in the form of the past participle of irregular verbs is still
a frequently encountered phenomenon. Variants belong to different verb types
and involve ablauted, suffixless forms and syncopated weak forms which may
or may not be (re-)suffixed with -en or -ed. In many of the relevant cases the
variants differ not only in their morphological shape, but also in their prosodic
properties, i.e. a monosyllabic form as opposed to a disyllabic form, equipped
with an additional participial suffix. The list of examples includes drunk/
drunken, shrunk/shrunken, sunk/sunken, struck/stricken, swelled/swollen, shaved/
shaven, fit/fitted, knit/knitted, quit/quitted, lit/lighted, chid/chidden, trod/trodden,
got/gotten, etc. In earlier forms of English (as well as in modern dialects), the
number of variants was considerably greater (cf. Fijn van Draat 1912a: 27ff.;
Stroheker 1913: 42ff.; Franz 1939: 166ff.; Bolinger 1965: 145ff.).
The following empirical study focuses on the past participles of the verb
drink and investigates the variable preservation of the Old English participial
suffix -en, which underwent a gradual process of degrammaticalization, having
been omissible in the participle since the thirteenth century (OED s.v. drink v.1).2
The database is composed of the Middle English section of the Helsinki Corpus
(HC; 1150–1500; 0.6 million words), the Early English Prose Fiction corpus
(EEPF; 1518–1700; 9.6 million words), the Eighteenth-Century Fiction corpus
(ECF; 1705–1780; 11.2 million words),3 the Nineteenth-Century Fiction corpus
(NCF; 1782–1903; 37.6 million words), and the British National Corpus (BNC)
imaginative prose section (1964–1993; 19.7 million words). All occurrences of
drunk and drunken are classified according to their syntactic functions. The first
category, labelled ‘simple attributive uses’, contains all cases in which the
participles immediately precede a noun and are not premodified themselves; cf.
example (1):
(1) The drúnken wrétch fell off the sofa, and fell on to the floor, where he
stayed; (William M. Thackeray: Catherine, 1839–1840; NCF)4

The second category comprises all uses that are not simple attributive ones. This
involves, in particular, the large number of non-attributive occurrences of
drunk and drunken, such as postnominal uses, as in example (2), predicative
uses, as in example (3), nominalized uses, as in example (4),5 and verbal uses,
as in example (5):
258 Julia Schlüter

(2) The beating of drums, the rattle of tomtoms, and the yells and howls of
the rebels, drúnk with opium and with bang, were enough to remind us
all the night of our dangerous neighbours across the stream.
(Sir Arthur Conan Doyle: The Sign of Four, 1890; NCF)
(3) “I would lie a wager they are all drúnk in an hour,” said he, (…)
(Meadows Taylor: Confessions of a Thug, 1839; NCF)
(4) Long horse-hair settles for the drúnk, with horse-hair pillows at each
end. (Thomas Hardy: Far from the Madding Crowd, 1874; NCF)
(5) At tea, two or three hours earlier, they had, in the freakishness of affec-
tion, drúnk from one cup.
(Thomas Hardy: Tess of the D’Urbervilles, 1891; NCF)

The syntactic classification adopted here closely correlates with certain prosodic
constellations: prenominal contexts constitute the rhythmically most critical
contexts for attributive material ending in a stressed syllable, since nouns in
English typically carry initial stress.6 Thus, Bolinger (1965: 146f.; cf. also Fijn
van Draat 1912a: 24; Franz 1939: 167) predicts that the disyllabic form pre-
serving the Old English participial suffix -en will be particularly favoured in
attributive contexts, where the suffix can serve to prevent a stress clash.7 An
illustrative example is given in sentence (1) above. By contrast, the participle in
non-attributive uses is usually either followed by an unstressed function word or
by a pause,8 so that stress clashes are rare. This is the case in examples (2) to (5),
which are representative of most of the relevant cases. The results of the count are
summarized in Figure 1, which lists the corpora in their chronological order.
As can be seen in the comparison of the percentages of drunk and drunken
in simple attributive and other contexts, the expectations are fully confirmed.
At every point in the history from Middle English to Late Modern English,
drunken was preserved in close to 100% of the critical attributive cases. It is only
in the late twentieth century that the older form showed signs of an incipient
demise in simple attributive uses.9 By contrast, in all other uses, its frequency
suddenly dropped from 40% in Middle English to values between 11 and 1% in
the later periods. The stark contrast between simple attributive and other uses
found as early as Middle English maintained a quasi-categorical status all
through Early and Late Modern English.
However, some qualifications of the binary opposition attributive vs. other
uses are in order. In contrast to the earlier centuries, the nineteenth century
witnessed the growth of more complex attributive structures. Thus, the Nine-
teenth-Century Fiction corpus yields 13 compounds in attributive uses involving
 Morphology recycled 259

3/3 98/98 26/26 83/84 158/158 181/183 238/276

=100% =100% =100% =99% =100% =99% =87%
100%

80%
percentage of drunken

60%
6/15
=40% simple attributive uses
40% other uses

36/434 63/560
20% 21/285 26/476
=8% 2/261 =11% 22/1250
=7% =5%
=1% =2%
0%

BNC wridom 1
(1964-1993)
EEPF
(1518-1700)

ECF
(1705-1780)

NCF
(1782-1830)

NCF
(1831-1860)

NCF
(1861-1903)
HC ME
(1150-1500)

Figure 1. The distribution of the participial variants drunk and drunken in a series of
prose corpora.

the past participle, three of which contain the monosyllabic variant drunk (half-
drunk, sleep-drunk and joy-drunk, as in example (6), alongside ten instances of
half-drunken). By virtue of the English compound stress rule, the lexical stress
in such complex lexemes regularly falls on the first element, so that even the
monosyllabic form of the participle causes no rhythmic problem. Similarly, in
two cases the participle is itself premodified by an adverb (newly drunk and
partially drunk, as in example (7)). This constellation affords the possibility of
shifting the stress leftwards to the premodifying adverb, so that the participle
merely retains secondary stress and the clash with the following noun is
mitigated (for the stress shift rule, cf. Selkirk 1984: 169). In the twentieth-
century corpus, the trend towards the monosyllabic form in complex attributive
structures is even more pronounced: in 16 out of 24 cases, drunk is used (10
compounds, among them 6 instances of half-drunk and 4 instances of punch-
drunk, and 6 adverbially premodified participles). Of the remaining 8 cases of
drunken, 3 are compounded (all of them half-drunken), and 5 adverbially
premodified. This distribution suggests that monosyllabic drunk is accepted as
fully grammatical in attributive structures, but only on the condition that these
are expanded by additional material and thereby rhythmically disencumbered.
260 Julia Schlüter

(6) she yields herself to the influences around her; the soft and sugared air,
the jóy-drunk lárks, the juicy grass fields thronged with bold dandelions
and faint ladies’-smocks. (Rhoda Broughton: Belinda, 1883; NCF)
(7) an exceedingly dirty and pártially drùnk mínister of justice asked me if I
would like to step in and hear a trial or so:
(Charles Dickens: Great Expectations, 1861; NCF)

Composite attributive structures in which attributive drunk(en) does not

directly precede the noun it modifies equally reduce the danger of a stress clash
since the participle is followed by a pause (sometimes indicated by a comma) or
by unstressed and or or, as is the case in example (8):
(8) the stranger who preceded them divided the press, shouldering from
him, by the mere weight and impetus of his motion, both drúnk and
sóber pássengers. (Sir Walter Scott: Guy Mannering, 1830; NCF)

Due to the fact that compounded, adverbially premodified and composite

attributive structures are usually unproblematic with respect to the alternation
of stressed and unstressed syllables, these are subsumed under the category
‘other uses’ of Figure 1. The two groups ‘simple attributive’ and ‘other uses’ are
thus highly correlated with contexts representing a high and a low potential for
stress clashes respectively. The obvious explanation for the continued preserva-
tion of drunken in simple attributive uses invokes the Principle of Rhythmic
Alternation, which favours an initially stressed disyllabic form preceding
stressed syllables and thus secures an ideal rhythmic pattern.
The wide discrepancy between the two curves in Figure 1 suggests that the
Principle of Rhythmic Alternation has led to the loss of free variation between
the participial variants during a long interim of almost five centuries. The
functional split has however not gone to completion: the nineteenth- and
twentieth-century data show that the use of drunk is licensed and widespread in
attributive functions wherever material expanding the attributive structure
conspires to secure an alternating rhythm. Thus, the morpheme -en, degram-
maticalized in Middle English (in the sense that it was abandoned as an
obligatory participial marker), has not entirely become regrammaticalized as a
categorical ending for attributive occurrences of the participle, but primarily
subserves rhythmic (i.e. extra-grammatical) preferences instead.
In conclusion, the Old English suffix -en has been able to maintain its
presence in Modern English thanks to its suitability as a rhythmic buffer in
attributive contexts. The apparent near grammaticalization observable in all
 Morphology recycled 261

corpus sections from Early Modern English onwards may be attributed to the
diachronic influence of the Principle of Rhythmic Alternation. Even so, rhythmic
alternation operates in conjunction with the English compound stress rule and the
stress shift rule to license the monosyllabic form in more complex attributive
structures. The evolution of the disyllabic form drunken illustrates a phenomenon
which Lass (1990; 1997:316ff.) refers to as ‘linguistic exaptation’ (cf. also Aitchison
1991: 148ff.; Brinton & Stein 1995; Vincent 1995): that is, a morpheme that used
to code a grammatical function loses this function, but the morphological
material itself is still around and is put to a different use — one which may be
just as systematic. Instead of marking past participles, the -en suffix has from
Middle English times onwards been recycled to buffer stress clashes.

2.2 Adverbs with and without the -ly suffix

The -ly suffix, used to derive adverbs from adjectives, was fully established in
this function in Middle English, but even today it may be missing in de-
adjectival adverbs in non-standard and even standard language, provided that
the circumstances favour its omission (cf. Fijn van Draat 1910: 96ff.; Stroheker
1913: 49; Franz 1939: 223; Nevalainen 1997). Syntactic and stylistic factors play
an important role in this respect, but here I focus on the rhythmic aspect of
-ly-less adverbs.
Exemplary cases of -ly omission include a number of intensifying adverbs,
popular in earlier forms of English, which boast an unstressed final syllable, e.g.
excessive, wondrous, wonderful and proper. My corpus studies have confirmed that
the -ly-less forms frequently served as adverbs, even before items beginning with
a stressed syllable, since they never caused a stress clash with the following element:
(9) The execution after the Route was excéssive blóody, the Romans remem-
bring how freely the Affricans had open’d their veines, were not
vngratefull in their returne.
(Roger Boyle: Parthenissa, Part 2, 1655; EEPF)

Parallel results have been obtained for several non-finally stressed adjectives
which may serve as adverbial intensifiers without the addition of a -ly suffix
even in Present-day English, e.g. devilish clever, hellish good, hopping mad,
boiling hot, etc. In these cases, the presence of -ly offers no rhythmic advantage,
since the base forms in -ish and -ing have an unstressed final syllable. Indeed, it
would create the undesirable structure of a stress lapse. It is for this reason,
262 Julia Schlüter

among others, that these patterns are still very productive in contemporary
usage (see Adams 1973: 98).
An interesting and longstanding case in point is provided by the alternation
between scarce and scarcely:10 both can be used as adverbs without a difference
in meaning (cf. Stroheker 1913: 50). The following corpus study considers only
those cases in which scarce and scarcely immediately precede and premodify full
verbs (excluding auxiliaries and cases with intervening material). These full verbs
can be subdivided into initially stressed ones, as in example (10), non-initially
stressed ones, as in example (11), particle verbs, which are stressed on the particle,
such as pull off in example (12), as well as verbs, such as kiss in (13), which are
normally initially stressed (or monosyllabic), but which produce at most a
mitigated stress clash in the corpus examples because their stress is absorbed by
a stronger emphasis on the following material and is thereby reduced:11
(10) (…) I thereupon told him my name, which he had scárcely héard, but I
found my selfe in his Armes, as a reward of that discovery;
(Roger Boyle: Parthenissa, Part 4, 1655; EEPF)
(11) The generous Atafernes did scárce abándon me one moment, knowing
my condition needed all the consolation, though it was above all service
of a Friend. (Roger Boyle: Parthenissa, Part 5, 1656; EEPF)
(12) that Maharball wanted tyme to answer this Civility, and had scárce pulld
óff his owne Scarfe and taken his Freinds, before they were come up;
(Roger Boyle: Parthenissa, Part 2, 1655; EEPF)
(13) I had scárce kìst his hánds as an acknowledgement of his favour and my
Ioy but Batiatus came in, (Roger Boyle: Parthenissa, Part 1, 1655; EEPF)

The analysis of the data is guided by the hypothesis that, as in the examples
quoted above, the suffixed variant scarcely will be favoured in contexts involving
a subsequent stressed syllable (e.g. example (10)). By contrast, the suffixless
variant scarce is expected to occur where an unstressed or secondarily stressed
syllable follows the adverb and there is thus no need to avoid a stress clash (e.g.
examples (11)–(13)). This hypothesis has been tested against a large corpus of
fictional texts dating from 1518 to Present-day English (including the same
corpora as the previous case study, except for the Middle English section, which
contained too few examples).
The results reveal that the use of the two variants differs widely between
individual authors: in the Early English Prose Fiction corpus (comprising the
sixteenth and seventeenth centuries), the majority (58 authors) have scarce only;
 Morphology recycled 263

four authors have scarcely only and the remaining 42 authors fluctuate between
both variants. In the Eighteenth-Century Fiction corpus, six authors use scarce
only; none uses scarcely only, but the vast majority of 26 authors use both scarce
and scarcely. In the Nineteenth-Century Fiction corpus, 66 authors alternate
between scarce and scarcely and 32 use scarcely exclusively.12 Finally, of the
several hundreds of authors whose works make up the imaginative prose
section of the BNC, only five use scarce as an adverb at all. All of these are
authors of novels set wholly or partly in earlier centuries, employing an archaic
style on purpose. While in the sixteenth to eighteenth centuries, scarcely appears
in only 13 to 17% of all adverbial uses, it makes up 76% in the nineteenth
century and reaches 97% (respectively 100%, discounting the historical novels)
in the late twentieth-century corpus. There is thus a constant rise in the
marking of the adverb in the course of five centuries.
In the works of most authors who alternate between the two forms of the
adverb, the distribution of the variants is neutral with regard to rhythm;
probably other factors than rhythm prevail in their selection. However, the
handful of authors given in Figure 2 seem to be distributing the variants
according to the requirements of the Principle of Rhythmic Alternation. In this
study, only the works of those authors are included which contain at least two
relevant instances of both scarce and scarcely. Even so, due to the low numbers
of occurrence, most results for individual authors are far from attaining
statistical significance. Nevertheless, the figures are highly suggestive and allow
interesting conclusions.
The solid curve, ‘scarcely preceding a stressed syllable’, refers to all instances
of the adverb immediately preceding initially stressed verbs; the dashed curve,
‘scarcely preceding an unstressed syllable’, subsumes both non-initially stressed
verbs, particle verbs and rhythmically overshadowed verbs following the adverb.
As it turns out, for every author scarcely occurs in a higher percentage of
cases preceding stressed syllables than preceding unstressed syllables. Converse-
ly, the percentage of scarce soars before unstressed syllables. Thus, conformity
with the Principle of Rhythmic Alternation is achieved in the majority of cases.
Figure 2 shows that individual authors behave very differently. In the text
by Brian Melbancke (sixteenth century) there are only seven instances of the
adverb preceding and modifying a verb: scarcely exclusively precedes initially
stressed verbs, while scarce precedes verbs that are either initially stressed, non-
initially stressed or rhythmically overshadowed. Like Melbancke, the anony-
mous seventeenth-century author of Fortunatus uses scarce preceding an
unstressed syllable. In the nineteenth century, the same is true of Charles
264 Julia Schlüter

preceding a stressed syllable

preceding an unstressed syllable

13/13 11/11 1/1 3/3

=100% =100% =100% =100%
100%
25/32
5/7 53/68=78% =78% 5/7
80% 2/3 =71% 3/5 =71%
=67%
percentage of scarcely

8/13 3/5 15/20 =60%

=62% =60% 4/6 =75%
60%
=67%
18/32
2/4 =56%
40% =50%
2/8 25/80
=25% =31% 10/40 5/13
=38% 11/53
20% =25% =21% 1/6
1/5=20% =17%
1/12=8%
0/3=0% 0/1=0% 0/1=0% 1/21=5%
0%
Anon.:Fortunatus (1682)

Robert Louis Stevenson (1883-1896)

Roger Boyle (1655-1669)

Philip Ayres (1696)

Fanny Burney (1782-1796)

Lady Caroline Lamb (1816)

Charles Lever (1844)

Charles Kingsley (1850-1866)

Richard D. Blackmore (1869)
Brian Melbancke (1583)

Lady Sydney Morgan (1807)

Philippa Wiat (1990)

P.C. Doherty (1992)

Percy Bysshe Shelley (1810-1811)

16th/17th centuries 18th 19th century 20th

c. century

Figure 2. The distribution of scarce and scarcely immediately preceding full verbs in the
works of selected authors in a series of prose corpora.

