English in Transition

WDE

G
Topics in English Linguistics
23

Editor

Herman Wekker

Mouton de Gruyter
Berlin · New York
English in Transition
Corpus-based Studies in
Linguistic Variation and Genre Styles

Edited by

Matti Rissanen
Merja Kytö
Kirsi Heikkonen

Mouton de Gruyter
Berlin · New York 1997
M o u t o n de Gruyter (formerly Mouton, The Hague)
is a Division of Walter de Gruyter & Co., Berlin.

© Printed on acid-free paper which falls within the guidelines of the

A N S I to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

English in transition : corpus-based studies in linguistic vari-

ation and genre styles / edited by Matti Rissanen, Merja
Kytö, Kirsi Heikkonen.
p. cm. — (Topics in English linguistics ; 23)
Includes bibliographical references and index.
ISBN 3-11-015632-6 (alk. paper)
1. English language - Early modern, 1500—1700 — Dis-
course analysis. 2. English language — Early modern,
1500—1700 — Variation. 3. English language - Early mod-
ern, 1500-1700 - Style, 4. Literary form. I. Rissanen,
Matti. II. Kytö, Merja. III. Heikkonen, Kirsi, 1962—
IV. Series.
PE881.E54 1997
420M '41 —dc21 97-25407
CIP

Die Deutsche Bibliothek — Cataloging-in-Publication Data

English in transition : corpus based studies in linguistic variation

and genre styles / ed. by Matti Rissanen ... - Berlin ; New
York : M o u t o n de Gruyter, 1997
(Topics in English linguistics ; 23)
ISBN 3-11-015632-6
D B N : 95.102774.3©
SG: 52

© Copyright 1997 by Walter de Gruyter & Co., D-10785 Berlin

All rights reserved, including those of translation into foreign languages. No part of this
book may be reproduced or transmitted in any form or by any means, electronic or mechan-
ical, including photocopy, recording, or any information storage and retrieval system, with-
out permission in writing from the publisher.
Printing: Ratzlow-Druck, Berlin.
Binding: Lüderitz & Bauer, Berlin.
Printed in Germany.
Preface

This book is one of three volumes reporting the results of the project
'English in transition: Change through variation', carried out in the Eng-
lish Department of the University of Helsinki. The first volume, Early
English in the computer age: Explorations through the Helsinki Corpus
(ed. by Matti Rissanen, Merja Kytö and Minna Palander-Collin, Mouton
de Gruyter, 1993) is now followed by two volumes, English in transition:
Corpus-based studies in linguistic variation and genre styles and Gram-
maticalization at work: Studies of long-term developments in English.
Both these volumes approach change in English from the angle of lin-
guistic variation. The articles deal with processes of change in morphol-
ogy, syntax and lexis, and pay special attention to the role played by tex-
tual and discourse factors across the centuries. From the methodological
point of view, diachronic variation analysis and the multi-feature ap-
proach aiming at the identification of co-occurrence patterns in genres are
the main frameworks adopted.
The aim of the present volume is to give new insights into the develop-
ment of some central verb constructions (with be and have), expository
apposition, and genre-specific features of expressions of affect and atti-
tude in text. The Grammaticalizaton at work volume sheds light on the
development of adverbs and indefinite pronouns and on the means of re-
flexivization, in relation to various grammaticalization processes.
All the studies in these volumes are based on the Helsinki Corpus of
English Texts; supplementary material has been drawn from other corpora
and concordances and from primary texts outside the corpora.
The 'English in transition' project was initiated in 1990 as a continu-
ation of an earlier project which produced the Helsinki Corpus. The core
team of both projects has been the same, consisting of the authors and
editors of the volumes. The editors would like to express their special
thanks to all research assistants of the projects and particularly to Arja
Nurmi and Päivi Koivisto-Alanko for their excellent work in producing
these volumes.
vi Preface

We are most grateful to the Academy of Finland for funding our project
for four years. We are indebted to the University of Helsinki for giving
us research premises, and to the English Department for up-to-date tech-
nical facilities, travel grants and other support. Our thanks are due to Mrs
Leena Sadeniemi for giving us expert advice in computer technology and
training us to use programs. Finally, we would like to thank the editors of
Pyfouton de Gruyter for accepting the two volumes now published in their
Topics in English Linguistics series.

Helsinki, June 1996

M.R. M.K. K.H.

Contents

Introduction 1
Matti Rissanen, Matti Kilpiö, Merja Kytö, Anneli Meurman-Solin,
Saara Nevanlinna, Päivi Pahta and Irma Taavitsainen

Be/have + past participle: The choice of the auxiliary with 17

intransitives from Late Middle to Modern English
Merja Kytö

On the forms and functions of the verb be from Old to Modern 87

English
Matti Kilpiö

Re-phrasing in Early English: The use of expository apposition 121

with an explicit marker from 1350 to 1710
Päivi Pahta and Saara Nevanlinna

Genre conventions: Personal affect in fiction and non-fiction 185

in Early Modern English
Irma Taavitsainen

Towards reconstructing a grammar of point of view: Textual 267

roles of adjectives and open-class adverbs in Early Modern
English
Anneli Meurman-Solin

Bibliography 345
Index of subjects 365
List of abbreviations

The following parameter codes included in the Helsinki Corpus appear in

the present volume as such or in abbreviated versions (see also Biblio-
graphy at the end of the volume).

Prototypical text category:

EX = 'expository'
IR = 'instruction religious'
IS = 'instruction secular'
IS/EX = 'instruction secular'/'expository'
NI = 'narration imaginative'
NN = 'narration non-imaginative'
STA = 'statutory'

Text type:

BIA = 'biography, autobiography'

BIBLE = 'Bible'
BIL = 'biography, life of a saint'
BIO = 'biography, other'
COME = 'drama, comedy'
CORO = 'correspondence, non-private'
CORP = 'correspondence, private'
DEPO = 'proceeding, deposition'
DIARY = 'diary'
DOC = 'document'
EDUC = 'educational treatise'
FICT = 'fiction'
GEO = 'geography'
HANDA = 'handbook, astronomy'
HANDM = 'handbook, medicine'
χ Abbreviations

HANDO = 'handbook, other'

HIST = 'history'
HOM = 'homily'
LAW = 'law'
LET PRIV = 'letter, private'
LET NON-PRIV = 'letter, non-private'
MYST = 'drama, mystery play'
NEWT = 'New Testament'
OLDT = O l d Testament'
PHILO = 'philosophy'
PREF = 'preface' or 'epilogue'
RELT = 'religious treatise'
ROM = 'romance'
RULE = 'rule'
SCI A = 'science, astronomy'
SCIM = 'science, medicine'
SCIO = 'science other'
SERM = 'sermon'
TRAV = 'travelogue'
TRI = 'proceeding, trial'

Other:

PROF = 'audience, professional'

NON-PROF = 'audience, non-professional'
INT = 'interaction', 'interactive'
INF = 'informal setting'
FOR = 'formal setting'
Χ, XX = 'unspecified'
Introduction

Matti Rissanen, Matti Kilpiö, Merja Kytö, Anneli Meurman-Solin,

Saara Nevanlinna, Päivi Pahta and Irma Taavitsainen

1. Methodological considerations: linguistics and

philology in interaction

The last two decades have seen a rapid increase and methodological de-
velopment in the studies of variation in language. The basic assumption
in these studies is "orderly heterogeneity", i.e. variability which is not
random but affected by linguistic and extralinguistic factors or constraints
(Weinreich—Labov—Herzog 1968; Samuels 1972; Labov 1994). Lan-
guage can be seen as meaning potential which is realized in choices be-
tween forms and expressions "meaning the same thing" (Halliday 1973:
51; see also Halliday 1978). Within this framework, it has been our aim
to establish how linguistic variation is patterned not only socially, re-
gionally and temporally, but according to genres defined by extralinguis-
tic criteria. When the variationist approach is adapted to discourse studies,
comparisons of text types defined by their linguistic properties become
the key (Schiffrin 1994: 314); the last two chapters of this book extend
the application to historical stylistics with the aim of charting genre styles
and genre conventions.
In diachronic studies, the variationist approach provides us with a good
opportunity to observe the actual process of change. We can trace the
birth and death of variant expressions, but perhaps more interestingly,
their changing frequencies and distributions within a variant field at sub-
sequent periods of time and in various genres, and we can analyse
changes in the intricate mesh of linguistic and extralinguistic factors con-
ditioning the occurrence of these variants. Within this approach our philo-
2 Matti Rissanen et al.