Kingsley and — with one exception — of Richard D. Blackmore and Robert

Louis Stevenson. As late as that century, Robert Louis Stevenson still uses scarce
as the most common, unmarked form preceding all types of verbs,13 but even so
makes a slight difference between environments with a following stressed or
unstressed syllable. Lady Morgan and Percy Bysshe Shelley in the nineteenth
century and Philippa Wiat and P. C. Doherty in the twentieth century never let
 Morphology recycled 265

adverbial scarce precede an initially stressed verb. In the works of all authors, the
Principle of Rhythmic Alternation only manifests itself in a more or less
reduced incidence of scarce preceding initially stressed verbs. But within these
clearly recognizable tendencies, there is a margin for free variation, which seems
to be determined by factors outside the domain of prosody (e.g. processing
tendencies).
One strategy which probably contributes to deviations from a perfect
alternating rhythm may be a striving for parallelism, favouring the same
morphological form for a sequence of two adverbs fulfilling the same syntactic
function. Consider examples (14) and (15), in which one non-initially stressed
and one initially stressed full verb are coordinated. A rhythmically-guided
choice of scarce in one case and scarcely in the other would have led to a striking
structural inconsistency:
(14) She had scárce recéived it, scárce pláced it in her bosom, when Lady Mar-
garet attached her. (Lady Caroline Lamb: Glenarvon, 1816; NCF)
(15) He then held out to her his hand, which she could scárcely appróach from
trembling, and scárcely kíss for weeping,
(Fanny Burney: Camilla, 1796; NCF)

Useful as it may be in terms of rhythmic alternation, the lack of codification of

adverbial forms runs counter to standardization tendencies gaining ground in
the eighteenth century (cf. Franz 1939: 224). These forces favour a single
morphological form for each syntactic function, but they are generally insensi-
tive to the requirements of the Principle of Rhythmic Alternation. As a result,
in Present-day English prose scarce regularly fills the adjective slots, while —
with very few exceptions — scarcely has become the dominant adverbial form.
These exceptions occur almost exclusively in cases where the omission of -ly
serves to create a deliberately archaizing style, but causes no rhythmic inconve-
nience. The latter fact supports the view that the avoidance of stress clashes is a
universal factor which is at work at all times, but visible only to the extent that
it is not neutralized by contrary forces.
In conclusion, the adverbial suffix -ly can function as an accentual buffer,
thus averting threatening successions of strongly stressed syllables. Conversely,
it may be omitted where a buffer element is unnecessary because an unstressed
syllable separating two stressed ones is already present. Unlike the case of the
participial suffix -en, this phenomenon cannot be properly described as an exaptive
change: while the participial suffix was on its way out when it was recycled for
rhythmical purposes, the adverbial suffix was actually in the process of establishing
266 Julia Schlüter

itself. Even so, it replicates the process by which essentially grammatical

morphemes receive a subsidiary functional motivation as rhythmic buffers.

2.3 The variable marking of infinitives dependent on manipulative verbs

In Early Modern English, it was possible for a number of manipulative (or
directive) verbs (e.g. make, bid, charge, command, entreat, forbid, let, see and
suffer) to be followed either by marked or by unmarked infinitives (cf. Stro-
heker 1913: 81ff.; Franz 1939: 539; Mustanoja 1960: 526ff.). The presence or
absence of the infinitive marker to was determined by a range of different
factors, notably semantic and syntactic ones.14 The following corpus analysis
focuses on infinitives dependent on the verb make in the passive (parallel results
have been found in connection with the verb bid; for both verbs cf. Rohdenburg
& Schlüter 2000: 484ff.). In the passive, the superordinate verb made is normally
immediately followed by the dependent infinitive and both verbs carry a lexical
stress. In contrast to the active (where object phrases intervene), passivization
is therefore liable to produce a high potential for stress clashes. Consider the
following examples:
(16) my eyes have before been máde to glísten by this soul-moving beauty;
(Samuel Richardson: Clarissa, 1748; ECF)
(17) my Father was the happiest Man in the World, and had nothing to vex
him, but the Enmity he was máde belíeve his Children had to him.
(Sarah Fielding: David Simple, 1744; ECF)

In both examples, the dependent verbs glisten and believe directly follow the super-
ordinate made. Since glisten in (16) carries initial stress, the infinitive marker to
comes in handy as a stress clash buffer. By contrast, believe in (17) has an un-
stressed initial syllable which makes an additional buffer syllable superfluous.
The present study investigates the influence exerted by the stress pattern of
the dependent verb on the variable presence of the infinitive marker. The
proposed hypothesis is that the marker will be absent relatively more often
when the dependent infinitive is non-initially stressed. Figure 3 presents the
results of a corpus analysis covering the familiar prose corpora from the sixteenth
to the nineteenth centuries. The data include only non-coordinated infinitives
or such that are initial in coordinated infinitival structures; i.e. all instances of
non-initial coordinated infinitives are excluded (the use of the infinitive marker
is governed by different principles here; cf. Ohlander 1941: 59; Rohdenburg &
Schlüter 2000: 450f.). All examples of made + infinitive have been categorized
 Morphology recycled 267

57/59 66/67 75/75 246/246 194/194

=97% =99% =100% =100% =100%
100%
46/46 97/97 97/97
=100% =100% =100%
percentage of marked infinitives

80% 46/51
=88%
17/24
60%
=71%

40%

20%

0%
EEPF (1518- ECF (1705- NCF (1782- NCF (1840- NCF (1870-
1700) 1780) 1839) 1869) 1903)

preceding an initially stressed infinitive

preceding a non-initially stressed infinitive

Figure 3. The variable marking of (initial or non-coordinated) infinitives dependent on

passive make in a series of corpora.

according to the presence or absence of initial stress on the infinitive.15

On average, in the sixteenth and seventeenth centuries 89% of the depen-
dent infinitives are marked by to. In the eighteenth century, this is already true
of 94%, and in the nineteenth century the corpus yields not a single occurrence
of an unmarked dependent infinitive. This shows that the diachronic range of
the corpus under consideration extends over the terminal phase of the to-less
construction. Nevertheless, in the earlier corpus sections, the curves for initially
and non-initially stressed infinitives take a clearly distinct course: as early as the
EEPF corpus, nearly all initially stressed infinitives are associated with the
infinitive marker. Two out of the total of three exceptions to this trend can be
explained with reference to the overriding prosodic movement of the sentence:
(18) as the yonge Fawne will be máde tàke bréad at a mans hand, when the old
Bucke will not by any meanes looke vpon a man:
(Austen Saker: Narbonus, 1580; EEPF)
(19) Being come to his House, he put me into the Garden to work, there I was
máde dràw Wáter, dig, and labour hard all day,
(Penelope Aubin: Charlotta Du Pont, 1739; ECF)
268 Julia Schlüter

In both cases, the dependent infinitives are immediately followed by object

expressions which absorb much of the stress from the infinitives, so that the
latter have a relatively weak stress in comparison to the adjacent syllables and
can themselves function as buffer syllables.
In contrast to the initially stressed infinitives, of the non-initially stressed
infinitives, seven (i.e. 29%) in the EEPF corpus and five (i.e. 12%) in the ECF
corpus are left unmarked. This contrast is explained by the Principle of Rhyth-
mic Alternation which does not require any buffer to be added before infinitives
that do not carry stress on their initial syllable (e.g. appear, confess, obey and
believe, as in example (17)). The corpus data prove that while the to-less
construction is petering out, rhythm and sentence prosody are two important
factors accounting for its latest occurrences.
As Figure 3 shows, the marked infinitive after passive made becomes
obligatory (grammaticalized) in the course of the eighteenth century. By
contrast, in the active uses of make, the unmarked infinitive is still unchallenged
up to Present-day English. A strong hypothesis (cf. Franz 1939: 537) attributes
this grammatical split between the active and the passive voice to the fact that
the dependent infinitive usually follows immediately upon the passive form
made. As a consequence, the probability of imminent stress clashes is high, and
the infinitive marker to is called for on rhythmical grounds. The situation is
different in the active, where make and the dependent infinitive are usually
separated by intervening object expressions (cf. example (20)):
(20) I would máke you smíle in the midst of your gravest airs, as I used to do.
(Samuel Richardson: Clarissa, 1748; ECF)

Despite the decisive influence of the Principle of Rhythmic Alternation, the

choice of marked or unmarked infinitives after made seems to be co-deter-
mined by additional factors. One of them is the processing difficulty introduced
by syntactic complexity, which may lead to the practically exceptionless
marking of infinitives (including non-initially stressed ones) as soon as adverbi-
al expressions intervene between made and the infinitive. This is the case in all
of the seven relevant corpus examples, e.g. sentence (21):
(21) Tracts, so numerous that it would be impossible to give their measure or
their value by any other calculation than that of their weight, were máde
by the ingenuity of the fair and pious contributors to assúme a very
tempting aspect,
(Frances Milton Trollope: The Vicar of Wrexhill, 1837; NCF)
 Morphology recycled 269

In this occurrence, the infinitive marker is not conditioned by rhythm (since

assume has an unstressed initial syllable), but is indispensable from the point of
view of language processing: it clarifies the syntactic relation between the super-
and subordinate verbs, which would otherwise easily get lost due to the bulky
adverbial insertion. This finding is predicted by the Complexity Principle (cf.
Rohdenburg 1996: 151), which stipulates that in cases of increased cognitive
complexity more explicit grammatical variants tend to be favoured (e.g.
to-infinitives are preferred over unmarked infinitives). Similar effects have been
described for active constructions in studies by Ohlander (1941), Mustanoja
(1960), Quirk & Svartvik (1970) and Fanego (1994).
In conclusion, corpus analyses have shown that the Principle of Rhythmic
Alternation accounts for most of the variation involving the variable infinitive
marker after passive made in the sixteenth to eighteenth centuries. In addition,
it explains why to-less constructions had the longest lease of life with non-
initially stressed infinitives at a time when this construction was already on its
way out. Thus, as long as the marked infinitive was not yet grammatically
obligatory, the particle to was redeployed for extra-grammatical purposes,
namely the avoidance of stress clashes.

2.4 The variable presence of the a-prefix in -ing participles

The Old English prepositions or prefixes ā, of, on, at and ge had all become
semantically bleached and phonetically reduced to Middle English a (cf. Fijn
van Draat 1912b: 508–514; Stroheker 1913: 25). Although as early as Middle
English the particle a was almost meaningless, it continued to accompany -ing
forms (‘a-prefixing’) and remained vigorous in Early and Late Modern English
(cf. Franz 1939: 559) and in modern English dialects (cf. Wolfram 1976; 1980
for Appalachian English).
According to Wolfram (1976; 1980), a-prefixing is subject to a number of
phonological restrictions (intersecting with syntactic and semantic constraints;
cf. also Nagucka 1984): firstly, the a is never prefixed to a vowel-initial base.
This tendency is accounted for by the phonetic principle of optimal syllable
structure, which disfavours the creation of hiatuses (cf. the ungrammatical
example in (22)). Secondly, a is only rarely prefixed to a verb following a vowel-
final element, thus equally avoiding hiatuses (cf. example (23), which has to be
considered rare). Thirdly, a is exclusively prefixed to initially stressed verbs (cf.
the ungrammatical example in (24)). This restriction serves to avoid the
sequence of two or more unstressed syllables (stress lapse).
270 Julia Schlüter

(22) *John was a-eatin’ his food. (Wolfram 1980: 125)

(23) He was just standin’ quietly a-hollerin’. (Wolfram 1976: 51)
(24) *He was a-retúrnin’ from his house. (Wolfram 1980: 126)

While all of these factors have been found to have explanatory potential in the
following corpus study, an additional phonological constraint will be investigat-
ed. This concerns the question of whether the -ing form follows a stressed or an
unstressed syllable. Since, according to Wolfram (1980: 126), only initially
stressed verbs are eligible for a-prefixing, a prefixless -ing form following a
stressed syllable would create a stress clash. The a-prefix can, however, serve as
a buffer separating the two stresses.
The hypothesis that will be tested in this study has been borrowed from Fijn
van Draat (1912b: 508f.): if the -ing form follows a stressed syllable, a-prefixing
should be more likely to occur than if the -ing form follows an unstressed
syllable. The database is composed of the same series of prose corpora that has
been used in the preceding studies, covering the sixteenth to twentieth centu-
ries. The analysis is restricted to a-prefixing in -ing forms following transitive
and intransitive uses of the verb set.16 Furthermore, only -ing forms that are
stressed on their initial syllable are counted, because non-initially stressed forms
never create a potential for stress clashes. In addition to the prototypical prefix
a-, other items are included that can fulfil the same function. This concerns the
prepositions on, upon, to, into and at.17 For convenience, all of these will be
referred to — somewhat imprecisely — as ‘a-prefixing’ (which is the common
designation in the literature).
The examples obtained from the corpus can be subdivided according to
their rhythmic contexts. Thus, there may be different kinds of objects preceding
the -ing form: personal pronouns are normally unstressed and do not produce
a stress clash even if the prefix is absent. By contrast, reflexive pronouns end in
a stressed syllable. In this case, the presence of the unstressed a-prefix (or a
comparable preposition) is advantageous. Compare examples (25) and (26):
(25) These Instructions (…) were sufficient to set them góing to the
Emperour. (Peter Bellon: The Court Secret, 1689; EEPF)
(26) To ease my mind a little, I set my sélf to wríting, and made these Verses
on my departure from Bracilla.
(Anonymous: The Player’s Tragedy, 1693; EEPF)

Among the full noun objects, one can distinguish between those ending in an
unstressed syllable (non-oxytonic nouns) and those ending in a stressed syllable
 Morphology recycled 271

(oxytonic nouns). In the former case (e.g. sentence (27)), the omission of the
prefix does not cause a stress clash; in the latter (e.g. sentence (28)), it does:
(27) (…), Ile set a cándle búrning in the midst of this roome where we all are,
open and easie to be seen as my hand: (Thomas Brewer: The Life and
Death of the Merry Deuill of Edmonton, 1631; EEPF)
(28) In this manner I may say, he set all whéels a góing at once, that might in
any kinde prejudice and disturb his own King Orontes;
(Sir Percy Herbert: The Princess Cloria, 1661; EEPF)

Furthermore, there are some instances in which the verb set, which generally
carries a strong stress, immediately precedes the dependent -ing form. This is
the case when the object of the transitive construction is pre- or postposed (cf.
example (29)), when the construction is passivized (cf. example (30)), or when
set is used intransitively (cf. example (31)).18 In all cases, the sequence set +
initially stressed verb in the -ing form creates a potential for stress clashes,
which may be avoided by using the intervening prefix as a buffer:19
(29) But I will be bold to say, that neither She, nor my Brother, nor even my
Father himself, knows what a heart they have sét a bléeding.
(Samuel Richardson: Clarissa, 1748; ECF)
(30) but such a tongue as his might lay matters too open, if once sét a-góing,
for you see he is not to be over-awed to any thing.
(Charles Johnstone: Chrysal, 1760; ECF)
(31) some of them let fly several shot at him from their Fuzees, while others
sét a rúnning after him, thinking to have overtaken him;
(Anonymous: Don Tomazo, 1680; EEPF)

Figure 4 summarizes the results of the corpus study in a highly simplified way:
it focuses exclusively on the contrast between the total percentage of all instanc-
es of -ing forms following unstressed syllables (including personal pronouns
and non-oxytonic nouns; represented by the dashed curve) compared to the
total percentage following stressed syllables (including reflexive pronouns,
oxytonic nouns and stressed forms of set itself; represented by the solid curve).
The data in Figure 4 show that in the sixteenth and seventeenth centuries
and in the first part of the nineteenth century, freedom of variation was
greatest: in these periods, the imbalance between the two contexts was consider-
able, with a divergence of 27 to 35%. By contrast, in the eighteenth century, the
presence of the prefix was almost obligatory (cf. Fijn van Draat 1912b: 510f.).
Even -ing forms following an unstressed syllable could hardly run counter to
272 Julia Schlüter

31/33 50/51 20/21

=94% =98% =95%
100%
percentage of prefixed -ing forms

80% 51/57
=89%

60% 18/27 25/42

=67% =60%
40%
23/86 17/73
=27% =23%
12/97
20% =12%
22/131
=17% 3/59
6/133=5% =5%
0%
EEPF
(1518-1700)

ECF
(1705-1780)

NCF
(1782-1839)

NCF
(1840-1869)

NCF
(1870-1903)

BNC, wridom
1 (1964-1993)
following an unstressed syllable
following a stressed syllable

Figure 4. The variable presence of the particle a/(up)on/(in)to/at preceding initially

stressed -ing forms dependent on the verb set in a series of prose corpora.

this trend, taking the prefix in 89% of all cases.20 Since the middle of the
nineteenth century, the a-prefix has become almost extinct, but traces of it can
still be found following stressed syllables, where the prefix retains its buffer
function. Despite these diachronic changes, in each chronologically defined
subsection of the corpus the proportion of a-prefixation is lower following
unstressed syllables than following stressed syllables. In other words, the
a-prefix is consistently more likely to occur in contexts where it can serve to
prevent a stress clash.
The corpus examples from the second half of the nineteenth century make
it clear that the a-prefix has acquired a social stigma: authors use it to character-
ize the speech of uneducated speakers as non-standard rather than to secure the
rhythmic pattern of the language (cf. example (32)). This stylistic use of
a-prefixing provides a further reason why rhythmic motivations are watered
 Morphology recycled 273

down in the nineteenth century, and it betrays the standardization pressures

affecting the a-prefix:
(32) (…) Gambado, who says, “be werry shy of a crupper if your ‘oss natural-
ly throws his saddle forward. It will certainlie make his tail sore, sét him a
kíckin’, and werry likely bring you into trouble.”
(Robert S. Surtees: Handley Cross, 1845; NCF)

Contrary to expectations, the curve for a-prefixing after stressed syllables in

Figure 4 rises once more in the late twentieth-century corpus: the avoidance of
stress clashes seems to be gaining importance as the percentage of what I have
provisionally called a-prefixing increases again. A close inspection of the corpus
examples shows that this effect is, however, not due to a resurgence of the
a-prefix, but to an increased use of the preposition to, as exemplified in (33):
(33) But oh, he had set her héart to rácing, the instant she’d first seen him.
(Sandra Marton: Roman Spring, 1993; BNC)

In fact, in stark contrast to the previous centuries (where the a-prefix constantly
accounted for about 80% of the elements preposed to the -ing forms), 18 out of
the total of 20 instances that have been referred to as a-prefixed contain the
preposition to, whereas only two make use of the a-prefix itself (one of them
clearly characterizing a speaker of vernacular English). Presumably, the fully-
fledged preposition has less of a non-standard or obsolescent flavour, so that
even present-day authors employ it quite freely whenever they require some
intervening buffer syllable.
In connection with the avoidance of stress clashes by means of the a-prefix,
there is some indication of the effects of another factor co-determining the
variable presence of the a-prefix: the so-called horror aequi Principle (cf.
Rohdenburg & Schlüter 2000: 461)21 describes the avoidance of the semantically
unmotivated adjacency of phonologically similar material. Compare examples
(34) and (35) on the one hand to examples (36) and (37) on the other:
(34) But though this strange Message sett all my wounds frésh a bléeding, yet I
had so much discretion left, as only to answere it with a Complement
(…) (Roger Boyle: Parthenissa, Part 3, 1655; EEPF)
(35) he entertain’d me with such passionate discourses of his Flame, that I
must acknowledge, they sett my old wounds frésh a bléeding,
(Roger Boyle: Parthenissa, Part 3, 1655; EEPF)
274 Julia Schlüter

(36) You shall be satisfy’d with a true Narration of the Disasters of a

miserable Wretch, injur’d by Fortune, and pursu’d by Fate; the Relation
of which, will set my Wóunds blèeding afrésh;
(Anonymous: Cynthia, 1687; EEPF)
(37) yet the motion set both his wóunds blèeding afrésh; and it was with diffi-
culty they again stopped the blood.
(Samuel Richardson: Clarissa, 1748; ECF)

In (36) and (37), the a-prefix preceding the -ing form is consistently avoided,
presumably on account of the horror aequi effect provoked by the adverb afresh,
which itself contains an a-prefix deriving from the same source as the verbal
a-prefix (cf. Stroheker 1913: 25). Additionally, the -ing form bleeding is in both
cases rhythmically overshadowed by the strongly stressed and focused adverb
afresh. As a consequence, the stress clash with the preceding noun wounds is
mitigated and thus rendered more acceptable. While in (34) and (35) rhythmic
alternation requires the presence of the verbal a-prefix and there is no horror
aequi effect to counteract this tendency, in (36) and (37) both sentence prosody
and the avoidance of identity effects facilitate the omission of the prefix in the
-ing form.
To sum up, the study has revealed that a-prefixing in English has taken a
changeable route of evolution, first gaining ground on and then losing ground
to the prefixless variant. In the midst of this indeterminacy, the Principle of
Rhythmic Alternation has at all times been a powerful factor influencing the
choice between prefixed and prefixless -ing forms following the verb set. In
interaction with other factors (e.g. standardization tendencies and horror aequi)
it explains part of the synchronic distribution and of the diachronic preserva-
tion of the a-prefix and functionally similar prepositions in sixteenth- to
twentieth-century English.

3. Conclusion

In this paper I have presented the results of empirical studies of four different
grammatical variation phenomena, all of which show that the presence or
absence of grammatical morphemes and markers may be determined — among
other factors — by the Principle of Rhythmic Alternation. This concerns (a) the
variable presence of the suffix -en in the participle drunk, (b) the variable
presence of the adverbial suffix -ly in scarce used as an adverb, (c) the variable
 Morphology recycled 275

marking of infinitives depending on the verb make in the passive and (d) the
variable prefixation of -ing forms with the prefix a- (or the insertion of certain
prepositions). A number of comparable findings have been obtained for the
variable use of the -er suffix, which could be optionally appended to the
irregular comparative worse in the sixteenth and seventeenth centuries (cf.
Schlüter 2001), for the -ed suffix in the participle lighted, which was partly
replaced by the monosyllabic variant lit (cf. Schlüter to appear), for variable
infinitive marking dependent on the verb bid in the passive (cf. Rohdenburg &
Schlüter 2000: 484ff.), and for the fate of the final vowels deriving from Old
English inflectional endings (cf. Minkova 1990; 1991: 155ff.).
In all of these cases we witness situations in which grammatical morphemes
or markers are lingering, without however being assigned a definite grammati-
cal function: they may be involved in processes of grammaticalization (like the
adverbial suffix or the infinitive marker), or degrammaticalization (like the
participial suffix -en), or simply in longstanding cases of grammatical change-
ability (like the a-prefix). Each of these situations is inherently linked up with
a more or less extensive period of variability during which the morphemes and
markers under consideration are neither obligatory nor unavailable. As has
been demonstrated, it is in precisely these phases of indeterminacy that the
Principle of Rhythmic Alternation gets a chance to overrule grammatical
motivations.
The effects of this principle have a synchronic as well as a diachronic
dimension. Irrespective of other factors, it has been shown that rhythmic
alternation sometimes constitutes the primary determinant of the synchronic
distribution of these variants; in other cases, it operates within a more narrow
margin defined by competing influences. Diachronically, the principle helps to
establish new morphological variants wherever they can improve the rhythm;
on the other hand, it contributes to the retention of obsolescent variants
wherever they can preserve rhythmically optimal configurations. In this way,
the Principle of Rhythmic Alternation can lead to the elimination of the
variability if one of the variants proves disadvantageous in the majority of its
occurrences (e.g. the to-less infinitive after make in the passive, or the monosyl-
labic form drunk before nouns). However, this regularization entails the loss of
flexibility in individual cases (e.g. even non-initially stressed infinitives depen-
dent on made are regularly marked nowadays; even non-initially stressed nouns
were almost systematically preceded by drunken as an attribute from the
sixteenth to the nineteenth centuries). As a result, rhythmic alternation is
observed in the greater part of the occurrences, but not in every single instance.
<DEST "sch-n*">

276 Julia Schlüter

The corpus studies have shown that the Principle of Rhythmic Alternation
interacts with other determinants of grammatical variation such as standardiza-
tion pressures, the horror aequi Principle, the Complexity Principle, semantics
and a number of other phonological factors such as optimal syllable structure
and considerations of sentence prosody. This list is by no means exhaustive. As
has been illustrated in some of the above-mentioned examples, these influences
operate synergetically or antagonistically, depending on the alternatives in
question. The considerable efficacy of the Principle of Rhythmic Alternation
revealed in this contribution gives only a foretaste of the wealth of insights that
can still be derived from the study of linguistic variation and change.

Notes

* This work is part of the research project “Determinants of Grammatical Variation in

English”, based at the University of Paderborn (Germany). I would like to thank the
Deutsche Forschungsgemeinschaft (grant No. RO 2271/1–1) for financial support and Teresa
Fanego as well as an anonymous reviewer for many useful comments on previous versions
of this paper.
1. Depending on the linguistic background of the researcher, the notion of ‘stress lapse’ is
variously defined as a sequence of two (cf. Plag 1999:156) or three (cf. Selkirk 1984:49; Kager
1995: 382) unstressed syllables. It is agreed that the presence of only one unstressed syllable
separating two stressed ones represents an ideal constellation. There is evidence in the above-
mentioned works that the avoidance of sequences of more than one unstressed syllable is a
matter of degree: two unstressed syllables are universally avoided less strongly than three.
2. The following orthographic variants were retrieved from the corpora: drun(c)k(e),
dron(c)k(e), dran(c)k(e), drūk(e) and drōk(e) as suffixless forms (for ease of reference
subsumed henceforth under the spelling drunk); drun(c)ken, dron(c)ken, drūken, drōken,
drunkyn, drunkun and drunkē as suffix-containing forms (subsumed under the spelling
drunken). In the corpus search, a wildcard was prefixed to these forms to include all
premodified and orthographically fused forms. All past tense forms were excluded from the
data in Figure 1, since, with the exception of three past-tense uses of drunken, which were
inflected for plural in the Middle English part of the corpus, they regularly involved the
suffixless form.
3. In those cases in which the ECF corpus included more than one edition of a publication,
only the earliest edition was taken into consideration.
4. In this example and elsewhere, acute accents indicate primary stress. Where appropriate,
grave accents will be used to mark reduced primary or secondary stress. The italics are my
addition.
 Morphology recycled 277

5. Contrary to the information provided, for example, by the OED ([Link]. drunk, drunken),
the variant drunken was also, although rarely, used as a nominalized adjective. This is the case
in four corpus examples; cf.:
(i) euery one at a banquet is compelled to drinke carouse, to the end the sober may not
disclose ye words or deeds of the drunken. (Henry Wotton: A Courtlie controuersie of
Cupids Cautels, 1578; EEPF)
6. Cf. Kelly & Bock (1988: 391): 89% of the 4,218 most frequent disyllabic English nouns are
initially stressed. Monosyllabic nouns (which are at least as frequent) have not been included
in this count, but note that they are inherently stressed syllables.
7. Franz (1939: 167) additionally remarks that the preservation of -en in attributive past
participles may initially have been due to the fact that in Old English, after the definite article
and demonstrative pronouns, attributive past participles took the endings of the weak
adjectival declension (e.g. se foresprecena here ‘the aforementioned army’).
8. A pause occurs, roughly speaking, where the participle is followed by a relatively wide
syntactic and prosodic disjuncture (for a more technical analysis, see Selkirk 1984: 301–320).
9. It is striking that drunk in simple attributive uses begins to establish itself in high-
frequency collocations such as drunk driver(s), drunk driving (which tend to be treated as
compounds by the stress-assigning rules) and drunk man/men, drunk husband(s), drunk woman/
women (which involve generic nouns). The threatening succession of two stressed syllables
is perhaps averted by allocating only reduced stress to these highly given combinations.
10. There is little evidence from the etymology as well as from the spelling variants listed by
the OED (s.v. scarce) that — as has been suggested by an anonymous reviewer — scarce could
be given a disyllabic pronunciation in earlier centuries. Even if the r could possibly be
syllabic, the corpus findings prove that there is a clear distributional difference between scarce
and scarcely, which can only be explained by their dissimilar rhythmic structure.
11. Concerning the reduction of stress due to the length and complexity of following
material, cf. Bolinger (1965: 149f.); Zwicky (1969: 429).
12. There is also one author who has only one adverbial use and happens to employ scarce in
this context, but this case has to be considered exceptional.
13. In this respect, he resembles Sir Walter Scott, who is equally Scottish-born. These results
— albeit sporadic — seem to suggest that the rate of adverb marking in Scottish English lags
somewhat behind that in the more southern varieties.
14. Effects of the following factors on the marking of verb-dependent infinitives in Middle
and Early Modern English have been demonstrated in the literature: (1) various iconic
tendencies (degree of transitivity or impingement of the matrix clause subject on the object,
(in-)directness of the relation between the events denoted by the matrix verb and infinitive,
(non-)actualization of the event denoted by the infinitive; cf. Fischer 1995; 1996; 1997); (2)
certain syntactic characteristics that may be subsumed under the concept of processing
complexity (active/passive contrasts, degree of separation of matrix verb and dependent
infinitive due to intervening material, non-canonical sentence structures, change of object in
coordinated infinitival structures; cf. Ohlander 1941; Mustanoja 1960: 522; Quirk & Svartvik
1970; Fanego 1994; Fischer 1995: 15; Rohdenburg & Schlüter 2000: 446–452). In contrast,
278 Julia Schlüter

rhythmic influences are frequently either treated as an aside (e.g. Ohlander 1941: 66; Fanego
1994: 201f.) or denied as significant constraining forces (e.g. Fischer 1995: 2; 1996: 267).
15. To find passive uses of the past participle made, a corpus search was carried out that
retrieved all forms and spelling variants of the auxiliary be (am, art, is, are, was, wast, wert,
were, be, bee, being, beeing, been, beene, bē) immediately preceding made.
16. The following verb forms were searched: set, sete, sett, sette, setting, sets, settes, seting, setst,
settest, setts, setteth, settyth, setting, settynge, setyng and setted, each one followed by an -ing
form at a maximal distance of eight words. Besides the ending -ing, the spelling variants -yng,
-inge and -ynge were included.
17. Cf. Visser (1973: 1894). These prepositions are subsumed under the same category
because, in terms of rhythm, they all function alike; even the disyllabic items upon and into
can take over a buffer function if they are completely destressed; this constellation creates a
ternary rhythm, which is not particularly objectionable. There is however some evidence that
the prepositions are treated as more independent items than the a-prefix: they may precede
-ing forms that do not satisfy Wolfram’s above-mentioned conditions. The corpus yields
forms like on retreating (EEPF), upon enquiring (ECF), to examining, to replacing, to
recovering, on debating, upon devising and upon inquiring (all NCF), which are initially
unstressed and in some cases begin with a vowel. Since these -ing forms constitute no
potential loci for stress clashes, they are excluded from the data in Figure 4.
18. The disyllabic form setting does not occur even once in these three constellations. Note
that the second unstressed syllable provided by the -ing suffix would secure an alternating
rhythm. Nevertheless, the sequence setting (a-)V-ing is eschewed due to the strong avoidance
effect provoked by repeated -ing forms (see in particular Bolinger 1979). Cf. also the remarks
below in relation to the horror aequi tendency.
19. In the case of object extraposition or passivization, the a-prefix may additionally be
motivated along the lines of the Complexity Principle (Rohdenburg 1996; cf. the relevant
remarks in Section 2.3) as an explicit marker of the syntactic relation between set and the
-ing form.
20. According to Fijn van Draat, the great rhythmic value of the prefix was no longer clearly
recognized in the eighteenth century and generally, “[s]ense of Rhythm was but weakly
developed” in this period. By contrast, for the nineteenth century, the author claims a “re-
awakened sense of Rhythm” (Fijn van Draat 1912b: 511, 512).
21. The term horror aequi was originally introduced by Brugmann (1909:146ff.) to designate
the perception of a cacophony created by the close adjacency of phonologically similar
elements.
 Morphology recycled 279