logical training has been a great asset as it emphasizes the context of ex-
pressions and the context of culture in interpreting them.
The importance of extralinguistic factors in the analysis of development
and change has necessitated an intensive study and discussion of genre
typologies from the point of view of the historical study of language; cf.
e.g. Douglas Biber's and Edward Finegan's studies of the 'dimensions'
characterizing texts and offering a new basis for their grouping. In the
structure of text corpora, the labelling of genres provides a general frame-
work for the discussion of text-related changes at different periods of
time. However, the varying importance of conventions or innovative pres-
sures in the evolution of each genre or group of genres may decrease the
usefulness of such classifications, as genres are also internally hetero-
geneous, and the pace and direction of change may be different in indi-
vidual texts representing a particular genre. A comparison of the oc-
currences and frequencies of variant expressions in different texts allows
the reconstruction of the various levels of past expression: written and
speech-based, literary and non-literary, formal and informal, etc. This
method is also necessary for all attempts to describe the relationship of
the standard(s) to regional or social dialects.
Of the five chapters included in this volume, three discuss morpho-
logical, syntactic or lexical questions with special attention paid to varia-
tion relating to text type, dialect etc., while two concentrate on the co-
occurrence patterns of linguistic features in various types of texts and on
questions of genre classification, genre markers and distinguishing fea-
tures. The main focus is on long-span diachrony, mostly from Early Mid-
dle to Modern English, i.e. from the time of the radical reorganization of
the structure of the language to the period of its gradual establishment.
The long time span and the wealth of primary data set specific demands
for the grammatical models used in this volume. The model should make
it possible to compare changing grammatical phenomena across time and,
at the same time, be comprehensive enough to provide researchers with
analytical tools for the very wide range of morphosyntactic issues involv-
ed in classifying linguistic features in genre studies. The model which has
proved most useful in this type of research is a structurally oriented one,
such as A comprehensive grammar of the English language by Quirk et
al. (1985) for Present-day English. This type of grammar provides an ad-
equate basis for analysis at a relatively low level of abstraction, enabling
the researcher to deal even with the more problematic borderline cases.
On the other hand, it is clear that, as a grammar of Present-day English,
Quirk et al. cannot be directly applied to different historical phases of
 Introduction 3

English. In our case, it has provided the basis which the writers of this
book have employed in different ways and to different degrees. We have
made use of various approaches, from traditional grammar to semantic,
pragmatic and textual theories.

2. New openings offered by the Helsinki Corpus and

other computerized material

Scholars working on variationist studies benefit from having access to

computerized collections of texts. Computerized diachronic corpora, in-
creasingly available in international distribution, make it relatively easy to
collect evidence of the occurrence of variant expressions; they also en-
courage the researcher to approach topics that would earlier have meant
an unreasonable amount of time-consuming routine work.
The studies reported in this volume are based on the Helsinki Corpus
of English Texts, the first large computerized corpus to cover the time-
span of several periods in the history of English. With its 1.5 million
words — c. 400 samples of texts dating from the 8th to the 18th century
— it offers reliable indicators concerning the structural and lexical devel-
opments of English for over a millennium. In many cases, the results are
tentative and they must — and fortunately can — be supplemented from
other corpora, concordances and primary texts. In the future, the useful-
ness of the Helsinki Corpus will be further increased by the addition of
word-class tagging and syntactic bracketing to the text samples. 1
Each text or group of related texts of the Helsinki Corpus is equipped
with a set of parameter values containing information on the text and its
author, if known. In Old and Early Middle English, this information is
concentrated mainly on the date and dialect of the text and on a fairly
loose description of the genre. In Late Middle and Early Modern English,
the genre selection is more extensive than in the earlier periods, with
samples from drama texts, private letters, law court records, diaries, prose
fiction, and so forth. 2 In these periods, sociolinguistic information is giv-
en on the authors of the texts (their rank, sex, and age) and, in the case of
letters, on the relationship between the writer and the receiver (intimate
family members are distinguished from more distant addressees; addres-
sees superior to authors are distinguished from those inferior to them).
4 Matti Rissanen et al.

A large and structured corpus such as the Helsinki Corpus, equipped

with textual parameter codings, has made it possible for us to combine ef-
fectively qualitative and quantitative analysis through variation. This ap-
proach has been extensively used in the analysis and interpretation of the
data (cf. McEnery—Wilson 1996: 62-63). Our background knowledge of
texts, also capitalized on in the compilation of the Helsinki Corpus, has
been utilized to place our observations of linguistic phenomena within a
larger framework. We have been able to reanalyse and re-evaluate earlier
genre and text-type classifications of the older periods of English and to
test the validity of the suggestions made concerning the co-occurrence of
linguistic features with various types of text. The parameter coding and
extensive textual basis have also allowed us to observe the role played by
dialectal distribution and the differences between prose and verse texts in
the survey of Old and Middle English variant forms. In the discussion of
Late Middle and Early Modern evidence, emphasis has been given to the
distributions shown by speech-based texts, or texts showing a relatively
high degree of orality. In a few cases, it has been possible to draw con-
clusions from sociolinguistic variables: the genre, degree of interactive-
ness, level of formality, sex, age, rank or education of the author, or the
relationship between the sender and receiver of letters.3 The insights de-
rived from recent trends in sociolinguistics have added a new angle to the
discussion of these factors (cf. e.g. Romaine 1982; Milroy—Milroy 1985;
Milroy 1992). We believe that, among these variables, genre is the most
complex one, as the varying types of genre can be claimed to derive from
a combination of other variables such as topic, audience, setting, text cat-
egory, etc.
The corpus-based framework lends itself to various statistical applica-
tions by which it is possible to verify the reliability of the results ob-
tained. In addition to traditional frequency surveys, a number of more ad-
vanced statistical analyses have been carried out for the purposes of some
studies included in the volume. Thus Merja Kytö in her study on the
be/have variation uses logistic regression analysis to assess the impact of
various factors on the use of the variant forms. This analysis tests out
how statistical models based on various combinations of factors account
for variation in the data and assess the significance of the individual fac-
tors and their interaction. Several statistical methods are combined in
Irma Taavitsainen's chapter on personal affect and genre styles. Factor
analysis is applied first to identify adjoining text types and possible text-
type markers. This method serves to indicate the overall patterns of genre
styles and how they relate to one another. Then t-tests and f-tests are used
 Introduction 5

to evaluate the significance of the features in telling fiction apart from the
adjoining genres. The combination of these methods yields results which
are then assessed in a larger sociohistorical context. In the chapter by An-
neli Meurman-Solin on the concept of point of view in texts, the different
informative value of mean frequencies and percentages is stressed and the
findings are presented by focusing on one feature or factor at a time and,
after a detailed analysis of this kind, by clustering them by both syntactic
and semantic criteria. This adds to the reliability of data and also allows
the mapping of significant correlations in a network-like pattern of di-
mensions that usefully describe focal features in genre styles and text
types.

3. Variation on the level of morphology, syntax and lexis:

the verbs be and have

The studies by Merja Kytö and Matti Kilpiö in this volume focus on the
use and development of the verbs be and have, which have played a vital
role in the shaping of the English verb phrase. The paths of be and have
are parallel, both having occurred as lexical and auxiliary verbs providing
variant expressions in the auxiliary function (e.g. compound tenses with
mutative verbs, and expression of obligation).
Owing to complexities in the development of these verbs, there is no
consensus about the role played by such crucial processes as grammati-
calization. With the verb be, for instance, the status of the verb in pro-
gressive forms has clearly changed from the copula in Old English into
an auxiliary from Middle English onwards. The uses of be and have
grammaticalize at different stages in the history of English, and the two
studies only concentrate on certain key periods in these developments.
The study of the be/have variation with mutative intransitives examines
the process by which have finally supersedes be in present and past per-
fect constructions; the study on be focuses on the forms and functions of
the verb, with an eye on the developments in its functional load. The for-
mer study covers the period from Late Middle to Modern English, the lat-
ter from Old to Modern English. Among the extralinguistic factors includ-
ed in the two studies are chronology, region and dialect, and foreign
influence; moreover, with the be/have variation, such factors as text type,
relationship of the text to spoken language, level of formality, orality and
6 Matti Rissanen et al.

authorial properties are taken into account. Among the linguistic factors
observed in both studies are tense and certain verb constructions (finite/
non-finite forms, -ing constructions). Furthermore, with the study of the
be/have variation, attention is paid to the status of the verb (stative and
mutative; action and process; frequent and rarer verbs), durative, iterative
and conditional contexts, negation, and object-like and other comple-
ments. With be, additional factors include developments in the morphol-
ogy of be, the function of the verb (its use as an auxiliary or a lexical
verb, copular or non-copular), and semantic-syntactic functions of ft-forms
as against non-0-forms (for Old English).
These studies have brought up new evidence pinning down trends of
development in greater detail than found in previous research, thanks to
the combination of quantitative and qualitative corpus linguistics and
philological assessments. With the be/have variation, the use of have is
shown to increase gradually from the Late Middle to the Early Modern
English period, gain in momentum in the late 1700s and supersede be in
the early 1800s. The more powerful extralinguistic factors influencing the
process of change include chronology and text type, and a number of lin-
guistic factors (relationship to tense, aspect, complementation etc.). In the
study dealing with be, the most striking morphological developments are
the rapid disappearance in the Early Middle English period of the co-
existent Old English present tense paradigms (beon/wesan), the diversifi-
cation of forms in Middle English and the subsequent process of simplifi-
cation and regularization as Early Modern English is reached. In the sur-
vey of the main functions of be, the remarkable stability seen in the rela-
tive share of the three main functions of be over the centuries is the most
important finding. Within the auxiliary category, the predominance of the
use of be as a passive auxiliary is the most noticeable feature.
These two studies have shown that the Helsinki Corpus, supplemented
by other diachronic corpora and other standard reference works (LALME,
MED, OED), offers a solid basis for the empirical approach aimed at ac-
counting for variation and change.
 Introduction 7