References

Corpora
British National Corpus. Corpus BNC 1, version 1.00.
Early English Prose Fiction. 1997. Electronic Book Technologies Inc./Chadwyck-Healey.
Cambridge. In association with the Salzburg Centre for Research on the English Novel
SCREEN.
Eighteenth-Century Fiction. 1996. Electronic Book Technologies Inc./Chadwyck-Healey.
Cambridge.
Nineteenth-Century Fiction. 1999–2000. Electronic Book Technologies Inc./Chadwyck-
Healey. Cambridge.
The Helsinki Corpus of English Texts. 1991. Compiled by Matti Rissanen, Ossi Ihalainen &
Merja Kytö. Helsinki: Department of English, University of Helsinki.
Secondary Sources
Adams, Valerie. 1973. An Introduction to Modern English Word Formation. London:
Longman.
Aitchison, Jean. 1991 [2nd ed.]. Language Change: Progress or Decay? Cambridge: Cambridge
University Press.
Bihl, Josef. 1916. Die Wirkungen des Rhythmus in der Sprache von Chaucer und Gower.
Heidelberg: Carl Winter.
Bolinger, Dwight L. 1965. “Pitch Accent and Sentence Rhythm”. Forms of English: Accent,
Morpheme, Order ed. by Dwight L. Bolinger, 139–180. Cambridge, Mass.: Harvard
University Press.
Bolinger, Dwight L. 1979. “The Jingle Theory of Double -ing”. Function and Context in
Linguistic Analysis: A Festschrift for William Haas ed. by David J. Allerton, Edward
Carney & David Holdcroft, 41–56. Cambridge: Cambridge University Press.
Brinton, Laurel J. & Dieter Stein. 1995. “Functional Renewal.” Historical Linguistics 1993.
Selected Papers from the 11th International Conference on Historical Linguistics, Los
Angeles, 16–20 August 1993 ed. by Henning Andersen, 33–47. Amsterdam & Philadel-
phia: John Benjamins.
Brugmann, Karl. 1909. Das Wesen der lautlichen Dissimilationen. (= Abhandlungen der
philologisch-historischen Klasse der königlich sächsischen Gesellschaft der Wissenschaften,
27/V.) Leipzig: B. G. Teubner.
Fanego, Teresa. 1994. “Infinitive Marking in Early Modern English.” English Historical
Linguistics 1992. Papers from the 7th International Conference on English Historical
Linguistics, Valencia, 22–26 September 1992 ed. by Francisco Fernández, Miguel Fuster
& Juan José Calvo, 191–203. Amsterdam & Philadelphia: John Benjamins.
Fijn van Draat, Pieter. 1910. Rhythm in English Prose. Heidelberg: Universitätsverlag Carl
Winter. Repr. Amsterdam: Swets & Zeitlinger, 1967.
Fijn van Draat, Pieter. 1912a. “Rhythm in English Prose: The Adjective”. Anglia 36.1–56.
Fijn van Draat, Pieter. 1912b. “Rhythm in English Prose: The Pronoun./The Preposition”.
Anglia 36.493–538.
Fischer, Olga. 1995. “The Distinction between to and Bare Infinitival Complements in Late
Middle English.” Diachronica 12.1–30.
280 Julia Schlüter

Fischer, Olga. 1996. “Verbal Complementation in Early ME: How Do the Infinitives Fit in?”
English Historical Linguistics 1994 ed. by Derek Britton, 247–270. Amsterdam &
Philadelphia: John Benjamins.
Fischer, Olga. 1997. “Infinitive Marking in Late Middle English: Transitivity and Changes in
the English System of Case.” Studies in Middle English Linguistics ed. by Jacek Fisiak &
Werner Winter, 109–134. Berlin & New York: Mouton de Gruyter.
Franz, Wilhelm. 1939. Die Sprache Shakespeares in Vers und Prosa, unter Berücksichtigung des
Amerikanischen entwicklungsgeschichtlich dargestellt. Fourth, revised and augmented
edition. Halle/Saale: Niemeyer. Repr. Tübingen: Max Niemeyer, 1986.
Kager, René. 1995. “The Metrical Theory of Word Stress”. The Handbook of Phonological
Theory ed. by John A. Goldsmith, 367–402. Oxford: Blackwell.
Kelly, Michael H. & J. Kathryn Bock. 1988. “Stress in Time”. Journal of Experimental
Psychology: Human Perception and Performance 14.389–403.
Lass, Roger. 1990. “How to Do Things with Junk: Exaptation in Language Evolution”.
Journal of Linguistics 26.79–102.
Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge
University Press.
Minkova, Donka. 1990. “Adjectival Inflexion Relics and Speech Rhythm in Late Middle and
Early Modern English.” Papers from the 5th International Conference on English Histori-
cal Linguistics, Cambridge, 6–9 April 1987 ed. by Sylvia Adamson, Vivien Law, Nigel
Vincent & Susan Wright, 313–337. Amsterdam & Philadelphia: John Benjamins.
Minkova, Donka. 1991. The History of Final Vowels in English: The Sound of Muting. Berlin
& New York: Mouton de Gruyter.
Mustanoja, Tauno F. 1960. A Middle English Syntax. Part I: Parts of Speech. Helsinki: Société
Néophilologique.
Nagucka, Ruta. 1984. “Explorations into Syntactic Obsoleteness: English a-X-ing and X-ing”.
Historical Syntax ed. by Jacek Fisiak, 363–381. Berlin, New York & Amsterdam: Mouton.
Nevalainen, Terttu. 1997. “The Processes of Adverb Derivation in Late Middle and Early
Modern English”. Grammaticalization at Work: Studies of Long-term Developments in
English ed. by Matti Rissanen, Merja Kytö & Kirsi Heikkonen, 146–189. Berlin & New
York: Mouton de Gruyter.
OED 2 on CD-ROM. 1995. Ed. by John A. Simpson & Edmund S. C. Weiner. Version 1.13.
Oxford: Oxford University Press/Rotterdam: AND Software B. V.
Ohlander, Urban. 1941. “A Study on the Use of the Infinitive Sign in Middle English.” Studia
Neophilologica 14.58–66.
Plag, Ingo. 1999. Morphological Productivity: Structural Constraints in English Derivation.
Berlin & New York: Mouton de Gruyter.
Quirk, Randolph & Jan Svartvik. 1970. “Types and Uses of Non-finite Clause Links in
Chaucer.” English Studies 51.393–411.
Rohdenburg, Günter. 1996. “Cognitive Complexity and Increased Grammatical Explicitness
in English”. Cognitive Linguistics 7.149–182.
Rohdenburg, Günter & Julia Schlüter. 2000. “Determinanten grammatischer Variation im
Früh- und Spätneuenglischen”. Sprachwissenschaft 25.444–496.
Schlüter, Julia. 2001. “Why Worser is Better: The Double Comparative in 16th to 17th
Century English”. Language Variation and Change 13.193–208.
</TARGET "sch">

Morphology recycled 281

Schlüter, Julia. To appear. “Phonological Determinants of Grammatical Variation in English:

Chomsky’s Worst Possible Case.” Determinants of Grammatical Variation in English ed.
by Günter Rohdenburg & Britta Mondorf. Berlin & New York: Mouton de Gruyter.
Selkirk, Elizabeth O. 1984. Phonology and Syntax: The Relation between Sound and Structure.
Cambridge, Mass. & London: MIT Press.
Stroheker, Friedrich. 1913. Doppelformen und Rhythmus bei Marlowe und Kyd. Heidelberg:
Carl Winter.
Vincent, Nigel. 1995. “Exaptation and Grammaticalization”. Historical Linguistics 1993.
Selected Papers from the 11th International Conference on Historical Linguistics, Los
Angeles, 16–20 August 1993 ed. by Henning Andersen, 433–445. Amsterdam & Philadel-
phia: John Benjamins.
Visser, Frederick Th. 1973. An Historical Syntax of the English Language. Part III, Second
Half: Syntactical Units with Two and with More Verbs. Leiden: E. J. Brill.
Wolfram, Walt. 1976. “Toward a Description of a-prefixing in Appalachian English”.
American Speech 51.45–56.
Wolfram, Walt. 1980. “A-prefixing in Appalachian English”. Locating Language in Time and
Space ed. by William Labov, 107–142. New York, London, Toronto, Sydney & San
Francisco: Academic Press.
Zwicky, Arnold M. 1969. “Phonological Constraints in Syntactic Descriptions”. Papers in
Linguistics 1.411–463.
<TARGET "ni" DOCINFO

AUTHOR ""

TITLE "Name index"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Name index

A
Adams, Valerie 262, 279 Bihl, Josef 256, 279
Aitchison, Jean 150, 153, 261, 279 Bock, J. Kathryn 277, 280
Aitken, Adam Jack 208, 227 Bolinger, Dwight L. 131, 153, 256–258,
Akimoto, Minoji v, vii, 2, 7, 9, 16, 17, 19 277–279
Akinnaso, F. Niyi 43, 64 Booij, Geert 126
Allen, Andrew S. 75, 76, 94 Bosworth, Joseph 108
Allen, Cynthia L. 2, 3, 23, 24, 38, 92, 94, Brinton, Laurel J. v, vii, 1, 2, 19, 67, 83,
137, 151, 153, 248, 252 92, 95, 185, 200, 261, 279
Altenberg, Bengt 23, 33, 38 Brown, Gillian 43, 64
Amos, Ashley Crandell 126 Brugmann, Karl 278, 279
Anttila, Raimo 71, 72, 76, 95 Brunner, Karl 124, 128
Aston, Guy 151, 153 Burchfield, Robert W. 134, 138, 153, 158,
Atkinson, Dwight 50, 64, 153 167, 179
Auwera, Johan van der 74–76, 91, 95, Burnard, Lou 151, 153
151 Bussmann, Hadumod 71, 95
Ayto, John 121, 122, 126 Bybee, Joan L. 152–154

B C
Baayen, Harald 161, 179 Cameron, Angus 126, 198, 201
Bailey, Richard W. 174, 179 Campbell, Alistair 124, 126
Ball, Christopher 198, 201 Campbell, Lyle 150, 154
Bammesberger, Alfred 124, 126 Cattell, Ray 19, 20
Barber, Charles 136, 153 Cawley, A. C. 38
Baron, Dennis 113, 126 Chafe, Wallace L. 17, 20, 43, 65
Bartsch, Renate 164, 179 Chen, Guohua 74, 75, 95
Bauer, Laurie 71–73, 86, 95, 123, 126, Claridge, Claudia 11, 20, 123
161, 167, 179 Clark, Victoria v, vii, 2, 3, 43, 174, 179
Bellert, Irena 163, 164, 177, 179 Clark Hall, John R. 108
Bergen, Linda D. van 182, 187, 199 Claudi, Ulrike 74–76, 90, 95, 154
Bertschinger, Max 132, 136, 153 Conrad, Susan 49, 50, 64, 65, 126, 153,
Biber, Douglas v, vii, 1–3, 43–45, 47–50, 179
64, 65, 124, 126, 131, 133, 151, 153, Cowie, A. P. 15, 19, 20
174, 177–179, 224, 227 Craigie, Sir William A. 228
Croft, William 131, 154
284 Name index

D Franz, Wilhelm 256–258, 261, 265, 266,

Dalton-Puffer, Christiane 121, 126, 160, 268, 269, 277, 280
161, 167, 177–179
Davies, Robertson 193, 200 G
Denison, David 10, 20, 86, 95, 137, 154, Gaaf, Willem van der 23–27, 30–33, 38,
189, 200, 205, 217, 225, 227, 244, 137, 154
246, 248, 249, 251, 252 Gamillscheg, Ernst 249, 252
Dennis, Leah 222, 227 Giacalone Ramat, Anna 68, 69, 90, 95
Deutschbein, Max 236, 239, 249–252 Givón, Talmy 92, 232, 236, 241, 243,
DeVito, Joseph A. 43, 65 244, 246, 248, 249, 251, 252
Devitt, Amy J. 204, 216, 225, 227 Goldberg, Adele E. 240, 252
Dieth, Eugen 37, 38 González Fernández-Corugedo, Santiago
Diewald, Gabriele 97 1
Dik, Simon 5, 7, 231, 232, 239, 240, 242, Goodluck, Helen 154
248–252 Görlach, Manfred 19, 20, 174, 179,
Don, Jan 123, 127 233–235, 246, 248, 249, 252
Dons, Ute 205, 225, 226, 228 Gradon, Pamela 32, 38
Downing, Angela 184, 200 Greenbaum, Sidney 20, 97, 128, 155, 180,
Dressler, Wolfgang U. 124, 127 229
Grice, Herbert Paul 140, 154
E Griffin, W. J. 65
Einenkel, Eugen 23, 38 Gumperz, John J. 43, 65
Elsness, Johan 207, 228
Enkvist, Nils E. 185, 200 H
Erickson, Jon 248 Haan, Pieter de 44, 65
Ernestus, Mirjam 199, 200 Hagège, Claude 68, 71, 87, 91, 95
Haiman, John 125, 127
F Halliday, M. A. K. 3, 7, 43, 65, 182, 184,
Faber, Pamela 5, 7, 231, 252 200, 232, 249, 250, 253
Fanego, Teresa v, vii, 1, 9, 93–95, 151, Hannay, Mike 191, 200
189, 200, 225, 269, 276–279 Harris, Alice 150, 154
Ferris, Connor 92, 95 Hasan, Ruqaiya 182, 200
Fijn van Draat, Pieter 256–258, 261, Haspelmath, Martin 73–75, 95, 243, 253
269–271, 278, 279 Hatcher, Anna Granville 23, 37, 39
Fillmore, Charles J. 131, 154, 249, 251, Healey, Antonette diPaolo 126, 198, 200,
252 253
Finegan, Edward 49, 50, 64, 126, 153, Heidermanns, Frank 233, 248, 249, 253
174, 179 Heine, Bernd 74–76, 90, 152–154
Fischer, Olga 75, 95–97, 137, 154, 186, Henry, Alison 249, 253
189, 198, 200, 246, 251, 252, Hickey, Raymond 1
277–280 Hofstetter, Walter 177
Follett, Wilson 158, 179 Hopper, Paul J. 2, 7, 71–73, 75, 76, 88,
Foster, Brian 168, 169, 179 89, 95–97, 152, 154, 232, 240, 243,
Foster, Robert 185, 200 248, 253
Fowler, H. W. 92, 93, 95 Horstmann, Carl 23, 39
 Name index 285

Houghton, Donald E. 157, 158, 161, 179 Kölbing, Eugen 37, 39

Howard, Godfrey 158, 179 Koopman, Willem 200
Huddleston, Rodney 69, 86, 96, 143, 154 Kornexl, Lucia v, vii, 2, 4, 111, 112,
Hundt, Marianne 133, 151, 154, 246, 125–127
248, 252, 253 Kortmann, Bernd 151
Hünnemeyer, Friederike 74–76, 90, 95, Krahe, Hans 114, 116, 124, 127, 233, 234,
154 253
Kroch, Anthony 37
I Kroll, Barbara 43, 65
Ihalainen, Ossi 253, 279 Krug, Manfred v, viii, 2, 131, 132, 136,
Ikegami, Yoshihiko 248, 253 137, 144, 152, 153, 155
Inghult, Göran 170, 171, 179 Kruisinga, Etsko 77, 96
Iwata, Seizi 244–246, 248, 251, 253 Kuhn, Sherman M. 39, 128, 254
Kurath, Hans 39, 128, 254
J Kuryłowicz, Jerzy 73–75, 96
Jacob-Flynn, Nicholas 177 Kytö, Merja 92, 96, 151, 155, 253, 279
Janda, Richard D. 75, 76, 92, 96
Jespersen, Otto 10, 20, 23, 27, 28, 30, 31, L
39, 77, 78, 79, 81, 92, 94, 96, 137, Labov, William 144, 155
150, 154, 236, 239, 249, 253 Langacker, Ronald W. 100, 108, 231, 239,
Johansson, Stig 64, 126, 151, 153, 154, 250, 253
179 Lass, Roger 75, 87, 90, 94, 96, 114, 128,
Joseph, Brian D. 75, 76, 92, 96, 155 261, 280
Leech, Geoffrey 20, 64, 97, 126, 128,
K 153–155, 179, 180, 229
Kager, René 276, 280 Leek, Frederike van der 137, 154
Kahlas-Tarkka, Leena 152, 154, 182, 201 Lehmann, Christian 69, 71–76, 88, 92,
Kaltman, Hannah 43, 65 96, 126
Kärre, Karl 124, 127 Lemmens, Marcel 191, 201
Kastovsky, Dieter v, vii, 4, 5, 99, 101, Lenker, Ursula vi, viii, 2, 4, 123, 157, 163,
104, 108, 111–117, 119, 124–127, 171, 178, 179
234, 250, 253 Lessau, Donald A. 70, 71, 76, 96
Kay, P. 43, 65 Lewis, Robert E. 39, 128, 254
Keizer, Evelien 191, 200 Lightfoot, David W. 243, 253
Kellner, Leon 27, 30, 39 Lipka, Leonhard 178, 180
Kelly, Michael H. 277, 280 Locke, Philip 184, 200
Kemenade, Ans van 1, 7, 94, 187–189, López-Couso, María José viii, 137, 155
200, 201 Los, Bettelou vi, viii, 2, 3, 181, 196, 198,
Kemmer, Suzanne 251, 253 201
Kirchner, G. 144, 154
Klages, Monika vi, vii, 2, 5, 231, 250, 253 M
Kluge, Friedrich 114, 116, 124, 127, 168, Mackin, R. M. 15, 19, 20
179 Mahler, Andreas 177
Knutson, A. 124, 127 Mair, Christian 133, 154, 189, 190, 201
Koefoed, Geert 125, 127 Mairal Usón, Ricardo 5, 7, 231, 252
286 Name index

Marchand, Hans
86, 96, 105, 106, 108, Ohlander, Urban
266, 269, 277, 278, 280
109, 111, 116–123, 126, 128, 161, Olson, David R.
43, 65
165, 167, 171, 180
Marle, Jaap van
125, 127 P
Matsumoto, Meiko
11, 20 Páez Urdaneta, Iraset
91, 97
Mayerthaler, Willi
127 Pagliuca, William
154
McCully, C. B.
1 Palmer, Frank R.
131, 132, 144, 152, 155,
McMahon, April M. S.
64, 65 241, 246, 251, 254
Meid, Wolfgang
114, 116, 124, 127, 233, Panagl, Oswald
127
234, 253 Paraschkewoff, Boris
163, 165, 168, 178,
Méndez-Naya, Belén
1, 137, 155 180
Meurman-Solin, Anneli
vi, viii, 5, 6, 203, Parr, Tony
191, 201
204, 208, 212, 213, 223, 225, 226, Paul, Hermann
124, 128
228 Pennanen, Esko V.
123, 128
Minkova, Donka
275, 280 Pérez-Guerra, Javier
viii
Mitchell, Bruce
38, 39, 184, 198, 201, 231 Perkins, Revere D.
154
Mittendorf, Ingo
6, 7, 203, 214, 216, 228 Pilch, Herbert
125, 128
Moessner, Lilo
205, 225, 226, 228 Plag, Ingo
160, 161, 167, 177–179, 276,
Möhlig, Ruth
vi, viii, 2, 5, 231 280
Moralejo-Gárate, Teresa
19 Plank, Frans
1
Moreno Cabrera, Juan C.
67, 72, 96 Poppe, Erich
6, 7, 203, 214, 216, 228
Mossé, Fernand
246 Poutsma, Hendrik
10, 19, 20, 77, 94, 97
Mugdan, Joachim
126 Prins, Anton Adriaan
16, 20
Mustanoja, Tauno F.
23, 29, 30, 39, 133, Pulgram, Ernst
161, 180
137, 155, 214, 223, 228, 266, 269,
277, 280 Q
Quine, Willard Van Orman
147, 155
N Quirk, Randolph
18, 20, 69–71, 77, 86,
Nagucka, Ruta
269, 280 87, 97, 121, 128, 131, 132, 143, 155,
Nate, Richard
178, 180 158, 162–164, 167, 175, 177, 178,
Nevalainen, Terttu
94, 96, 108, 109, 261, 180, 205, 225, 229, 269, 277, 280
280
Newmeyer, Frederick J.
72, 75, 89, 91, R
94, 96 Radford, Andrew
232, 254
Noël, Dirk
190, 201 Rahn, Walter
161, 175, 180
Noonan, Michael
196, 201 Ramat, Paolo
68–72, 74, 90, 91, 97, 157,
Norris, R. C.
65 169, 180
Nunberg, Geoffrey
9, 11, 12, 18–20 Raumolin-Brunberg, Helena
46–48, 59,
Nurmi, Arja
220, 228 65, 182, 201
Reid, Thomas B. W.
254
O Renouf, Antoinette
161, 179
O’Connor, Mary Catherine
43, 65 Reppen, Randi
50, 64
O’Donnell, Roy C.
43, 65 Ricca, Davide
157, 169, 180
Ochs, Elinor
43, 65 Rissanen, Matti
37, 181–183, 186, 198,
Ogura, Michiko
238, 250, 254 201, 202, 235, 246, 253, 254, 279
 Name index 287

Rohdenburg, Günter 266, 269, 273, 275, T

277, 278, 280 Taavitsainen, Irma 1
Rothwell, William 249, 254 Tabor, Whitney 69, 97
Roy, Arundhati 176, 180 Tanabe, Harumi 11, 20
Rydén, Mats 227, 229 Thompson, Sandra A. 232, 240, 243, 248,
253
S Tieken-Boon van Ostade, Ingrid 1
Saeed, John I. 250, 254 Toller, T. Northcote 108
Sag, Ivan A. 20 Traugott, Elizabeth Closs 2, 7, 17, 20,
Sand, Andrea 154 68–73, 75, 87, 89, 91, 94–97, 131,
Sanders, Gerald 123, 126, 128 147, 151–155, 213, 254
Sauer, Hans 117, 119, 125, 128 Trench, Richard Chevenix 173, 180
Schaefer, Christiane 233, 249, 254 Trommelen, Mieke 123, 127
Scheffer, Johannes 227, 229
Schlüter, Julia vi, viii, 4, 255, 266, 273, V
275, 277, 280, 281 Valera, Salvador 123, 129, 178, 180
Schmid, Hans-Jörg 121, 126, 128 Varantola, Krista 44, 65
Schultze, Dirk 123 Venezky, Richard L. 198, 200, 253
Schwenter, Scott A. 69, 97 Vennemann, Theo 6, 7
Searle, John 143, 155 Vincent, Nigel 261, 281
Seebold, Elmar 168, 179 Visser, Frederikus Th. 11, 15, 17, 20, 131,
Selkirk, Elizabeth O. 4, 7, 255, 259, 276, 133, 143, 144, 155, 227, 229, 234,
277, 281 239, 244, 246, 249, 252, 254, 278,
Seoane-Posse, Elena 1, 19, 20, 183, 192, 281
199, 202
Siemund, Rainer 154 W
Sievers, Eduard 124, 128 Warner, Anthony R. 1, 189, 202, 227,
Siewierska, Anna 183, 192, 202, 236, 241, 229
249, 254 Wartburg, Walther von 254
Simpson, John A. 20, 39, 96, 128, 155, Wasow, Thomas 20
180, 201, 254, 280 Weiner, Edmund S. C. 20, 39, 96, 128,
Skeat, Walter W. 26, 39 155, 180, 201, 254, 280
Söderlind, Johannes 11, 20, 152, 155 Weinreich, Otto 125, 129
Stein, Dieter 95–97, 261, 279 Welte, Werner 170, 180
Štekauer, Pavol 123, 128 Wierzbicka, Anna 251, 254
Stevens, Martin 38 Wischer, Ilse 68–70, 72–74, 87, 95–97
Stone, Louise W. 254 Wolfram, Walt 269, 270, 278, 281
Strang, Barbara M. H. 218, 229 Wright, Joseph 154
Stroheker, Friedrich 256, 257, 261, 266, Wright, Susan 219, 229
269, 274, 281 Wurff, Wim van der 200
Summers, Della 128 Wurzel, Wolfgang U. 127
Svartvik, Jan 20, 97, 128, 155, 180, 229,
269, 277, 280
Sweetser, Eve E. 151–153, 155
Szymanek, Bogdan 121, 129
</TARGET "ni">

288 Name index

Y
Yule, George 43, 64

Z
Zonneveld, Wim 123, 127
Zwicky, Arnold M. 277, 281
<TARGET "si" DOCINFO

AUTHOR ""

TITLE "Subject index"

SUBJECT "Current Issues in Linguistic Theory, Volume 223"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Subject index

A
ablative verb see verb types domain 4, 157–178 passim; see also
active form in passive sense 6, 203, 215, -wise
216, 220, 221, 227 see also in German 5, 159, 165,
progressive: passival 168–171, 174, 175, 177
adjectival modification 89 linking 177
adjective 2, 11, 76–94 passim, 99, 104, 166, sentence 5, 157–178 passim
168, 169, 171, 172, 178, 244, 249 stance 177
attributive 2, 3, 44, 46–48, 50–52, 57, subjuncts 178
61, 67, 76–94 passim, 257–261, adverbial modification 84, 89, 163
277 Ælfric 183–197 passim
comparative form 84, 275 ambiguity
compound 165, 177 semantic 79, 92, 150, 165
coordinated 51–52 syntactic 150, 165
nominalized 123, 165, 257, 277 American English see varieties of English
participial 211 anglicization of Scots 204, 206, 214, 225
possessive see determiner: possessive a-prefix(ing) 256, 269, 270, 272–274, 278,
separative 92 281
adjuncts see adverbial ARCHER Corpus 44, 49–51, 64, 115, 133,
adverb 142, 152, 187, 205, 259, 274 see 139, 140, 145, 151, 153
also adverbial argument 137, 139, 231, 239, 246, 248,
comparative form 84, 163, 275 249 see also predicate frame
in -(c)ally see -(c)ally auxiliary 74, 131–153 passim, 278 see also
in -ly 84, 94, 171, 256, 261–265 passim modality
intensifying 84, 205, 261 auxiliarihood 131
manner 84, 87, 162, 165, 166, 167, 169, modal 2, 131–153 passim, 182
173, 176, 177, 244, 245, 251 modalhood 131, 139, 141
marking 277
temporal 2, 76–90 passim, 92, 93 B
adverbial 157–178 passim, 268, 269, 277 borrowing 120, 235, 238 see also loan
adjuncts 166, 167, 169, 173, 178, 190, British National Corpus (BNC) 83, 92, 93,
191 138, 139, 146, 151, 160, 161, 167,
circumstance 177 177, 257, 263, 273, 279
conjuncts 176, 177
disjuncts 169, 177, 178
290 Subject index

C cognitive scale 249

-(c)ally 5, 159, 163, 164, 171, 172, 174, 177 cognitive structure 5, 99–101, 107, 151
case 25–27, 248 cohesion 3, 184, 186, 189, 191, 194 see also
accusative 136, 231, 232, 248 information structure
dative 75, 137, 195, 231, 232, 248 anaphoric reference 181, 184
genitive 75, 166, 167, 231, 232 colloquial English 44, 121, 143, 144, 157
nominative 113–115, 197 complement see also object
categorial indeterminacy 87, 114 infinitival 131, 142, 144, 195
Celtic influence 6, 203, 213, 214, 216, 224 nominal 132, 138, 141, 142, 144
see also progressive complex preposition 69, 70, 152
Chadwyck-Healey Literature Online 92, in Spanish 69
257–278 passim, 279 complexity, syntactic see syntactic
change see also lexical diffusion complexity
from above 213 Complexity Principle 269, 276, 278
from below 144 composite predicate 2, 9–20 passim
grammatical 149 compound stress rule 259, 261
lexical 149 conjuncts see adverbial
locus of 89 Construction Grammar 131
semantic 2, 16, 69, 84, 87, 93, 131–153 conventionalization 2, 68, 140, 146
passim conversion 71, 86, 90, 94, 118, 120, 121,
syntactic 131–153 passim, 181–199 123 see also zero-derivation
passim, 248 coordination
clausal modifier see noun phrase of adjectives 51–52
modification of adjectives and adverbs 85
clause of clauses 183–184
adverbial 205, 221, 223 of NP and VP 142
coordinated 183–184 co-referentiality 236, 238 see also verb: co-
dependent 142, 184, 194–198, 219, 221, referential intransitive/co-referential
223 reflexive
directives 143–144 Corpus of Scottish Correspondence
-ed clause 46, 48, 58–61 (CSC) 204, 206, 228
-ing clause 46, 48, 58–61, 226
main 3, 4, 139, 142, 181, 183, 184, 193, D
219, 221 Danish -vis
non-finite 206, 211, 226 de- 5, 102, 105–108
purpose 132, 139, 141, 142 decategorialization 2, 69, 88–91
relative 31, 32, 38, 46, 48, 50, 57 decliticization 75 see also clitic
subjunctive that-clause 4, 181, 184, Defoe, Daniel 133, 134, 141, 142, 145, 151
194–198 degrammaticalization 67, 69, 71, 72, 74,
to-infinitive 139 see also infinitive 75, 76, 86, 89–91, 257, 260, 275
clitic 74–76, 187 see also declitization deixis 30
COBUILD CD-ROM 9, 10, 14, 19 demorphologization 75, 76, 91, 92
cognitive complexity 126, 255, 269, 277 determiner 18, 23–38 passim
see also Complexity Principle definite 16–18, 24, 31, 33–36, 38
cognitive linguistics 100, 112, 149
 Subject index 291

demonstrative 24, 29, 30, 33, 34, 36–38, parentheticals 90

277 exaptation 90, 261, 265
indefinite 12, 16–18, 24, 28–30, 33, 36, Exceptional Case-Marking construction
204 189, 190
possessive 33, 37
detransitivization 5, 231–252 passim; see F
also transitivization; verb: -fashion 166
(in)transitive fossilization 16, 68, 72
Dickens, Charles 133, 145, 151, 199, 260 Freiburg-Lob Corpus (FLOB) 133, 146,
Dictionary of the Older Scottish Tongue 151, 160
209, 228 French 5, 16, 32, 120, 125, 178, 249
direct speech 222 see also indirect speech; Old 178, 235, 238
style of expression functional approach 5, 158, 231–252
dis- 102, 105, 106, 108 passim
discourse function 3, 33, 34, 35, 38, 189 functional dimension 48–51 see also
discourse, type of 6, 43, 171, 175, 176, Multi-Dimensional analysis
178, 203, 212, 216, 221–224 see also functional explanation 44, 161, 168, 177
register; style of expression functional factors 63, 255, 266
discreteness of linguistic categories 1, 114, Functional Grammar 232
124 functional shift 86, 87 see also conversion;
disjuncts see adverbial zero-derivation
double genitive construction 3, 23–38 functional typology of participants see
passim; see also possession; pronoun: participants, functional typology of
possessive
‘membership in a set’ meaning of 3, 28, G
29 genre see register
partitive meaning of 3, 28, 29, 34, 35 German 71, 231, 251 see also adverbial:
drunk 4, 257–260, 274–277 domain; pronoun: indefinite
drunken 4, 256–261, 275–277 nineteenth-century 170, 171
dummy element 186, 190, 198 Old High 125, 249
Dutch see pronoun: indefinite; Verb- German -(er)weise 159, 165, 168, 169, 178
Second German -mäßig 159, 170, 174, 175
Dutch -(er)wijs 165, 169 German -technisch 175
German Weise
E Germanic 113–116, 124, 165, 178, 183,
Early English Prose Fiction Corpus 257, 233, 243, 249
262 Proto-Germanic 124, 233
Eighteenth-Century Fiction Corpus 257, gerund 89, 94, 203, 205, 211, 213, 273 see
263 also verbal noun
en-/em- 5, 102, 105, 106, 108 be + a + -ing 214–216, 221, 226
-en see participial suffix -en be + in + -ing 214–216, 226
English Dialect Dictionary 144 gradience 70, 94 see also categorial
episode boundary 185, 186 indeterminacy
epistemic intersective 86–87
meaning 153
292 Subject index

grammatical variation 4, 45, 255–278 indirect speech 218, 221, 222 see also
passim direct speech; style of expression
grammaticalization 1, 2, 3, 67–94 passim, inference 2, 73, 142, 147, 150, 151, 152
131,147, 150, 152, 153, 178, 260, infinitive 3, 256, 266–269, 277
268, 275 see also bare 132, 266–269
degrammaticalization; to- 3, 131, 132, 139, 141, 142, 195, 196,
unidirectionality 197
of indefinite man 182 infinitive marker 132, 256, 266–269, 275
of English progressive 251 infinitive suffix -ian 233
of middles 246 information structure 3, 63, 181, 183–199
primary 73 passim, 255 see also episode
secondary 73 boundary
Greek stems see word-formation end focus 189
Gregory’s Dialogues 227 focus position 184
Guardian, The 157–178 passim given information 3, 184, 189, 190
new information 184, 189
H topic 184, 189
hapax legomena 161, 167, 177 see also topicalization 17, 101, 187, 194
word-formation (un)marked theme 3, 181–199 passim
Helsinki Corpus 20, 37, 79, 92, 96, 133, in-phrases 60–63
136, 137, 151, 152, 155, 204, 207, intersective gradience 86–87
225, 228, 232, 253, 257, 279 inversion 184, 187, 188 see also Verb-
Helsinki Corpus of Older Scots 204, 224, Second
228
horror aequi Principle 273, 276 L
Lancaster-Oslo-Bergen Corpus (LOB)
I 133, 146, 151, 154, 160
ICAME Collection of English Language language contact 224, 235 see also Celtic
Corpora 159, 161, 165, 171, 175, 177 influence
iconicity 90, 123, 125 see also re- Latin 26, 38, 120, 183, 192, 193
iconization lexical diffusion 64
idiolectal variation 6, 203, 205, 210, 219, lexical word 68–74, 87
224, 226 lexicalization 2, 67–94 passim; see also
idiomatization 2, 9–19 passim, 73 see also univerbation
composite predicate; opacity lexicalized items 12, 123, 165, 166,
abstractness 18 167, 176, 177
concreteness 18 primary 74
idiomatic combinations 9–19 passim, secondary 74, 87
240 likewise 165, 166, 174, 176, 177
syntactic fixing 12, 16, 17, 88 loan 111, 119, 125, 136, 238 see also
impersonal construction 137, 183 borrowing
impersonal passive see passive from French 238
implicature 140 see also inference from German 168, 169
i-mutation 104, 113, 116 locative verb see verb types
Independent, The 157–178 passim
 Subject index 293