4. Variation in re-phrasing: expository apposition across

the centuries

The chapter by Päivi Pahta and Saara Nevanlinna forms a bridge between
the studies concentrating on the syntactic and morphological develop-
ments described above and those discussing the characteristics of genres
and text types. It examines characteristics of expository apposition — the
grammatical category connected with re-phrasing. Re-phrasing as a com-
municative phenomenon occurs in both written and spoken media, and in
both planned and unplanned discourse. It can generally be analysed as the
writer's or speaker's attempt to reformulate an utterance in order to
achieve successful communication. On closer inspection, the decision to
re-phrase may be based on various factors, including stylistic and didactic
considerations, the author's assessment of the addressee's ability to pro-
cess given information, or the author's wish to add to the flow of dis-
course by providing more information about the topic of discussion.
This chapter focuses on the development and use of expository apposi-
tion with an explicit marker in Late Middle and Early Modern English.
Apposition is seen as a broad notional category containing both nominal
and non-nominal phrases, clauses and sentences. There are no previous
detailed studies of apposition in this period, and the theoretical framework
adopted as the starting-point in this study is the recent discussion of appo-
sition in Present-day English by Meyer (1987 and 1992), where apposi-
tion is seen as a syntactic, semantic and pragmatic relation. The main em-
phasis in Pahta and Nevanlinna's study is on the semantic characteristics
of appositional constructions and their distribution across different text
types. Attention is also paid to the devices used in linking appositional
units, i.e. explicit markers of expository apposition.
The study shows that the use of expository apposition links up with
some of the most central lexical phenomena of the Middle English and
Early Modern English periods, such as dialectal variation and the adop-
tion and accommodation of loan-words. It also indicates a clear tendency
for some text types to favour the use of appositional constructions in gen-
eral, and certain semantic and syntactic types in particular. This is so
throughout the period, although there is internal variation within most text
8 Matti Rissanen et al.

types. Most of the markers of expository apposition available in Present-

day English were found to exist even in the late medieval period, with
many others which have since gone out of use. In this respect, the results
obtained point to a difference in the use of coordinative apposition, par-
ticularly binomial constructions with the markers and and or.
In the course of the analysis, the use of traditional philological tools
(knowledge of textual background, cultural status of texts, etc.) proved
helpful. With certain limitations, the corpus-based approach offered a
fruitful way of collecting data for the study of appositional constructions.
Considering the open-class nature of apposition as a linguistic phenom-
enon, the results obtained in this study show the way for further work on
the topic.

5. Expressions of personal affect and stance marking:

identifying genre-specific choices

The last two chapters of this volume, by Irma Taavitsainen and Anneli
Meurman-Solin, have a somewhat different problem-setting, but the ap-
proach combining the quantitative and qualitative methods applies here as
well. Their main topic is identifying genre-specific features in the lin-
guistic choices related to participant roles by analysing expressions of
affect and attitude in texts. Taavitsainen discusses the use of personal pro-
nouns, exclamations, direct questions and other expressions of personal
affect, while Meurman-Solin's study focuses on the frequencies and dis-
tributions of adjectives and open-class adverbs as stance markers.
In recent years genres have been looked at from many different per-
spectives. Besides thorough comparative studies of features of individual
genres (such as fiction in Fludernik 1993 and 1996) or a wide range of
genres in a particular time period (such as Renaissance genres in Le-
walski 1986), we find the interdisciplinary approach (for example in Sell
and Verdonk 1994) and the important advances in discourse analysis
(Coulthard 1994) particularly relevant. Taavitsainen and Meurman-Solin
approach the problems of genre studies from a variationist's and socio-
linguist's point of view. Rather than restrict the focus to the dimension of
written genres as against genres reflecting usages more typical of spoken
language, they aim at pointing out clusters of features which position
texts on a number of other dimensions, particularly those which reflect
 Introduction 9

idiolectal or genre-specific characteristics of participant roles, or focus on

genre markers.
Traditional genre labels such as 'history', 'private letter', 'autobio-
graphy' or 'sermon' are used in the majority of recently produced text
corpora. A typology of this kind has also been adopted as a working tool
in the Helsinki Corpus and its supplement, the Helsinki Corpus of Older
Scots (see note 1). The labelling in these corpora is based on extralinguis-
tic factors such as the social and communicative function of texts or their
subject matter. In a number of studies based on the Helsinki Corpus and/
or its supplement of Scots (see the bibliographies in the chapters by Taa-
vitsainen and Meurman-Solin in this volume), medieval and Renaissance
prose genres have been shown to be linguistically relatively heteroge-
neous. This heterogeneity is partly due to the compilers' decision to po-
larize the samples: they have intentionally selected the representatives of
a genre from different stylistic traditions when such stylistic variation has
been established in earlier research (Nevalainen—Raumolin-Brunberg
1989: 99). This should always be taken into consideration when general-
izing from the results.
Patterns of co-occurring features, illustrated in Taavitsainen's and Meur-
man-Solin's studies, provide evidence for a classification of texts into text
types. Each text type may thus comprise texts which represent different
genres; in addition, intertextuality phenomena between genres complicate
the issues. Biber's pioneering work (1988) in corpus-based stylistics with
its multifeature and multidimensional statistical assessments is strictly lin-
guistic; our innovation is the firm philological anchoring, limiting the
comparisons to texts that share common features and that belong to re-
lated genres (Taavitsainen) and recategorizing multifunctional linguistic
features by means of a thorough analysis of their varying syntactic and
semantic properties in different time periods (Meurman-Solin). We have
also profited from other studies which tackle related questions. The two
studies aim at making it applicable to the analysis of early prose texts by
selecting features other than those in Biber's factors, and by introducing
a more detailed semantic subcategorization of a more comprehensive set
of realizations of some features included in his factor analysis, and by de-
veloping the research tasks for different aims. Because of the emphasis on
semantic features, the relevant examples are carefully selected by qualita-
tive reading and analysed in the wider context of running text. Both
studies thus highlight the importance of combining the quantitative ap-
proach of corpus linguistics with a detailed analysis of discourse function
and meaning, central in the philological tradition.
10 Matti Rissanen et al.

To compensate for the low number of texts per genre and subperiod,
the Helsinki Corpus offers relatively long samples of texts, so that statisti-
cally significant amounts of data can be recorded in them. This is also
true of numerous lexical items, which often have an insignificant role in
corpora where the sample size is smaller (cf. Biber 1988, Devitt 1989).
Further study will show whether the texts in the corpus are prototypical
representatives of the different genres, but they can perhaps be claimed to
have an established position in their sociocultural context. The findings in
the two chapters identify basic similarities and differences between texts
and genres, which may be selected as diagnostic features in subsequent
studies based on a larger corpus of texts.
Taavitsainen's study tests the methods provided by corpus linguistics
for a new aim, using structuralist literary criticism and a variationist ap-
proach as her theoretical basis. Her focus is on personal affect which is
an optional component of participant relations. Personal affect offers a
fruitful object of study since it is realized in various ways in texts, and
the co-occurrence patterns of its linguistic features show a great deal of
variation. Personal affect has not been defined in an adequate way in the
literature, and one of the aims of the study is to define its quality in these
texts in more detail.
Several quantifying methods are employed in this study and they are
combined with qualitative analysis. Factor analysis is applied first to re-
veal the co-occurrence patterns of linguistic features marking subjectivity.
It arranges these features in a hierarchical order according to their impor-
tance in reflecting the underlying dimensions of variation. It also indi-
cates textual affinities and points out genres that are close to one another.
Further statistical tests are then applied to a smaller number of genres
with close affinities. The aim is to test how significant the features with
highest factor loadings are as markers between adjoining text types, with
special reference to the corpus parameter of imaginative versus non-
imaginative narration. Qualitative analysis is used to complement quanti-
tative assessments by relating the texts to their sociohistorical background
and by considering the means available in the language for expressing
personal affect at that time. In the conclusions, the quality of personal af-
fect in fiction and the adjoining text types are defined in terms of "surge"
features of personal affect (i.e. expressions, such as interjections and ex-
pletives, conveying intensified personal charge between the participants
of communication), interactive versus egocentric focus, indexical features
of proximity, and space-building modality.
 Introduction 11

In Meurman-Solin's study the hypothesis is that the frequencies and

distributions of adjectives and open-class adverbs representing various se-
mantic categories provide evidence of the relative importance of descrip-
tive features as against the importance of stance marking in texts (cf.
Biber and Finegan 1989). The study discusses the correlation patterns be-
tween these linguistic features and a number of extralinguistic factors
such as genre, text category, degree of interactiveness, level of formality,
the audience or the author's sex. Moreover, it illustrates how the choices
made between adverb and adjective realizations, and, in the case of the
latter, between integrated and fragmented structures, reflect varying de-
grees of informational density, and identify the relative prominence of the
author's voice in the individual texts and in the different genres and text
categories. Methodologically, the study aims at solving problems related
to the traditional ways of grouping texts. It presents an approach in which
a detailed analysis of individual features in each idiolect or text leads to
the accumulation of evidence as regards co-occurrence patterns of these
features; it is the cumulative effect of the carefully analysed patterns that
then serves to identify text types. Ideally, these classifications can be
meaningfully related to extralinguistic variables, so that ultimately the in-
terrelatedness of the varying social functions of texts and the linguistic
expression of participant roles becomes evident.
Finally, as concerns the theoretical framework adopted, the studies pre-
sented in this volume clearly show the power of the two methods, lin-
guistic and philological: accounting for language change in all its com-
plexity becomes a possible — and a rewarding — task when based on the
principles of corpus linguistics, variationist approach, and genre studies.