Longman Spoken and Written English N

Corpus (LSWE) 45 Natural Morphology 124, 127
-ly suffix 256, 261–265 passim nineteenth-century English 5, 60, 64, 92,
107, 145, 146, 148, 149, 171, 173,
M 174, 249, 258, 263, 264, 267, 271,
made + infinitive 266 272, 273, 278
make 256, 266–269, 275 Nineteenth-Century Fiction Corpus 257,
marked theme see information structure 258, 263
metaphor(ical) 68, 100, 101, 104, 106, nominal features 16
119, 152, 238, 251 Norwegian -vis
metonymy 68, 118, 119, 151, 152, 235 noun
Microfiche Concordance to Old English agent 4, 78, 84, 111–126 passim
108, 232 in -er 111, 117, 118, 120, 122, 123
Middle English 4, 5, 11, 31, 119, 121, 133, in OE -end 114, 117, 125
136–137, 152, 182, 189, 196, 234, in OE -ere 117
235, 238, 240, 244, 250, 251, 256, OE n-stems 113, 114, 115, 117, 118,
258, 260, 261, 269, 277 122–124
Early 3, 29, 35, 40, 112, 117, 122, 125, attitudinal 4, 120–122
238, 254 attributive 121–123
Late 3, 106, 107, 227, 235, 252 common 54, 61, 124, 168
Middle English Dictionary (MED) 23, 232 deverbal 9, 12, 16
Middle Welsh 216 see also Celtic influence evaluative 120, 122
modality 178 (non-)oxytonic 270, 271
epistemic 153 of rank or role 79, 81, 83, 84, 89
necessity 131, 132, 135, 144–146, 149, pejorative 121
152 used as title 47, 52, 54, 57
obligation 2, 132, 135, 144 zero derivational 4, 12, 16, 111–126
volition 2, 131–153 passim; see also passim
want noun phrase 3, 10, 43–64 passim, 138, 142
Modern English 260 complex 3, 43–64 passim
Early 2, 5, 19, 20, 23, 31, 37, 67, 77, 84, informational structure of 43, 50
87, 90, 94, 101, 105, 107, 119, 131, noun phrase modification 43–64 passim
133, 136, 139, 140, 144, 151, 178, -ed clauses 46, 48, 58–61
199, 202, 228, 229, 252, 261, 266, -ing clauses 46, 48, 58–61
277 noun-noun sequences 51, 53, 54, 57,
Late 4, 9, 131, 140, 256–278 passim 61–63
modification/modifier see noun phrase postmodification 46–48, 57–64
modification premodification 46, 47, 51–57, 61–64
MoodPhrase 186–188, 198
morpheme 74 O
bound morpheme 72 object 3, 18, 101, 138, 139, 187, 194–196,
monomorphemic form 72 198, 231–252 passim, 268, 271, 277
Multi-Dimensional analysis 48–51 see also complement
dimension 48–51 control 195–196
294 Subject index

co-referential 235 see also verb: Recipient 240

co-referential reflexive Theme 250
gerundial 138 Trajector 99
Goal- 231, 232, 236, 238, 241, 242, 244, Zero 239, 245, 247
246, 247, 250 ZeroExperiencer 240, 250
Observer, The 160 participial suffix -en 256–261, 265,
OE verbal prefix ge- 234 274–275, 277
OED Online 25, 152, 171 see also Oxford participle see also clause: -ed clauses/-ing
English Dictionary (OED) clauses
of-phrases 46, 58–63 present 203–227 passim
Old English 4, 5, 20, 29, 37–38, 101–108 being + -ing 217
passim, 111–126 passim, 165–166, in -and 204–227 passim
168, 181–198 passim, 231–252 in -en(e) /-in(e) 208–214
passim, 257, 258, 260, 269, 275, 277 in -ing 203–227 passim, 256,
pre- 114, 116, 124 269–274 passim, 278
Old Frisian 249 in nominal function 214, 226
Old Icelandic 249 past 4, 210, 225 see also participial
Old Norse 136, 139 suffix -en
Old Saxon 249 passival see progressive: passival
opacity/opaqueness 12, 17, 68, 69, 71, 72, passive 2, 4, 6, 9–19 passim, 185, 186,
73, 74, 182 190–194, 197, 198, 242, 277 see also
optimal syllable structure 269, 276 progressive: passival; voice – aspect
ornative verb see verb types system
otherwise 165, 166, 174, 176, 177 double 10, 12
Oxford English Dictionary (OED) 9, 25, 92, impersonal 4, 181, 183, 189, 192
232 see also OED Online inner 2, 9–19 passim
of make 256, 266–269, 275, 278
P outer 2, 9–19 passim
participants, functional typology of 92, prepositional 9–19 passim, 189
231–252 passim progressive 236, 246, 251–252
affected 231, 232, 240, 243, 247, 248 passive sense see active form in passive
Agent 99, 108, 139, 142, 198, 231, 236, sense; progressive: passival
237, 240, 242–247, 249, 251 passivization 2, 9–19 passim
Cause 137 Penn-Helsinki Parsed Corpus of Middle
Causer 108 English (PPCME) 37, 39
Experiencer 137 phonogenesis 75, 76, 89, 96
Force/Instrument 242 possession
Goal 231, 232, 236–239, 241, 242, 244, (in)alienable 100
246, 247, 250 (pro)nominal possessors 23–38 passim
Instrument 108, 251 pragmatic inference see inference
Landmark 100 pragmatic marker 92
non-agentive 241, 248 predicate frame 236, 239, 242, 244, 247,
Patient 192, 248 249 see also argument; verb
Processed 241, 242, 246–248 preposition to + -ing 273
ProcessedExperiencer 240, 250
 Subject index 295

prepositional phrase 9–19 passim, 46–64 R

passim; see also in-phrases; reanalysis 16, 17, 68, 69, 89
of-phrases; with-phrases recategorialization 86, 90
Present-day English 5, 9, 15, 17, 31, 43–64 regrammaticalization 198, 260
passim, 115, 133, 135, 140–146, 149, re-iconization 117 see also iconicity
157–178 passim, 183, 184, 189, 191, register 43–64 passim, 136, 174, 178, 183,
193, 197, 198, 261, 262, 265, 268 199 see also colloquial English; style
Principle of Rhythmic Alternation 4, of expression
255–278 passim; see also stress clash; conversation 43–48, 64
stress lapse drama 51–64 passim
rhythmic buffer 260, 261, 265, 266, essay 10
268, 270–273, 278 fiction 43–64 passim
privative verb see verb types informational 48, 51, 62, 63, 174
PRO letters 10, 205–221 passim, 226–227
arbitrary/generic 4, 181, 195–96, 198 medical prose 49–52, 54, 56–58, 59,
object-controlled 195, 197 60–62, 64, 174
productivity see word-formation narrative 208, 219, 221–224, 226
progressive 6, 201–229 passim; see also news 43–64 passim
gerund; participle: present; passive: oral 64, 216
progressive religion 178
Celtic influence on 6, 203, 213, 214, science 5, 49, 50, 107, 173–176, 178,
216 215, 225
in dependent clauses 218, 219, 221–224 speech-based 6, 49, 50, 220–224
passival 246, 251–252 trial proceedings 207, 208, 221–224,
pronoun 227
indefinite 3, 24, 181–199 passim written 64, 174
Dutch men 181, 191–193, 196, 197, reversative verb see verb types
199 Robinson Crusoe 133, 134, 141, 142, 145,
German man 181, 196 151
inclusion/exclusion of interlocutors
183, 192, 193 S
ME me(n) 182, 197 scarce(ly) 256, 262–266, 274, 277
OE man 3, 4, 181–199 passim Scots
PDE one 181, 182, 183, 185, 192, anglicization of Scots 204, 206, 214, 225
193, 197, 199 Older 203–227 passim
possessive 23–38 passim S-curve 64, 148
(un)strengthened 25–28, 36, 37 semantic categories of verbs see verb types
reflexive 236, 250, 270, 271 semantic role see participants, functional
prosodic movement see Principle of typology of
Rhythmic Alternation; sentence semanticization 73, 91
prosody sentence prosody 267, 268, 274, 276 see
also Principle of Rhythmic
Q Alternation
quantifier 24, 28, 33 set 270–275
296 Subject index

Shakespeare, William 31, 33, 78, 79, 81, topicalization 17, 101, 187, 194
85, 133, 151, 189, 235 transformational grammar 232
Spanish 178 transitivity 232, 248, 277
standardization pressures 265, 273, 274, transitivization 231–252 passim; see also
276 see also varieties of English: non- detransitivization; transitivity; verb:
standard transitive
states of affairs see also verb types translation
dynamic 232, 237, 239, 241, 242, 245, from Latin 26, 183, 192, 193
246, 248, 250, 251
Process 241, 242 U
stative 239 unidirectionality 75, 89, 95 see also
telic 237 see also telicity grammaticalization
State 245, 247 univerbation 68, 72, 89 see also
(un)controlled 237, 241–243, 245, 246 lexicalization
stress clash 258, 260–262, 266, 270–272,
274 see also Principle of Rhythmic V
Alternation variation see grammatical variation;
stress lapse 261, 269, 276 see also Principle idiolectal variation
of Rhythmic Alternation varieties of English see also colloquial
stress shift rule 259, 261 English
-style 167 American 49, 50, 93, 157, 158, 159,
style of expression see also direct speech; 175–177, 252
indirect speech; register Australian 159
compressed 3, 62, 63, 174 Belfast 249, 253
expanded 62 East African 159, 160, 177
graphic presentation 223 Indian 159
subject 139, 184–199 passim New Zealand 159, 160, 175
expletive 186, 190, 198 non-standard 93, 261, 272, 273
nominal 187, 188, 198 Scottish 277
pronominal 187, 188, 198 verb see also predicate frame
subjectification 147, 155 catenative 2
subject-to-object raising see Exceptional co-referential intransitive 5, 236–238,
Case-Marking construction 240, 242, 247, 249–251
subjuncts see adverbial co-referential reflexive 235, 236–238,
Swedish 165, 169, 181 249–251
syntactic complexity 43–64 passim, 221, deadjectival 104, 106, 244, 249
223, 268, 277 see also noun phrase: denominal 104–107
complex ditransitive 240
ergative use 236, 240–244, 247–249,
T 251
telicity 237, 242, 243, 245, 246 generic use 232, 236, 239–241, 247,
text type see register 250, 251
textual cohesion see cohesion intransitive 217–218, 231–252 passim,
theme see information structure 270
topic see information structure medio-passive 244
</TARGET "si">

Subject index 297

middle use 5, 236, 240, 244–248, 251 relational 218, 227

modal see auxiliary reversative 5, 99–108 passim
strong 233 state 217, 227
transitive 137, 139, 142, 217, 218, 227, stative 17, 233, 234, 242
231–252 passim, 270 transfer 240, 245
transitive – intransitive 233–235 verbal noun 205, 212 see also gerund
weak 233 ‘V+NP+P’ construction 9–19 passim
verb formation 99–108 passim, 233–234 voice – aspect system 246 see also passive
French influence on 101, 106, 107–108 volition see modality
see also de-, dis-, en-/em-
Latin influence on 101, 105–108 W
Verb-Second 181, 186–191 see also want 131–153 passim
inversion West-Saxon 182
in Dutch 189–191 -wise 4, 5, 157–178 passim
loss of 3, 181, 188, 189, 193, 194, 197 with-phrases 60–63
verb types word class
ablative 5, 99–109 grammatical 68–74, 87, 91
body-care 238, 247 lexical 68–74, 87
body movement 238, 247 word-formation 99–108 passim, 111–126
causative 234, 241, 247, 249, 250 passim; see also hapax legomena;
causative – inchoative 99, 233, 243 verb formation
change-of-state 245 expressive 121
cognition 245 Greek bases used in 5, 171, 173, 174,
contact 251 177, 178
dynamic 17 (Neo-)Latin bases used in 5, 163, 171,
evaluative 194, 196 173, 174, 177, 178
experience 240 productivity 4, 5, 99–108 passim,
inchoative 233, 234, 241–244, 249 111–126 passim, 158–161,
locative 5, 99–108 passim 165–167, 169, 176–177, 246
motion 217, 218, 226, 227, 235, 251 Romance bases used in 106, 178
ornative 5, 99–108 passim
of commanding 194–196, 266 Z
of permitting 194–196 zero-derivation 4, 102–105, 111–126
of persuading and urging 194–197 passim
privative 5, 99–108 passim
 CURRENT ISSUES IN LINGUISTIC THEORY

E. F. K. Koerner, Editor
Department of Linguistics, University of Ottawa
OTTAWA, Canada K1N 6N5
koerner@[Link]

The Current Issues in Linguistic Theory (CILT) series is a theory-oriented series which welcomes
contributions from scholars who have significant proposals to make towards the advancement of
our understanding of language, its structure, functioning and development. CILT has been
established in order to provide a forum for the presentation and discussion of linguistic opinions
of scholars who do not necessarily accept the prevailing mode of thought in linguistic science. It
offers an alternative outlet for meaningful contributions to the current linguistic debate, and
furnishes the diversity of opinion which a healthy discipline must have. In this series the following
volumes have been published thus far or are scheduled for publication:

1. KOERNER, Konrad (ed.): The Transformational-Generative Paradigm and Modern Linguistic

Theory. 1975.
2. WEIDERT, Alfons: Componential Analysis of Lushai Phonology. 1975.
3. MAHER, J. Peter: Papers on Language Theory and History I: Creation and Tradition in
Language. Foreword by Raimo Anttila. 1979.
4. HOPPER, Paul J. (ed.): Studies in Descriptive and Historical Linguistics. Festschrift for Winfred
P. Lehmann. 1977.
5. ITKONEN, Esa: Grammatical Theory and Metascience: A critical investigation into the
methodological and philosophical foundations of ‘autonomous’ linguistics. 1978.
6. ANTTILA, Raimo: Historical and Comparative Linguistics. 1989.
7. MEISEL, Jürgen M. & Martin D. PAM (eds): Linear Order and Generative Theory. 1979.
8. WILBUR, Terence H.: Prolegomena to a Grammar of Basque. 1979.
9. HOLLIEN, Harry & Patricia (eds): Current Issues in the Phonetic Sciences. Proceedings of the
IPS-77 Congress, Miami Beach, Florida, 17-19 December 1977. 1979.
10. PRIDEAUX, Gary D. (ed.): Perspectives in Experimental Linguistics. Papers from the Univer-
sity of Alberta Conference on Experimental Linguistics, Edmonton, 13-14 Oct. 1978. 1979.
11. BROGYANYI, Bela (ed.): Studies in Diachronic, Synchronic, and Typological Linguistics:
Festschrift for Oswald Szemérenyi on the Occasion of his 65th Birthday. 1979.
12. FISIAK, Jacek (ed.): Theoretical Issues in Contrastive Linguistics. 1981. Out of print
13. MAHER, J. Peter, Allan R. BOMHARD & Konrad KOERNER (eds): Papers from the Third
International Conference on Historical Linguistics, Hamburg, August 22-26 1977. 1982.
14. TRAUGOTT, Elizabeth C., Rebecca LaBRUM & Susan SHEPHERD (eds): Papers from the
Fourth International Conference on Historical Linguistics, Stanford, March 26-30 1979. 1980.
15. ANDERSON, John (ed.): Language Form and Linguistic Variation. Papers dedicated to Angus
McIntosh. 1982.
16. ARBEITMAN, Yoël L. & Allan R. BOMHARD (eds): Bono Homini Donum: Essays in
Historical Linguistics, in Memory of [Link] Kerns. 1981.
17. LIEB, Hans-Heinrich: Integrational Linguistics. 6 volumes. Vol. II-VI n.y.p. 1984/93.
18. IZZO, Herbert J. (ed.): Italic and Romance. Linguistic Studies in Honor of Ernst Pulgram.
1980.
19. RAMAT, Paolo et al. (eds): Linguistic Reconstruction and Indo-European Syntax. Proceedings
of the Colloquium of the ‘Indogermanischhe Gesellschaft’. University of Pavia, 6-7 September
1979. 1980.
20. NORRICK, Neal R.: Semiotic Principles in Semantic Theory. 1981.
21. AHLQVIST, Anders (ed.): Papers from the Fifth International Conference on Historical
Linguistics, Galway, April 6-10 1981. 1982.
22. UNTERMANN, Jürgen & Bela BROGYANYI (eds): Das Germanische und die Rekonstruktion
der Indogermanischen Grundsprache. Akten des Freiburger Kolloquiums der Indogermanischen
Gesellschaft, Freiburg, 26-27 Februar 1981. 1984.
23. DANIELSEN, Niels: Papers in Theoretical Linguistics. Edited by Per Baerentzen. 1992.
24. LEHMANN, Winfred P. & Yakov MALKIEL (eds): Perspectives on Historical Linguistics.
Papers from a conference held at the meeting of the Language Theory Division, Modern
Language Assn., San Francisco, 27-30 December 1979. 1982.
25. ANDERSEN, Paul Kent: Word Order Typology and Comparative Constructions. 1983.
26. BALDI, Philip (ed.): Papers from the XIIth Linguistic Symposium on Romance Languages,
Univ. Park, April 1-3, 1982. 1984.
27. BOMHARD, Alan R.: Toward Proto-Nostratic. A New Approach to the Comparison of Proto-
Indo-European and Proto-Afroasiatic. Foreword by Paul J. Hopper. 1984.
28. BYNON, James (ed.): Current Progress in Afro-Asiatic Linguistics: Papers of the Third
International Hamito-Semitic Congress, London, 1978. 1984.
29. PAPROTTÉ, Wolf & René DIRVEN (eds): The Ubiquity of Metaphor: Metaphor in language
and thought. 1985 (publ. 1986).
30. HALL, Robert A. Jr.: Proto-Romance Morphology. = Comparative Romance Grammar, vol.
III. 1984.
31. GUILLAUME, Gustave: Foundations for a Science of Language.
32. COPELAND, James E. (ed.): New Directions in Linguistics and Semiotics. Co-edition with
Rice University Press who hold exclusive rights for US and Canada. 1984.
33. VERSTEEGH, Kees: Pidginization and Creolization. The Case of Arabic. 1984.
34. FISIAK, Jacek (ed.): Papers from the VIth International Conference on Historical Linguistics,
Poznan, 22-26 August. 1983. 1985.
35. COLLINGE, N.E.: The Laws of Indo-European. 1985.
36. KING, Larry D. & Catherine A. MALEY (eds): Selected papers from the XIIIth Linguistic
Symposium on Romance Languages, Chapel Hill, N.C., 24-26 March 1983. 1985.
37. GRIFFEN, T.D.: Aspects of Dynamic Phonology. 1985.
38. BROGYANYI, Bela & Thomas KRÖMMELBEIN (eds): Germanic Dialects:Linguistic and
Philological Investigations. 1986.
39. BENSON, James D., Michael J. CUMMINGS, & William S. GREAVES (eds): Linguistics in a
Systemic Perspective. 1988.
40. FRIES, Peter Howard (ed.) in collaboration with Nancy M. Fries: Toward an Understanding
of Language: Charles C. Fries in Perspective. 1985.
41. EATON, Roger, et al. (eds): Papers from the 4th International Conference on English Historical
Linguistics, April 10-13, 1985. 1985.
42. MAKKAI, Adam & Alan K. MELBY (eds): Linguistics and Philosophy. Festschrift for Rulon S.
Wells. 1985 (publ. 1986).
43. AKAMATSU, Tsutomu: The Theory of Neutralization and the Archiphoneme in Functional
Phonology. 1988.
44. JUNGRAITHMAYR, Herrmann & Walter W. MUELLER (eds): Proceedings of the Fourth
International Hamito-Semitic Congress. 1987.
45. KOOPMAN, W.F., F.C. Van der LEEK , O. FISCHER & R. EATON (eds): Explanation and
Linguistic Change. 1986
46. PRIDEAUX, Gary D. & William J. BAKER: Strategies and Structures: The processing of relative
clauses. 1987.
47. LEHMANN, Winfred P. (ed.): Language Typology 1985. Papers from the Linguistic Typology
Symposium, Moscow, 9-13 Dec. 1985. 1986.
48. RAMAT, Anna G., Onofrio CARRUBA and Giuliano BERNINI (eds): Papers from the 7th
International Conference on Historical Linguistics. 1987.
49. WAUGH, Linda R. and Stephen RUDY (eds): New Vistas in Grammar: Invariance and
 Variation. Proceedings of the Second International Roman Jakobson Conference, New York
University, Nov.5-8, 1985. 1991.
50. RUDZKA-OSTYN, Brygida (ed.): Topics in Cognitive Linguistics. 1988.
51. CHATTERJEE, Ranjit: Aspect and Meaning in Slavic and Indic. With a foreword by Paul
Friedrich. 1989.
52. FASOLD, Ralph W. & Deborah SCHIFFRIN (eds): Language Change and Variation. 1989.
53. SANKOFF, David: Diversity and Diachrony. 1986.
54. WEIDERT, Alfons: Tibeto-Burman Tonology. A comparative analysis. 1987
55. HALL, Robert A. Jr.: Linguistics and Pseudo-Linguistics. 1987.
56. HOCKETT, Charles F.: Refurbishing our Foundations. Elementary linguistics from an ad-
vanced point of view. 1987.
57. BUBENIK, Vít: Hellenistic and Roman Greece as a Sociolinguistic Area. 1989.
58. ARBEITMAN, Yoël. L. (ed.): Fucus: A Semitic/Afrasian Gathering in Remembrance of Albert
Ehrman. 1988.
59. VAN VOORST, Jan: Event Structure. 1988.
60. KIRSCHNER, Carl & Janet DECESARIS (eds): Studies in Romance Linguistics. Selected
Proceedings from the XVII Linguistic Symposium on Romance Languages. 1989.
61. CORRIGAN, Roberta L., Fred ECKMAN & Michael NOONAN (eds): Linguistic Categoriza-
tion. Proceedings of an International Symposium in Milwaukee, Wisconsin, April 10-11, 1987.
1989.
62. FRAJZYNGIER, Zygmunt (ed.): Current Progress in Chadic Linguistics. Proceedings of the
International Symposium on Chadic Linguistics, Boulder, Colorado, 1-2 May 1987. 1989.
63. EID, Mushira (ed.): Perspectives on Arabic Linguistics I. Papers from the First Annual
Symposium on Arabic Linguistics. 1990.
64. BROGYANYI, Bela (ed.): Prehistory, History and Historiography of Language, Speech, and
Linguistic Theory. Papers in honor of Oswald Szemérenyi I. 1992.
65. ADAMSON, Sylvia, Vivien A. LAW, Nigel VINCENT and Susan WRIGHT (eds): Papers from
the 5th International Conference on English Historical Linguistics. 1990.
66. ANDERSEN, Henning and Konrad KOERNER (eds): Historical Linguistics [Link] from
the 8th International Conference on Historical Linguistics,Lille, August 30-Sept., 1987. 1990.
67. LEHMANN, Winfred P. (ed.): Language Typology 1987. Systematic Balance in Language.
Papers from the Linguistic Typology Symposium, Berkeley, 1-3 Dec 1987. 1990.
68. BALL, Martin, James FIFE, Erich POPPE &Jenny ROWLAND (eds): Celtic Linguistics/
Ieithyddiaeth Geltaidd. Readings in the Brythonic Languages. Festschrift for T. Arwyn Watkins.
1990.
69. WANNER, Dieter and Douglas A. KIBBEE (eds): New Analyses in Romance Linguistics.
Selected papers from the Linguistic Symposium on Romance Languages XVIIII, Urbana-
Champaign, April 7-9, 1988. 1991.
70. JENSEN, John T.: Morphology. Word structure in generative grammar. 1990.
71. O’GRADY, William: Categories and Case. The sentence structure of Korean. 1991.
72. EID, Mushira and John MCCARTHY (eds): Perspectives on Arabic Linguistics II. Papers from
the Second Annual Symposium on Arabic Linguistics. 1990.
73. STAMENOV, Maxim (ed.): Current Advances in Semantic Theory. 1991.
74. LAEUFER, Christiane and Terrell A. MORGAN (eds): Theoretical Analyses in Romance
Linguistics. 1991.
75. DROSTE, Flip G. and John E. JOSEPH (eds): Linguistic Theory and Grammatical Description.
Nine Current Approaches. 1991.
76. WICKENS, Mark A.: Grammatical Number in English Nouns. An empirical and theoretical
account. 1992.
77. BOLTZ, William G. and Michael C. SHAPIRO (eds): Studies in the Historical Phonology of
Asian Languages. 1991.
78. KAC, Michael: Grammars and Grammaticality. 1992.
79. ANTONSEN, Elmer H. and Hans Henrich HOCK (eds): STAEF-CRAEFT: Studies in Ger-
manic Linguistics. Select papers from the First and Second Symposium on Germanic Linguistics,
University of Chicago, 24 April 1985, and Univ. of Illinois at Urbana-Champaign, 3-4 Oct. 1986.
1991.
80. COMRIE, Bernard and Mushira EID (eds): Perspectives on Arabic Linguistics III. Papers from
the Third Annual Symposium on Arabic Linguistics. 1991.
81. LEHMANN, Winfred P. and H.J. HEWITT (eds): Language Typology 1988. Typological
Models in the Service of Reconstruction. 1991.
82. VAN VALIN, Robert D. (ed.): Advances in Role and Reference Grammar. 1992.
83. FIFE, James and Erich POPPE (eds): Studies in Brythonic Word Order. 1991.
84. DAVIS, Garry W. and Gregory K. IVERSON (eds): Explanation in Historical Linguistics. 1992.
85. BROSELOW, Ellen, Mushira EID and John McCARTHY (eds): Perspectives on Arabic
Linguistics IV. Papers from the Annual Symposium on Arabic Linguistics. 1992.
86. KESS, Joseph F.: Psycholinguistics. Psychology, linguistics, and the study of natural language.
1992.
87. BROGYANYI, Bela and Reiner LIPP (eds): Historical Philology: Greek, Latin, and Romance.
Papers in honor of Oswald Szemerényi II. 1992.
88. SHIELDS, Kenneth: A History of Indo-European Verb Morphology. 1992.
89. BURRIDGE, Kate: Syntactic Change in Germanic. A study of some aspects of language change
in Germanic with particular reference to Middle Dutch. 1992.
90. KING, Larry D.: The Semantic Structure of Spanish. Meaning and grammatical form. 1992.
91. HIRSCHBÜHLER, Paul and Konrad KOERNER (eds): Romance Languages and Modern
Linguistic Theory. Selected papers from the XX Linguistic Symposium on Romance
Languages,University of Ottawa, April 10-14, 1990. 1992.
92. POYATOS, Fernando: Paralanguage: A linguistic and interdisciplinary approach to interactive
speech and sounds. 1992.
93. LIPPI-GREEN, Rosina (ed.): Recent Developments in Germanic Linguistics. 1992.
94. HAGÈGE, Claude: The Language Builder. An essay on the human signature in linguistic
morphogenesis. 1992.
95. MILLER, D. Gary: Complex Verb Formation. 1992.
96. LIEB, Hans-Heinrich (ed.): Prospects for a New Structuralism. 1992.
97. BROGYANYI, Bela & Reiner LIPP (eds): Comparative-Historical Linguistics: Indo-European
and Finno-Ugric. Papers in honor of Oswald Szemerényi III. 1992.
98. EID, Mushira & Gregory K. IVERSON: Principles and Prediction: The analysis of natural
language. 1993.
99. JENSEN, John T.: English Phonology. 1993.
100. MUFWENE, Salikoko S. and Lioba MOSHI (eds): Topics in African Linguistics. Papers from
the XXI Annual Conference on African Linguistics, University of Georgia, April 1990. 1993.
101. EID, Mushira & Clive HOLES (eds): Perspectives on Arabic Linguistics V. Papers from the Fifth
Annual Symposium on Arabic Linguistics. 1993.
102. DAVIS, Philip W. (ed.): Alternative Linguistics. Descriptive and theoretical Modes. 1995.
103. ASHBY, William J., Marianne MITHUN, Giorgio PERISSINOTTO and Eduardo RAPOSO:
Linguistic Perspectives on Romance Languages. Selected papers from the XXI Linguistic Sympo-
sium on Romance Languages, Santa Barbara, February 21-24, 1991. 1993.
104. KURZOVÁ, Helena: From Indo-European to Latin. The evolution of a morphosyntactic type.
1993.
105. HUALDE, José Ignacio and Jon ORTIZ DE URBANA (eds): Generative Studies in Basque
Linguistics. 1993.
106. AERTSEN, Henk and Robert J. JEFFERS (eds): Historical Linguistics 1989. Papers from the 9th
International Conference on Historical Linguistics, New Brunswick, 14-18 August 1989. 1993.
107. MARLE, Jaap van (ed.): Historical Linguistics 1991. Papers from the 10th International
Conference on Historical Linguistics, Amsterdam, August 12-16, 1991. 1993.
108. LIEB, Hans-Heinrich: Linguistic Variables. Towards a unified theory of linguistic variation.
1993.
109. PAGLIUCA, William (ed.): Perspectives on Grammaticalization. 1994.
110. SIMONE, Raffaele (ed.): Iconicity in Language. 1995.
111. TOBIN, Yishai: Invariance, Markedness and Distinctive Feature Analysis. A contrastive study of
sign systems in English and Hebrew. 1994.
112. CULIOLI, Antoine: Cognition and Representation in Linguistic Theory. Translated, edited and
introduced by Michel Liddle. 1995.
113. FERNÁNDEZ, Francisco, Miguel FUSTER and Juan Jose CALVO (eds): English Historical
Linguistics 1992. Papers from the 7th International Conference on English Historical Linguis-
tics, Valencia, 22-26 September 1992.1994.
114. EGLI, U., P. PAUSE, Chr. SCHWARZE, A. von STECHOW, G. WIENOLD (eds): Lexical
Knowledge in the Organisation of Language. 1995.
115. EID, Mushira, Vincente CANTARINO and Keith WALTERS (eds): Perspectives on Arabic
Linguistics. Vol. VI. Papers from the Sixth Annual Symposium on Arabic Linguistics. 1994.