Notes

1. For information on the Helsinki Corpus, see Rissanen et al. (1993) and
Kytö (1996).
Among the satellite corpora projects closely related to the Helsinki
Corpus is the Helsinki Corpus of Older Scots, compiled by Anneli Meur-
man-Solin and currently in international distribution (see Meurman-Solin
1995). In addition, there are three projects in progress which will result in
new diachronic corpora: the Corpus of Early American English (see Kytö
1993); the Corpus of Early English Correspondence (see Nevalainen—
Raumolin-Brunberg 1996) and the Corpus of Early English Medical Writ-
ing (Taavitsainen—Pahta 1997).
12 Matti Rissanen et al.

The Penn-Helsinki Parsed Corpus of Middle English, compiled by An-

thony Kroch and Ann Taylor, is based on the prose texts drawn from the
Middle English section of the Helsinki Corpus accompanied by a number
of supplementary texts. The texts have been annotated for syntactic analy-
sis. A new version of this corpus, with a more versatile linguistic coding
and further texts, is in preparation. The Brooklyn-Geneva-Amsterdam-Hel-
sinki Parsed Corpus of Old English, based on the Old English section of
the Helsinki Corpus, is another morphologically and syntactically anno-
tated corpus due to appear within the next five years. The team of com-
pilers include Susan Pintzuk, Willem Koopman and others.
Further information on historical corpora of English, either completed or
in preparation, is found in Kytö et al. (1994), Kytö — Rissanen (1996), and
Hickey et al. (1997).
2. The following text types, defined on the basis on extralinguistic and situa-
tional criteria, are represented in the Helsinki Corpus: law, documents,
handbooks, science, homilies, sermons, rules, religious treatises, the Bible,
philosophy, prefaces, history, geography, travelogue, (auto)biography, fic-
tion, romances, depositions, trial records, drama (mystery plays and com-
edies), private and official letters, educational treatises, and diaries. There
are also texts which have not been given any particular text type definition
(e.g. Old and Middle English verse texts).
3. The articles in the present volume do not presuppose detailed knowledge of
the contents and structure of the Helsinki Corpus. However, as always,
some knowledge of the material will make it easier to appreciate the find-
ings presented. For those interested in learning more about the corpus, Ris-
sanen et al. (1993) and Kytö (1996) will offer a good starting-point.
The typographical conventions in the examples cited from the Helsinki
Corpus are explained in Kytö (1996). For convenience, the references to the
Helsinki Corpus source texts are listed in the Bibliography at the end of the
present volume.

References

Biber, Douglas
1988 Variation across speech and writing. Cambridge—New York etc:
Cambridge University Press.
Biber, Douglas—Edward Finegan
1989 "Styles of stance in English: Lexical and grammatical marking of
evidentiality and affect", TEXT 9/1: 93-124.
 Introduction 13

Coulthard, Malcolm (ed.)

1994 Advances in written text analysis. London—New York: Routledge.
Devitt, Amy J.
1989 Standardizing written English. Diffusion in the case of Scotland 1520
-1659. Cambridge—New York etc.: Cambridge University Press.
Fludernik, Monika
1993 The fictions of language and the languages of fiction. London—New
York: Routledge.
1996 Towards a 'natural' narratology. New York: Routledge.
Halliday, Μ. A. K.
1973 Explorations in the functions of language. London: Edward Arnold.
[Reprinted in 1974.]
1978 Language as social semiotic. The social interpretation of language
and meaning. London: Edward Arnold.
Helsinki Corpus
1991 The Helsinki Corpus of English Texts. Helsinki: Department of Eng-
lish, University of Helsinki.
Hickey, Raymond—Merja Kytö—Ian Lancashire—Matti Rissanen (eds.)
1997 Tracing the trail of time. Proceedings from the Second Diachronic
Corpora Workshop. Amsterdam and Atlanta, GA: Rodopi.
Kytö, Merja
1993 "Early American English", in: Matti Rissanen—Merja Kytö—Minna
Palander-Collin (eds.), 83-91.
Kytö, Merja—Matti Rissanen—Susan Wright (eds.)
1994 Corpora across the centuries: Proceedings of the First International
Colloquium on English Diachronic Corpora, St Catherine's College
Cambridge, 25-27 March 1993. Amsterdam—Atlanta, GA: Rodopi.
Kytö, Merja (comp.)
1996 Manual to the diachronic part of the Helsinki Corpus of English
Texts: Coding conventions and lists of source texts. (Third edition.)
Helsinki: Department of English, University of Helsinki.
Kytö, Merja—Matti Rissanen
1996 "English historical corpora: Report on developments in 1995", ICAME
Journal 20: 117-132.
Labov, William
1994 Principles of linguistic change. Vol. 1. (Language in Society 20.)
Oxford: Blackwell.
Lewalski, Barbara Kiefer (ed.).
1986 Renaissance genres: Essays on theory, history and interpretation.
Cambridge, Mass.—London, England: Harvard University Press.
14 Matti Rissanen et al.

McEnery, Tony—Andrew Wilson

1996 Corpus linguistics. (Edinburgh Textbooks in Empirical Linguistics.)
Edinburgh: Edinburgh University Press.
Meurman-Solin, Anneli
1995 "A new tool: The Helsinki Corpus of Older Scots (1450-1700)",
/CAME Journal 19: 49-62.
Meyer, Charles F.
1987 "Apposition in English", Journal of English Linguistics 20/1: 101-
121.
1992 Apposition in contemporary English. Cambridge: Cambridge Univer-
sity Press.
Milroy, James
1992 Linguistic variation and change. On the historical sociolinguistics of
English. (Language in Society 19.) Oxford—Cambridge, Mass.:
Blackwelk
Milroy James—Lesley Milroy
1985 "Linguistic change, social network and speaker innovation", Journal
of Linguistics 21: 339-384.
Nevalainen, Terttu—Helena Raumolin-Brunberg
1989 "A corpus of Early Modern Standard English in a socio-historical
perspective", Neuphilologische Mitteilungen 90/1: 67-111.
Nevalainen, Terttu—Helena Raumolin-Brunberg (eds.)
1996 Sociolinguistics and language history. Studies based on the Corpus
of Early English Correspondence. Amsterdam—Atlanta, GA: Rodopi.
Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik
1985 A comprehensive grammar of the English language. London—New
York: Longman.
Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.)
1993 Early English in the computer age: Explorations through the Helsinki
Corpus (Topics in English Linguistics 11.) Berlin—New York: Mou-
ton de Gruyter.
Romaine, Suzanne
1982 Socio-historical linguistics, its status and methodology. (Cambridge
Studies in Linguistics 34.) Cambridge—London etc.: Cambridge Uni-
versity Press.
Samuels, M. L.
1972 Linguistic evolution, with special reference to English. (Cambridge
Studies in Linguistics 5.) Cambridge: Cambridge University Press.
 Introduction 15

Schiffrin, Deborah
1994 Approaches to discourse. Oxford—Cambridge, Mass.: Blackwell.
Sell, Roger D.—Peter Verdonk (eds.)
1994 Literature and the new interdisciplinary, poetics, linguistics, history.
Amsterdam—Atlanta, GA: Rodopi.
Taavitsainen, Irma—Päivi Pahta
1997 "Corpus of Early English Medical Writing 1375-1750", ICAME Jour-
nal 21: 71-78.
Weinreich, Uriel—William Labov—Marvin I. Herzog
1968 "Empirical foundations for a theory of language change", in: W. P.
Lehmann—Yakov Malkiel (eds.), Directions for historical linguistics:
A symposium. Austin—London: University of Texas Press, 95-195.
Be/have + past participle: The choice of
the auxiliary with intransitives from
Late Middle to Modern English