116. MILLER, D. Gary: Ancient Scripts and Phonological Knowledge. 1994.

117. PHILIPPAKI-WARBURTON, I., K. NICOLAIDIS and M. SIFIANOU (eds): Themes in
Greek Linguistics. Papers from the first International Conference on Greek Linguistics, Reading,
September 1993. 1994.
118. HASAN, Ruqaiya and Peter H. FRIES (eds): On Subject and Theme. A discourse functional
perspective. 1995.
119. LIPPI-GREEN, Rosina: Language Ideology and Language Change in Early Modern German. A
sociolinguistic study of the consonantal system of Nuremberg. 1994.
120. STONHAM, John T. : Combinatorial Morphology. 1994.
121. HASAN, Ruqaiya, Carmel CLORAN and David BUTT (eds): Functional Descriptions. Theorie
in practice. 1996.
122. SMITH, John Charles and Martin MAIDEN (eds): Linguistic Theory and the Romance
Languages. 1995.
123. AMASTAE, Jon, Grant GOODALL, Mario MONTALBETTI and Marianne PHINNEY:
Contemporary Research in Romance Linguistics. Papers from the XXII Linguistic Symposium
on Romance Languages, El Paso//Juárez, February 22-24, 1994. 1995.
124. ANDERSEN, Henning: Historical Linguistics 1993. Selected papers from the 11th International
Conference on Historical Linguistics, Los Angeles, 16-20 August 1993. 1995.
125. SINGH, Rajendra (ed.): Towards a Critical Sociolinguistics. 1996.
126. MATRAS, Yaron (ed.): Romani in Contact. The history, structure and sociology of a language.
1995.
127. GUY, Gregory R., Crawford FEAGIN, Deborah SCHIFFRIN and John BAUGH (eds):
Towards a Social Science of Language. Papers in honor of William Labov. Volume 1: Variation
and change in language and society. 1996.
128. GUY, Gregory R., Crawford FEAGIN, Deborah SCHIFFRIN and John BAUGH (eds):
Towards a Social Science of Language. Papers in honor of William Labov. Volume 2: Social
interaction and discourse structures. 1997.
129. LEVIN, Saul: Semitic and Indo-European: The Principal Etymologies. With observations on
Afro-Asiatic. 1995.
130. EID, Mushira (ed.) Perspectives on Arabic Linguistics. Vol. VII. Papers from the Seventh
Annual Symposium on Arabic Linguistics. 1995.
131. HUALDE, Jose Ignacio, Joseba A. LAKARRA and R.L. Trask (eds): Towards a History of the
Basque Language. 1995.
132. HERSCHENSOHN, Julia: Case Suspension and Binary Complement Structure in French. 1996.
133. ZAGONA, Karen (ed.): Grammatical Theory and Romance Languages. Selected papers from
the 25th Linguistic Symposium on Romance Languages (LSRL XXV) Seattle, 2-4 March 1995.
1996.
134. EID, Mushira (ed.): Perspectives on Arabic Linguistics Vol. VIII. Papers from the Eighth
Annual Symposium on Arabic Linguistics. 1996.
135. BRITTON Derek (ed.): Papers from the 8th International Conference on English Historical
Linguistics. 1996.
136. MITKOV, Ruslan and Nicolas NICOLOV (eds): Recent Advances in Natural Language
Processing. 1997.
137. LIPPI-GREEN, Rosina and Joseph C. SALMONS (eds): Germanic Linguistics. Syntactic and
diachronic. 1996.
138. SACKMANN, Robin (ed.): Theoretical Linguistics and Grammatical Description. 1996.
139. BLACK, James R. and Virginia MOTAPANYANE (eds): Microparametric Syntax and Dialect
Variation. 1996.
140. BLACK, James R. and Virginia MOTAPANYANE (eds): Clitics, Pronouns and Movement.
1997.
141. EID, Mushira and Dilworth PARKINSON (eds): Perspectives on Arabic Linguistics Vol. IX.
Papers from the Ninth Annual Symposium on Arabic Linguistics, Georgetown University,
Washington D.C., 1995. 1996.
142. JOSEPH, Brian D. and Joseph C. SALMONS (eds): Nostratic. Sifting the evidence. 1998.
143. ATHANASIADOU, Angeliki and René DIRVEN (eds): On Conditionals Again. 1997.
144. SINGH, Rajendra (ed): Trubetzkoy's Orphan. Proceedings of the Montréal Roundtable
“Morphophonology: contemporary responses (Montréal, October 1994). 1996.
145. HEWSON, John and Vit BUBENIK: Tense and Aspect in Indo-European Languages. Theory,
typology, diachrony. 1997.
146. HINSKENS, Frans, Roeland VAN HOUT and W. Leo WETZELS (eds): Variation, Change,
and Phonological Theory. 1997.
147. HEWSON, John: The Cognitive System of the French Verb. 1997.
148. WOLF, George and Nigel LOVE (eds): Linguistics Inside Out. Roy Harris and his critics. 1997.
149. HALL, T. Alan: The Phonology of Coronals. 1997.
150. VERSPOOR, Marjolijn, Kee Dong LEE and Eve SWEETSER (eds): Lexical and Syntactical
Constructions and the Construction of Meaning. Proceedings of the Bi-annual ICLA meeting in
Albuquerque, July 1995. 1997.
151. LIEBERT, Wolf-Andreas, Gisela REDEKER and Linda WAUGH (eds): Discourse and Per-
spectives in Cognitive Linguistics. 1997.
152. HIRAGA, Masako, Chris SINHA and Sherman WILCOX (eds): Cultural, Psychological and
Typological Issues in Cognitive Linguistics. 1999.
153. EID, Mushira and Robert R. RATCLIFFE (eds): Perspectives on Arabic Linguistics Vol. X.
Papers from the Tenth Annual Symposium on Arabic Linguistics, Salt Lake City, 1996. 1997.
154. SIMON-VANDENBERGEN, Anne-Marie, Kristin DAVIDSE and Dirk NOËL (eds): Recon-
necting Language. Morphology and Syntax in Functional Perspectives. 1997.
155. FORGET, Danielle, Paul HIRSCHBÜHLER, France MARTINEAU and María-Luisa
RIVERO (eds): Negation and Polarity. Syntax and semantics. Selected papers from the Collo-
quium Negation: Syntax and Semantics. Ottawa, 11-13 May 1995. 1997.
156. MATRAS, Yaron, Peter BAKKER and Hristo KYUCHUKOV (eds): The Typology and
Dialectology of Romani. 1997.
157. LEMA, José and Esthela TREVIÑO (eds): Theoretical Analyses on Romance Languages.
Selected papers from the 26th Linguistic Symposium on Romance Languages (LSRL XXVI),
Mexico City, 28-30 March, 1996. 1998.
158. SÁNCHEZ MACARRO, Antonia and Ronald CARTER (eds): Linguistic Choice across Genres.
Variation in spoken and written English. 1998.
159. JOSEPH, Brian D., Geoffrey C. HORROCKS and Irene PHILIPPAKI-WARBURTON (eds):
Themes in Greek Linguistics II. 1998.
160. SCHWEGLER, Armin, Bernard TRANEL and Myriam URIBE-ETXEBARRIA (eds): Ro-
mance Linguistics: Theoretical Perspectives. Selected papers from the 27th Linguistic Symposium
on Romance Languages (LSRL XXVII), Irvine, 20-22 February, 1997. 1998.
161. SMITH, John Charles and Delia BENTLEY (eds): Historical Linguistics 1995. Volume 1:
Romance and general linguistics. 2000.
162. HOGG, Richard M. and Linda van BERGEN (eds): Historical Linguistics 1995. Volume 2:
Germanic [Link] papers from the 12th International Conference on Historical
Linguistics, Manchester, August 1995. 1998.
163. LOCKWOOD, David G., Peter H. FRIES and James E. COPELAND (eds): Functional
Approaches to Language, Culture and Cognition. 2000.
164. SCHMID, Monika, Jennifer R. AUSTIN and Dieter STEIN (eds): Historical Linguistics 1997.
Selected papers from the 13th International Conference on Historical Linguistics, Düsseldorf, 10-
17 August 1997. 1998.
165. BUBENÍK, Vit: A Historical Syntax of Late Middle Indo-Aryan (Apabhramsa). ´ 1998.
166. LEMMENS, Maarten: Lexical Perspectives on Transitivity and Ergativity. Causative construc-
tions in English. 1998.
167. BENMAMOUN, Elabbas, Mushira EID and Niloofar HAERI (eds): Perspectives on Arabic
Linguistics Vol. XI. Papers from the Eleventh Annual Symposium on Arabic Linguistics,
Atlanta, 1997. 1998.
168. RATCLIFFE, Robert R.: The “Broken” Plural Problem in Arabic and Comparative Semitic.
Allomorphy and analogy in non-concatenative morphology. 1998.
169. GHADESSY, Mohsen (ed.): Text and Context in Functional Linguistics. 1999.
170. LAMB, Sydney M.: Pathways of the Brain. The neurocognitive basis of language. 1999.
171. WEIGAND, Edda (ed.): Contrastive Lexical Semantics. 1998.
172. DIMITROVA-VULCHANOVA, Mila and Lars HELLAN (eds): Topics in South Slavic Syntax
and Semantics. 1999.
173. TREVIÑO, Esthela and José LEMA (eds): Semantic Issues in Romance Syntax. 1999.
174. HALL, T. Alan and Ursula KLEINHENZ (eds): Studies on the Phonological Word. 1999.
175. GIBBS, Ray W. and Gerard J. STEEN (eds): Metaphor in Cognitive Linguistics. Selected papers
from the 5th International Cognitive Linguistics Conference, Amsterdam, 1997. 2001.
176. VAN HOEK, Karen, Andrej KIBRIK and Leo NOORDMAN (eds): Discourse in Cognitive
Linguistics. Selected papers from the International Cognitive Linguistics Conference, Amster-
dam, July 1997. 1999.
177. CUYCKENS, Hubert and Britta ZAWADA (eds): Polysemy in Cognitive Linguistics. Selected
papers from the International Cognitive Linguistics Conference, Amsterdam, 1997. 2001.
178. FOOLEN, Ad and Frederike van der LEEK (eds): Constructions in Cognitive Linguistics.
Selected papers from the Fifth International Cognitive Linguistic Conference, Amsterdam, 1997.
2000.
179. RINI, Joel: Exploring the Role of Morphology in the Evolution of Spanish. 1999.
180. MEREU, Lunella (ed.): Boundaries of Morphology and Syntax. 1999.
181. MOHAMMAD, Mohammad A.: Word Order, Agreement and Pronominalization in Standard
and Palestinian Arabic. 2000.
182. KENESEI, István (ed.): Theoretical Issues in Eastern European Languages. Selected papers from
the Conference on Linguistic Theory in Eastern European Languages (CLITE), Szeged, April
1998. 1999.
183. CONTINI-MORAVA, Ellen and Yishai TOBIN (eds): Between Grammar and Lexicon. 2000.
184. SAGART, Laurent: The Roots of Old Chinese. 1999.
185. AUTHIER, J.-Marc, Barbara E. BULLOCK, Lisa A. REED (eds): Formal Perspectives on
Romance Linguistics. Selected papers from the 28th Linguistic Symposium on Romance Lan-
guages (LSRL XXVIII), University Park, 16-19 April 1998. 1999.
186. MIŠESKA TOMIĆ, Olga and Milorad RADOVANOVIĆ (eds): History and Perspectives of
Language Study. 2000.
187. FRANCO, Jon, Alazne LANDA and Juan MARTÍN (eds): Grammatical Analyses in Basque
and Romance Linguistics. 1999.
188. VanNESS SIMMONS, Richard: Chinese Dialect Classification. A comparative approach to
Harngjou, Old Jintarn, and Common Northern Wu. 1999.
189. NICHOLOV, Nicolas and Ruslan MITKOV (eds): Recent Advances in Natural Language
Processing II. Selected papers from RANLP ’97. 2000.
190. BENMAMOUN, Elabbas (ed.): Perspectives on Arabic Linguistics Vol. XII. Papers from the
Twelfth Annual Symposium on Arabic Linguistics. 1999.
191. SIHLER, Andrew L.: Language Change. An introduction. 2000.
192. ALEXANDROVA, Galina M. and Olga ARNAUDOVA (eds.): The Minimalist Parameter.
Selected papers from the Open Linguistics Forum, Ottawa, 21-23 March 1997. 2001.
193. KLAUSENBURGER, Jurgen: Grammaticalization. Studies in Latin and Romance morphosyntax.
2000.
194. COLEMAN, Julie and Christian J. KAY (eds): Lexicology, Semantics and Lexicography.
Selected papers from the Fourth G. L. Brook Symposium, Manchester, August 1998. 2000.
195. HERRING, Susan C., Pieter van REENEN and Lene SCHØSLER (eds): Textual Parameters in
Older Languages. 2000.
196. HANNAHS, S. J. and Mike DAVENPORT (eds): Issues in Phonological Structure. Papers from
an International Workshop. 1999.
197. COOPMANS, Peter, Martin EVERAERT and Jane GRIMSHAW (eds): Lexical Specification
and Insertion. 2000.
198. NIEMEIER, Susanne and René DIRVEN (eds): Evidence for Linguistic Relativity. 2000.
199. PÜTZ, Martin and Marjolijn H. VERSPOOR (eds): Explorations in Linguistic Relativity. 2000.
200. ANTTILA, Raimo: Greek and Indo-European Etymology in Action. Proto-Indo-European *aǵ-
.2000.
201. DRESSLER, Wolfgang U., Oskar E. PFEIFFER, Markus PÖCHTRAGER and John R.
RENNISON (eds.): Morphological Analysis in Comparison. 2000.
202. LECARME, Jacqueline, Jean LOWENSTAMM and Ur SHLONSKY (eds.): Research in
Afroasiatic Grammar. Papers from the Third conference on Afroasiatic Languages, Sophia
Antipolis, 1996. 2000.
203. NORRICK, Neal R.: Conversational Narrative. Storytelling in everyday talk. 2000.
204. DIRVEN, René, Bruce HAWKINS and Esra SANDIKCIOGLU (eds.): Language and Ideology.
Volume 1: cognitive theoretical approaches. 2001.
205. DIRVEN, René, Roslyn FRANK and Cornelia ILIE (eds.): Language and Ideology. Volume 2:
cognitive descriptive approaches. 2001.
206. FAWCETT, Robin: A Theory of Syntax for Systemic-Functional Linguistics. 2000.
207. SANZ, Montserrat: Events and Predication. A new approach to syntactic processing in English
and Spanish. 2000.
208. ROBINSON, Orrin W.: Whose German? The ach/ich alternation and related phenomena in
‘standard’ and ‘colloquial’. 2001.
209. KING, Ruth: The Lexical Basis of Grammatical Borrowing. A Prince Edward Island French case
study. 2000.
210. DWORKIN, Steven N. and Dieter WANNER (eds.): New Approaches to Old Problems. Issues
in Romance historical linguistics. 2000.
211. ELŠÍK, Viktor and Yaron MATRAS (eds.): Grammatical Relations in Romani. The Noun
Phrase. 2000.
212. REPETTI, Lori (ed.): Phonological Theory and the Dialects of Italy. 2000.
213. SORNICOLA, Rosanna, Erich POPPE and Ariel SHISHA-HALEVY (eds.): Stability, Variation
and Change of Word-Order Patterns over Time. 2000.
214. WEIGAND, Edda and Marcelo DASCAL (eds.): Negotiation and Power in Dialogic
Interaction. 2001.
215. BRINTON, Laurel J.: Historical Linguistics 1999. Selected papers from the 14th International
Conference on Historical Linguistics, Vancouver, 9-13 August 1999. 2001.
216. CAMPS, Joaquim and Caroline R. WILTSHIRE (eds.): Romance Syntax, Semantics and L2
Acquisition. Selected papers from the 30th Linguistic Symposium on Romance Languages,
Gainesville, Florida, February 2000. 2001.
217. WILTSHIRE, Caroline R. and Joaquim CAMPS (eds.): Romance Phonology and Variation.
Selected papers from the 30th Linguistic Symposium on Romance Languages, Gainesville,
Florida, February 2000. n.y.p.
218. BENDJABALLAH, S., W.U. DRESSLER, O. PFEIFFER and M. VOEIKOVA (eds.): Morphol-
ogy 2000. Selected papers from the 9th Morphology Meeting, Vienna, 24–28 February 2000. 2002.
219. ANDERSEN, Henning (ed.): Actualization. Linguistic Change in Progress. 2001.
220. SATTERFIELD, Teresa, Christina TORTORA and Diana CRESTI (eds.): Current Issues
in Romance Languages. Selected papers from the 29th Linguistic Symposium on Romance
Languages (LSRL), Ann Arbor, 8-11 April 1999. 2002.
221. D'HULST, Yves, Johan ROORYCK and Jan SCHROTEN (eds.): Romance Languages and
Linguistic Theory 1999. Selected papers from ‘Going Romance’ 1999, Leiden, 9-11 December.
2001.
222. HERSCHENSOHN, Julia, Enrique MALLÉN and Karen ZAGONA (eds.): Features and
Interfaces in Romance. Essays in honor of Heles Contreras. 2001.
223. FANEGO, Teresa, María José LÓPEZ-COUSO and Javier PÉREZ-GUERRA (eds.): English
Historical Syntax and Morphology. Selected papers from 11 ICEHL, Santiago de Compostela, 7-
11 September 2000. 2002.
224. FANEGO, Teresa, Belén MÉNDEZ-NAYA and Elena SEOANE (eds.): Sounds, Words, Texts
and Change. Selected papers from 11 ICEHL, Santiago de Compostela, 7-11 September 2000.
n.y.p.
225. SHAHIN, Kimary N.: Postvelar Harmony. n.y.p.
226. LEVIN, Saul: Semitic and Indo-European. Volume II: comparative morphology, syntax and
phonetics; with observations on Afro-Asiatic. n.y.p.
227. FAVA, Elisabetta (ed.): Clinical Linguistics. Theory and applications in speech pathology and
therapy. n.y.p.
228. NEVIN, Bruce E. (ed.): The Legacy of Zellig Harris. Language and information into the
21st century. Volume 1: philosophy of science, syntax and semantics. n.y.p.
229. NEVIN, Bruce E. and Stephen JOHNSON (eds.): The Legacy of Zellig Harris. Language
and information into the 21st century. Volume 2: computability of language and computer
applications. n.y.p.
230. PARKINSON, Dilworth B. and Elabbas BENMAMOUN (eds.): Perspectives on Arabic
Linguistics XIII-XIV. Papers from the Thirteenth and Fourteenth Annual Symposia on
Arabice Linguistics. n.y.p.