Merja Kytö

1. The be/have variation in the history of English

In many languages the choice of the auxiliaries be and have (and their
equivalents) has been subject to variation in perfective constructions with
intransitives indicating "transition" or "change" (cf. they are arrived ver-
sus they have arrived). In some Germanic and Romance languages the
development has resulted in the generalization of one construction (e.g.
the almost total dominance of "be" in Present-day Danish and "have" in
Spanish and Portuguese); in others both variant forms occur in certain
grammatically, stylistically or regionally restricted contexts (as, e.g., in
Present-day Swedish, German, French and Italian).1 This cross-linguistic
variation reflects the differences in the systemic realizations of the dis-
tinction between state (favouring be) and action/process (favouring have).
Over the successive stages of development, various linguistic and extra-
linguistic factors have influenced the choice of one form or the other.
In the Old English period the be/have + past participle construction de-
noted "state" in intransitive and transitive uses (cf. hie wceron gecumene;
hie hcefdon hine gebundenrte). The past participle originally functioned as
an adjective and was sometimes inflected; it could also be preceded by an
object complement (Mitchell 1985:1, §709). However, the grammatical
concord was gradually lost and the past participle was placed immediately
after the auxiliary. Have originally occurred with transitive verbs only,
but early on came to be used with intransitives, too. In Early Middle Eng-
lish be prevailed with mutative verbs, but have started gaining ground
slowly in uses with the emphasis on "action" and the notion of per-
18 Merja Kytö

fectivity (Ryden—Brorström 1987: 16-18). Signs of the rise of have are

apparent from around the 1400s on (for a recent corpus-based study, see
Elsness 1989: 100; 1991: 276-283).
The development of be and have perfects show features attributed to
the process of grammaticalization, but scholars disagree about the exact
chronology and nature of the process. A recent discussion on the topic
can be found in Denison 1993: 340—368 (who uses the term "grammati-
cisation"). With have perfect the relevant factors have been the loss of in-
flections in the participle, the word order, and certain VP types (accord-
ing to the semantics of OE habban and the valency of the main verb).
The process of grammaticalization would have reached a stage of fulfil-
ment "when the have perfect became available for any lexical verb which
did not conjugate with be", i.e. when it became an auxiliary verb, the
suggested (but debatable) date for this being the late Old English period
(Denison 1993: 352). The other possible stages of grammaticalization in-
clude the point when the construction became a tense equivalent (pro-
bably in late Old English); when it had developed its present-day meaning
and superseded be (probably in the seventeenth century); when it became
used with all non-auxiliary verbs (in late Modern English) (Denison,
1993: 352). There is, similarly, disagreement over the grammaticalization
process of the be perfect (for a summary, see Denison 1993: 360—361);
the factors regarded as having influenced the process include, e.g., the in-
creasing use of be as the auxiliary of the passive and the relatively light
functional load of have, the possibility of neutralizing the present tense
third-person singular forms into the clitic 's, and the prescriptions of nor-
mative grammarians.
In this study the emphasis is on the variational method, which means
excluding from analysis the contexts in which there is no choice of one
variant form or the other, e.g. the uses with transitive verbs, which have
always been associated with the have perfect. However, many of the fac-
tors pointed out in previous studies have been of great help when specify-
ing the scope of the study and the distributional factors considered.2

2. The scope of the study

This study focuses on the development of the be/have paradigm in the

history of English from the Late Middle English period up until modern
 Be/have + past participle 19

times. The reasons for setting Late Middle English as a starting point for
observing the development are two-fold. Firstly, as pointed out above, a
change starts to take place in the use of the auxiliaries in the case of the
mutatives in the fifteenth century, a justification for including the data
from the Late Middle English period in the analysis. Secondly, the early
stages of the increasing use of have coincide with the crucial transition
from the Middle to the Early Modern English period. As for the time-
span covered, pursuing the development up until modern times is of great
interest, as it was only in the early 19th century that have finally super-
seded be (Ryden—Brorström 1987). Be is, moreover, still available with
some verbs in present-day received usage (they are/have gone, the sun
is/has set; Ryden—Brorström 1987: 211) and in some dialects (see, e.g.
Kallen 1989: 18-23; Melchers 1992; Filppula 1994).
Though the be/have variation has been dealt with in a good number of
studies, many of the previous works have largely been based on idiolectal
corpora of single authors or restricted to cover only certain periods or a
few text types. The present study, based on material that covers a period
of some 650 years and representative of a variety of authors and text
types, should provide new evidence on the role of the linguistic and extra-
linguistic factors in the process of change.
The study will be carried out within the framework of socio-historical
variation analysis (see the Introduction in this volume). The distribution
patterns of the variant forms will be followed across the successive pe-
riods of time distinguished for the corpora studied (see below). Conclu-
sions will be based on simple frequency tables and results yielded by the
logistic regression analysis used to assess the combined effects of the fac-
tors and their interaction (see section 9 below).

3. The sources for data

In variational terms, the be/have paradigm presents a two-term notional

case, involving various aspects of syntactic, semantic and lexical change.
For obvious reasons, this paradigm is a particularly rewarding topic with-
in the framework of computer-assisted corpus analysis. Relevant examples
with the construction are, if not overwhelming in number, still frequent
enough for a thorough-going analysis; they are also relatively easy to re-
trieve from machine-readable material on the basis of lexical forms. As
20 Merja Kytö

there is no single corpus available at the moment that would cover the
time-span included in this study, several diachronic corpora will need to
be consulted. The differences in the size and structure of the corpora will
be taken into consideration when weighing the evidence. The corpora that
were used as sources for data are introduced below.

The Helsinki Corpus of English Texts (HC; Late ME, EModE)

The Late Middle English data and a good deal of the Early Modern Eng-
lish data have been drawn from the Helsinki Corpus (see Rissanen—
Kytö—Palander-Collin 1993; Kytö 1996); the subperiods included in the
Helsinki Corpus material are ME3 (1350-1420) and ME4 (1420-1500),
and EModE 1 (1500-1570), EModE2 (1570-1640) and EModE3 (1640-
1710). All types of texts represented in the corpus will be considered, but
priority will be given to those that have counterparts in the other corpora
studied.

The Century of Prose Corpus (COPC)

As regards the rise of have, the Century of Prose Corpus covers a crucial
period of development, extending from 1680 to 1780. Part A of the cor-
pus includes extracts from works of 20 major prose authors of the period
(three selections of 5000 words representing each author), while Part Β
consists of writings (sample length 2000 words) drawn from the pens of
one hundred authors writing as journalists, scholars, men of letters and so
forth (Milic 1990: 27-29; 1994: 70). Text type definitions are given by
the compiler for texts included in Part Β only. The first three decades of
writing in the Century of Prose Corpus (1680-1710) coincide with the
last three decades of the Early Modern English section of the Helsinki
Corpus. The whole century covered also coincides with the early decades
represented by the ARCHER Corpus (1650-1990, see below). Contrary to
the conventions adopted for the other corpora included in this study, the
spelling of the Century of Prose Corpus has been modernized.

The ARCHER Corpus

This study owes a great deal to the compilers of the ARCHER Corpus, or
A Representative Corpus of Historical English Registers, for access to rel-
evant data drawn from parts of the pilot version of the corpus. When
complete, this corpus will include c. 1,000 texts and c. 1.7 million words
 Be/have + past participle 21

(sample size at least 2,000 words). The aim of the corpus is "to enable
the analysis of historical change in the range of written and speech-based
registers of English from 1650 to the present" (Biber et al. 1994a: 3; see
also Biber et al. 1994b). Out of the text types found in the corpus, six
have been included in the present study: fiction prose, drama, journals,
letters, science and sermons. The six text types are of special interest for
this study, as they all bear affinities to those found in the Helsinki Corpus
and parts of the Century of Prose Corpus. Texts by British and American
authors have been considered; while the British authors are represented
more or less evenly over the subperiods distinguished, the texts by Ameri-
can authors are all from three subperiods, 1750-1799, 1850-1899 and
1950-1990 (all scientific writings are by British authors). The corpus is
being tagged for grammatical and functional categories; for the purposes
of the present study, the data was drawn from the untagged version on the
basis of lexical forms. 3

Summary of the corpora studied

The main characteristics of the size and the structure of the corpora in-
cluded in the study are given in Table 1.

Table 1. The structure and size of the (sub)corpora studied.

Helsinki Corpus Text type Words Total

ME3 ( 1 3 5 0 - 1 4 2 0 ) fiction 14,300

diary
private letter
official letter 5,000
drama = mystery play
science 3,600
sermon 18,900
document 13,900
handbook 19,200
philosophy 12,600
homily 7,300
other 94,900

Total 189,700
22 Merja Kytö

Table 1. Continued.

Helsinki Corpus Text type Words Total

ME4 (1420-1500) fiction 8,800

diary -

private letter 19,500

official letter 3,200
drama = mystery play 20,100
science 6,400
sermon 25,000
document 10,500
handbook 20,100
philosophy -

homily -

other 100,300

Total 213,900

Total Late ME 403,600

EModEl fiction 11,600

(1500-1570) diary 13,100
private letter 10,600
official letter 6,300
comedy 10,600
science 12,900
sermon 9,500
document
handbook 10,000
philosophy 9,900
other 95,600

Total 190,100
 Be/have + past participle 23

Table 1. Continued.

Helsinki Corpus Text type Words Total

EModE2 fiction 12,500

(1570-1640) diary 12,500
private letter 11,600
official letter 5,700
comedy 11,800
science 13,000
sermon 10,300
document -

handbook 12,300
philosophy 6,900
other 93,200

Total 189,800

EModE3 fiction 12,000

(1640-1710) diary 11,200
private letter 13,100
official letter 5,900
comedy 12,700
science 11,300
sermon 12,500
document -

handbook 11,400
philosophy 8,800
other 72,100

Total 171,000

Total EModE 550,900

24 Merja Kytö

Table 1. Continued.

COPC Text type Words Total

Part A expository prose 300,000

(1680-1780) by 20 authors
(15,000 words each)

Total 300,000

Part Β fiction 20,000

(1680-1780) letter 20,000
science 20,000
biography 20,000
periodical 20,000
education 20,000
essay 20,000
history 20,000
polemics 20,000
travel 20,000

Total 200,000

Total COPC 500,000

ARCHER Corpus Text type Words Total

ARCHER 1 fiction 29,500

(1650-1700) journal 21,900
letter 14,900
drama 32,600
science 20,900
sermon 11,400

Total 131,200

Total ARCHER 1 131,200

 Be/have + past participle 25

Table 1. Continued.

ARCHER Corpus Text type Words Total

ARCHER2a fiction 44,000

(1700-1750) journal 22,000
letter 20,700
drama 24,700
science 21,400
sermon 6,500

Total 139,300

ARCHER2b fiction 86,200

(1750-1800) journal 45,300
letter 32,200
drama 44,400
science 20,900
sermon 26,900

Total 255,900

Total ARCHER2 395,200

ARCHER3a fiction 53,900

(1800-1850) journal 22,800
letter 17,200
drama 34,400
science 21,400
sermon 4,600

Total 154,300
26 Merja Kytö

Table 1. Continued.

ARCHER Corpus Text type Words Total

ARCHER3b fiction 79,700

(1850-1900) journal 46,600
letter 28,300
drama 77,200
science 22,400
sermon 29,100

Total 283,300

Total ARCHER3 437,600

ARCHER4a fiction 12,400

(1900-1950) journal 22,700
letter 15,200
drama 35,700
science 22,500
sermon 4,300

Total 112,800

ARCHER4b fiction 68,100

(1950-1990) journal 45,100
letter 32,100
drama 68,400
science 22,900
sermon 28,000

Total 264,600

Total ARCHER4 377,400

Total ARCHER 1 341,400

TOTAL HC+COPC+ARCHER 2 795,900

 Be/have + past participle 27

4. On the criteria of inclusion

Selecting examples for a study of the be/have variation with intransitives

poses some problems. There are two main questions related to (1) the
type and history of the main verb and (2) the rich functional range of the
auxiliary be. Further selectional criteria to be considered include the role
of object-like complements (e.g. nouns of extent or measure), and some
relatively infrequent constructions in which the variational role of the
auxiliaries is blurred.

The type of the verb

The main verbs are an open-ended class, characterized by a number of se-

mantic features not always too precise to determine. The studies on the
history of the be/have variation traditionally deal with the use of the more
frequent verbs indicating motion (e.g. arrive, come, go, pass, ride), pro-
cess or change (e.g. alter, change, improve, turn), happening (e.g. befall,
chance, hap(pen)), appearing or originating (e.g. appear, arise, become,
begin) and finishing or disappearing (e.g. cease, decay, decline, expire,
die).4 The inclusion of the less frequent verbs (e.g. pirouette, penetrate)
has been the researcher's decision.
Most intransitive verbs considered in previous studies have here been
included for preliminary analysis (for some individual verbs excluded, see
below); along the lines of variational method, some categories dominated
by one or other variant form have been excluded from further analysis on
the basis of the distribution patterns obtained. Similarly, as seems sensible
within the framework of variation analysis, some verbs (e.g. follow, mis-
go), which on the basis of the previous studies have been used with the
form have only, have been excluded altogether (for a provisional list of
verbs and their be/have history, see Visser 1952: 661-663).
Some verbs are frequently indeterminate in nature as regards the tran-
sitive versus intransitive dimension. For this reason, such verbs as agree,
assemble, end, finish, gather, meet, marry, wed and die {be dead) have
been excluded from the analysis.
In some studies attention has also been paid to non-mutative or stative
verbs (e.g. cling, lie, rest, stay; see Visser 1963-1973: 2044-2084). This
category has also been considered at the first stages of the present study.
28 Merja Kytö

The functional range of the verb be

The use of be as a copula as well as a perfective and a passive marker

renders the functional load of the verb heavy (Ryden—Brorström 1987:
24; Elsness 1991: 266-268). With some verbs it is not always possible to
distinguish between the copula and the perfective uses (cf. he is changed
'he is different' versus he is changed 'he has become different'); see ex-
ample (l). 5

(1) The cannon, however, did not stand proof, and the Indians, who
made a close attack, were beaten off and the garrison relieved.
The fort is now totally decayed, and Captain Zane, the only in-
habitant at or near the place, makes use of it for firewood.
(ARCHER2b/Journal/Peter Muhlenberg)6

Nor can one always distinguish between active (perfective) and passive
(present/past tense) uses with verbs which can appear in both transitive
and intransitive contexts {he is changed 'he has become different' versus
he is changed 'he has been made different'); for a discussion, see Ry-
den—Brorström (1987: 24), who for this reason exclude an example such
as 'My prospect of getting to London this spring is rather darkened*
(Wordsworth II: 772, 1836) from their data; for further points, see also
Kakietek 1987 [1976]: 310-311); for a corpus example, see (2).

(2) .. and such is their methode, that rests not so much vppon eui-
dence of truth prooued by arguments, authorities, similitudes,
examples; as vpon particular confutations and solutions of
euerie scruple, cauillation & objection: ... so that the Fable and
fiction of Scylla seemeth to be a liuely Image of this kinde of
Philosophie or knowledge, which was transformed into a come-
ly Virgine for the vpper parts; (Bacon, Advancement of Learn-
ing 20V)

Ambiguous instances such as (1) and (2) have been excluded from this
study; the basic criterion of inclusion has been that in its context the con-
struction conveys the notion of perfectivity.7
All instances of dynamic intransitives (e.g. go, come, rise, arrive) have
been included, even though the construction in some examples may come
closer to a stative than a perfective meaning (cf. they are/have gone, the
sun is/has set).8
 Be/have + past participle 29

Object-like complements

Though this study is primarily concerned with intransitive uses as in (3a),

there are verbs that can take complements in the form of nouns of extent
or measure as in (3 b).9

(3) a. My godfathyr has be syke byt he ys whell mendyd, thankyd be

God. [Thys same] day my Loord ys comyn to London to aske
the Kyng leue to go to the Rodys for he ys sent for.
(Richard Cely, The Cely Letters 107)

b. ... and when mattens vhos done thay whente to a kynnyswho-

man off the 3ewnge genttyllwhomane; and I sent to them a
pottell of whyte romnay, and thay toke hyt thankefully, for thay
had cwm a myle a fote that mornyng; (Richard Cely, The Cely
Letters 152)

With verbs there is variation in whether prepositions or prepositional

phrases are used to link the destination or other complement with the
verb; cf. (4a) with enter + into + object pronoun and (4b) with enter +
object pronoun.10

(4) a. They have not perhaps received precisely what they expected
when their Christian life began, for the kingdom of heaven can-
not be really known until a man has entered into it;
(ARCHER3b/Sermon/Robert Dale)

b. Feb. 26th The works of Ciudad Rodrigo having been completely

put in order, and a garrison of Spaniards having entered it, the
army was ordered to proceed towards Estremadura.
(ARCHER3a/Journal/George Simmons)

In the interest of the emerging individual verb profiles, instances such as

(3b) will be included in the analysis. Out of the examples illustrated in
(4b), two frequent verbs, enter and pass, have been sampled systematical-
ly for full coverage; as for the other verbs, some of them presenting am-
biguous borderline cases between transitive and intransitive uses, the in-
stances with object-like complements have been omitted from the analy-
sis.
30 Merja Kytö

Special constructions

The construction have been + past participle was first attested in the 14th
century and disappeared from usage after the 1850s (for discussion and
references, see Ryden—Brorström 1987: 24-26; Denison 1993: 361, 363,
with reference to the "double perfect" construction). This category would,
strictly speaking, have presented a third variant form to compete with be
and have, but as no more than a score of examples were found (an in-
stance given in (5)), the category has been excluded from the statistics.

(5) Ay, at home, you're as sure of finding an old Mistress, as a

Creditor that expects you to pay him an old Debt, in good Hu-
mor too, I warrant; I was afraid, Madam, you had not been
come home yet. (ARCHER2a/Drama/Catharine Trotter)

Another marginal category is presented by the instances of coordinated

past participles (see example (6)); as the choice of the verb may have de-
pended on verbs other than those included in the study, these instances
have been omitted from the analysis.

(6) For Sr. Allmighty God I take to my record, I have not meant,
intended, or gone about, ne also have willed mine Officers, to
do any thing concerning the said Suppressions, but under such
forme & manner, as is & hath largely been to the full satisfac-
tion, recompence, & joyous contentation of any person which
hath had, or could pretend to have right or interest in the same
... (Wolsey, Original Letters (Illustrative of English History)
(1846) II 20)

Similarly, the examples with the reduced form 's, which could stand ei-
ther for is or has, are excluded from the analysis; see (7); the same holds
for the examples with the form 'd in contexts where it could stand either
for had or should/would.

(7) Every body, no body, I can't tell who; such a Mixture of the
pert and formal! but the most conspicuous Fop was Beau, what
d'you call him, with the fine lac'd Liveries? he's so lately come
into the Town, that I don't know his Name yet, ...
(ARCHER2 a/Dram a/John Mottley)
 Be/have + past participle 31

5. Distributional factors

5.1. The main factors considered

Over the centuries, the distinction between state/result (indicated by be)

and action (indicated by have) seems to have been one of the main distri-
butional factors influencing the choice of the auxiliary.11 The other fac-
tors, extralinguistic and linguistic, taken into consideration in the present
study are the following:

Extralinguistic factors·.
- time
- region (British versus American English in 1750-1990)
- text type
- the relationship of the text to spoken language (speech-based/
written/script)
- the level of formality (formal/informal)
- oral versus non-oral genres
- author properties (men/women writers)
- for foreign influence, see the section on main verbs and loan
words, below

Linguistic factors:
- stative versus mutative verbs (or stand, live, etc. versus come,
arrive, run, grow, become, etc.)
- action versus process verbs (or come, arrive, etc. versus grow,
become, etc.)
- frequent verbs (e.g. come and go) versus other less frequent
verbs; individual verb profiles
- past perfect
- perfect infinitive
- -ing constructions
- durative and iterative contexts
- optative / conditional contexts
- negation
- object-like and other complements
- loan verbs
32 Merja Kytö

After introducing the overall distributions of the data, I will discuss the
role of the two fundamental factors, Stative versus mutative and action
versus process uses. Next, the influence of a number of major extra-lin-
guistic factors will be dealt with, followed by a discussion of the role of
further linguistic factors. As many of the linguistic factors considered
may be effective simultaneously, attention will also be paid to possible
interaction and combined effects of factors. This will be done by discus-
sing the distribution patterns obtained and, more systematically, by apply-
ing the principles of logistic regression analysis to the data.

5.2. Overall distributions of variant forms

The data included in the analysis total 2868 examples. Out of the 194
different verbs, come and go are the most frequent (with 512 and 445
examples, respectively; see Appendix 1). Ten verbs are represented by
more than 30 examples, i.e. pass (178), become (165), fall (131), grow
(90), arrive (62), get (63), return (56), enter (43), run (39), depart (32).
Of the 194 different verbs, 71 are represented by one occurrence only.
The overall figures point, as might be expected, to a gradual rise in the
use of have over the centuries.12 While be still dominates in the Late
Middle English and Early Modern English periods, during the eighteenth
century have is already used in more than 50% of the instances; during
the nineteenth century have supersedes be (see Table 2; in the interest of
chronology, the list is organized so that the results obtained for the
Century of Prose Corpus (1680-1780) follow the first subperiod of the
ARCHER Corpus (1650-1700)). The percentages are calculated from the
total of examples found with be or have within a subperiod. With two-
way tables the chi-square test is used to test the significance of the effect
of the time factor (or period division) on the distributions obtained; the
chi-square figures are nearly significant when ρ < 0.05, significant when
ρ < 0.01 and highly significant when ρ < 0.001, e.g. in Table 2, the influ-
ence of the time factor on the distributions of the variant forms is highly
significant.13
The breakdown figures point to the post-1750s as the final turn-over pe-
riod in the history of the paradigm (see Table 3).
The results are, by and large, convergent with the trend of development
presented by Ryden—Brorström (1987) in their study of be/have variation
in eighteenth-century letters and drama. However, the figures obtained for
have are somewhat higher than expected on the basis of previous studies:
 Be/have + past participle 33

Table 2. Be/have variation with intransitives in 1350-1990 (the percentages are

calculated from the total of examples found with be or have).

be have Total

Late ME (1350-1500) 254 106 (29%) 360

EModE (1500-1710) 363 188 (34%) 551
ARCHER1 (1650-1700) 95 62 (39%) 157
COPC (1680-1780) 135 273 (67%) 408
ARCHER2 (1700-1800) 193 250 (56%) 443
ARCHER3 (1800-1900) 79 447 (85%) 526
ARCHER4 (1900-1990) 27 396 (94%) 423

Total 1146 1722 (60%) 2868

Chi-square = 668.0, df = 6, Ρ < 0.001

Table 3. Be/have variation with intransitives (per subperiods).

be have Total

ME3 (1350-1420) 134 50 (27%) 184

ME4 (1420-1500) 120 56 (32%) 176
EModE 1 (1500-1570) 105 49 (32%) 154
EModE2 (1570-1640) 152 74 (33%) 226
EModE3 (1640-1710) 106 65 (38%) 171
ARCHER 1 (1650-1700) 95 61 (39%) 156
COPC (1680-1780) 135 273 (67%) 408
ARCHER2a (1700-1750) 80 71 (47%) 151
ARCHER2b (1750-1800) 113 180 (61%) 293
ARCHER3a (1800-1850) 40 149 (79%) 189
ARCHER3b (1850-1900) 39 298 (88%) 337
ARCHER4a (1900-1950) 10 95 (90%) 105
ARCHER4b (1950-1990) 17 301 (95%) 318

Total 1146 1722 (60%) 2868

Chi-square = 685.0, df = 12, Ρ < 0.001

34 Merja Kytö

the overall ratio of have has been estimated at some 10% around 1600,
and at 20% around 1700 (Ryden—Brorström 1987: 200; Ryden 1991:
346-347). The varied textual background of the material analysed (see
below) could partly explain the difference. There may also have been dif-
ferences in the way of applying the criteria of exclusion when selecting
examples. As pointed out above, the examples with be are particularly
problematic; the instances with have are much easier to deal with. As my
principles for excluding the ambiguous examples with be were rather
strict, this may be reflected in the figures obtained for have.

6. Two main axes: stative versus mutative and action

versus process

6.1. Stative versus mutative verbs

Stative verbs (lie, rest, stand, stay, etc.) are known to have favoured the
use of have. This holds for the data included in this study as well (see
Table 4; percentages are given for the better represented categories of
have only):

Table 4. Stative verbs.

be have Total

Late ME (1350-1500) 2 22 (92%) 24

EModE (1500-1710) 2 35 (95%) 37
ARCHER 1 (1650-1700) 0 6 6
COPC (1680-1780) 0 11 11
ARCHER2 (1700-1800) 0 9 9
ARCHER3 (1800-1900) 0 8 8
ARCHER4 (1900-1990) 0 11 11

Total 4 102 (96%) 106

 Be/have + past participle 35

As the 106 examples attested with stative verbs represent only some four
percent of the data and are distributed fairly evenly over the subperiods,
the figures obtained for the mutative verbs (come, arrive, run, etc.) come
close to the overall figures given in Tables 2 and 3 (see Table 5).

Table 5. Mutative verbs.

be have Total

ME3 (1350-1420) 132 38 (22%) 170

ME4 (1420-1500) 120 46 (28%) 166
EModEl (1500-1570) 104 42 (29%) 146
EModE2 (1570-1640) 151 54 (26%) 205
EModE3 (1640-1710) 106 57 (35%) 163
ARCHER 1 (1650-1700) 95 55 (37%) 150
COPC (1680-1780) 135 262 (66%) 397
ARCHER2a (1700-1750) 80 65 (45%) 145
ARCHER2b (1750-1800) 113 177 (61%) 290
ARCHER3a (1800-1850) 40 147 (79%) 187
ARCHER3b (1850-1900) 39 292 (88%) 331
ARCHER4a (1900-1950) 10 94 (90%) 104
ARCHER4b (1950-1990) 17 291 (94%) 308

Total 1142 1620 (59%) 2762

Chi-square = 744.7, df = 12, Ρ < 0.001

Within the group of mutative verbs, a number of verbs came to be used

with have at an early stage. These can perhaps be best described by the
absence of the semantic feature that would assign them to verbs indicat-
ing motion or process. In Visser's classification of intransitive verbs these
verbs fall in the groups of verbs indicating appearing, originating (e.g. ap-
pear, begin), or finishing, ceasing, disappearing (e.g. cease, expire, do,
perish), or happening, befalling and like concepts (e.g. befall, hap(pen)
chance·, Visser 1963-1973: 2044-2084). The figures obtained for the use
of be and have with these verbs are given in Table 6; the verbs included
in the figures are appear, befall, begin, cease, chance, disappear, do, ex-
pire, fail, hap, happen, occur, perish, start, stop, vanish.
36 Merja Kytö

Table 6. Be versus have with appear, befall, begin, cease, chance, disappear,
do, expire, fail, hap, happen, occur, perish, start, stop, and vanish.

be have Total

Late ME (1350-1500) 5 6 (55%) 11

EModE (1500-1710) 8 22 (73%) 30
ARCHER1 (1650-1700) 1 7 (88%) 8
COPC (1680-1780) 1 70 (99%) 71
ARCHER2 (1700-1800) 1 46 (98%) 47
ARCHER3 (1800-1900) 0 63 (100%) 63
ARCHER4 (1900-1990) 2 64 (97%) 66

Total 18 278 (94%) 296

As the use of have with stative verbs and verbs indicating appearing, fin-
ishing, happening and like notions comes close to a knock-out feature,
they will be excluded from further analysis. The remaining sections of the
study will thus focus on the choice of be and have in the 2466 examples
with mutative verbs expressing action or process.

6.2. Action versus process uses

The fundamental differences between the use of be and have with verbs
indicating action (typically motion as in, e.g. arrive, return, enter) or
process (typically change of state as in, e.g. grow, become, wax) have
been pointed out in previous studies. 14 The main difference is that, by
and large, the process verbs are slower to adopt have than the motion
verbs. There are, of course, some verbs, e.g. turn and get, which appear
to have a double function, indicating both motion and process; there are
also some phrase-like uses such as fall ill/sick, go mad/crazy. However,
one or the other function seems to dominate the meaning profile of an in-
dividual verb, and instances of indeterminate meaning are relatively rare
(for a discussion, see the individual verb profiles below). The distribu-
tions of the variant forms with action and process uses are given in Table
7; when coding the examples, the "use" rather than the "main meaning"
of the verb has been followed, i.e. (8a) has been coded as an instance of
action use and (8b) as that of a process use.
 Be/have + past participle 37

(8) a. He was therefore no sooner got to his apartment, but he sent the
royal veil to Imoinda; that is, the ceremony of invitation: (Behn,
Oroonoko 158)

b. The church might have got tired of him: ... he forestalled any
disagreeable circumstances which might have attended that
event, by getting tired of the church: (ARCHER3b/Fiction/John
Cooke)

In discussion and tables, the terms action/process verbs/uses will be used

indiscriminately to cover the coding principle adopted.

Table 7. Action versus process verbs.

Action Process

be have Total be have Total

ME3 (1350-1420) 102 35 (26%) 137 30 1 (3%) 31

ME4 (1420-1500) 101 41 (29%) 142 14 1 (7%) 15
EModEl (1500-1570) 91 34 (27%) 125 12 4 (25%) 16
EModE2 (1570-1640) 127 43 (25%) 170 20 4 (17%) 24
EModE3 (1640-1710) 80 40 (33%) 120 23 6 (21%) 29
ARCHER 1 (1650-1700) 69 45 (39%) 114 25 3 (11%) 28
COPC (1680-1780) 80 166 (67%) 246 28 3 (10%) 31
ARCHER2a (1700-1750) 52 46 (47%) 98 54 26 (33%) 80
ARCHER2b (1750-1800) 70 124 (64%) 194 42 23 (35%) 65
ARCHER3a (1800-1850) 33 107 (76%) 140 7 18 (72%) 25
ARCHER3b (1850-1900) 34 205 (86%) 239 5 46 (90%) 51
ARCHER4a (1900-1950) 9 66 (88%) 75 1 16 (94%) 17
ARCHER4b (1950-1990) 11 176 (94%) 187 4 63 (94%) 67

Total 859 1128 (57%) 1987 265 214 (45%) 479

Chi-square = 512.8 Chi-square = 215.8

df = 12, Ρ < 0.001 df = 12, Ρ < 0.001

The relative figures obtained for the action uses follow closely the figures
obtained for the total of the examples with mutative verbs: have occurs in
over 60% of the instances from the late 1700s on (the figures for the
Century of Prose Corpus even point to an earlier breakthrough of the
38 Merja Kytö

form; see Table 5 above). However, the process uses retain the verb be in
nearly or over 70% of instances up until the early nineteenth century. A
levelling process then takes place and over the first half of the nineteenth
century have supersedes be, coming closer to the figures obtained for ac-
tion uses.

7. Extralinguistic factors

7.1. Regional variation (1750-1990)

The material included in the ARCHER Corpus contains some 328,000

words sampled from the latter halves of the eighteenth, nineteenth and
twentieth centuries: the data offers the possibility to compare the rate and
directions of change in British and American English (the overwhelming
use of the form be in the British English data from the pre-1700 period
makes comparisons of the earlier regional variation less interesting). Ta-
ble 8 presents the figures obtained for the distributions of forms in the
two varieties.

Table 8. Be versus have in British and American English (action and process
uses).

British English American English

Action/process be have Total be have Total

ARCHER 1 (1650-1700) 94 48 (34%) 142 _ — _

ARCHER2b (1750-1800) 66 68 (51%) 134 46 79 (63%) 125
ARCHER3b (1850-1900) 21 136 (87%) 157 18 115 (86%) 133
ARCHER4b (1950-1990) 7 162 (96%) 169 8 77 (91%) 85

Total 188 414 (69%) 602 72 271 (79%) 343

Chi-square = 182.2 Chi-square =30.17

df = 3, Ρ < 0.001 df = 3, Ρ < 0.001
 Be/have + past participle 39

According to the results obtained, there are signs of an innovative ten-

dency in late eighteenth-century American English. The breakdown fig-
ures for action and process verbs point to the process verbs, in particular,
as the catalyst for the faster rate of change (see Table 9).

Table 9. Action versus process verbs in British and American English.

British English American Engish

be have Total be have Total

Action verbs

ARCHER 1 (1650--1700) 69 45 (39%) 114 _ _ _

ARCHER2 (1750--1800) 41 62 (60%) 103 29 62 (68%) 91
ARCHER3 (1850--1900) 16 113 (88%) 129 18 92 (84%) 110
ARCHER4 (1950-1990) 3 124 (98%) 127 8 52 (87%) 60

Total 129 344 (73%) 473 55 206 (79%) 261

Chi-square = 125.8 Chi -square = 10.00

df = 3, Ρ < 0.001 df : = 3, Ρ < 0.01

Process verbs

ARCHER 1 (1650--1700) 25 3 (11%) 28 — _ _

ARCHER2 (1750--1800) 25 6 (19%) 31 17 17 (50%) 34
ARCHER3 (1850--1900) 5 23 (82%) 28 0 23 (100%) 23
ARCHER4 (1950--1990) 4 38 (90%) 42 0 25 (100%) 25

Total 59 70 (54%) 129 17 65 (79%) 82

Chi-square = 67.58 Chi -square = 30.28

df = 3, Ρ < 0.001 d f := 3, Ρ < 0.001
40 Merja Kytö

7.2. Text type variation

7.2.1. Organizing data

As pointed out above (see Table 1), the corpora studied differ in structure
and size. Moreover, there are differences in the way that texts have been
selected, defined and grouped to represent one type or another (e.g. dra-
ma, (private) letters). The subsections drawn from the Helsinki Corpus
and the Century of Prose Corpus present a more varied selection of text
types than the material drawn from the ARCHER Corpus. 15 To facilitate
comparisons, the textual distributions of the variant forms will be organiz-
ed as follows:

- the tables will be organized to preserve the original text type defini-
tions assigned by the compilers of the corpora, and the affinities
within related text types will be taken into account when interpreting
the figures.

- with the (sub)types that could be thought to display some degree of

diachronic continuity, the results obtained will be organized as fol-
lows:

- the types of texts followed across the corpora include fiction,

science, sermons, drama, diaries/journals, and (private) letters.
- the results obtained for the diaries in the Helsinki Corpus will be
compared to those obtained for the journals included in the
ARCHER Corpus. 16
- in the Helsinki Corpus a distinction is made between private let-
ters (or correspondence between close family members) and offi-
cial letters; this distinction is preserved in the present study. In
the other corpora studied no such rigid distinction is made. 17

- the extracts of representative writers in Part A of the Century of

Prose Corpus will be grouped together under the heading "COPC A"
(no attempt has been made to distinguish types of text in further
detail).
 Be/have + past participle 41

- the term "other" groups together the results obtained for texts which
have no direct counterparts in the other corpora or which contribute
only a relatively small number of examples to the study (e.g., law,
documents, official letters, homilies, prefaces and rules from the Hel-
sinki Corpus, and essays, periodicals and polemics from Part Β of the
Century of Prose Corpus).

7.2.2. The text types followed across the corpora

The textual distributions obtained for the variant forms will be presented
in four tables (Tables 10 to 12, and 15). Table 10 gives the overall figures
for the (sub)types of texts represented across the corpora studied (neither
drama nor sermons are included in the Century of Prose Corpus): have
occurs in over 50% of the instances recorded for the texts assigned the la-
bels fiction, diary/journal, (private) letter, drama,18 science and sermon.
Interestingly, the relative figures are highest for diary/journal and fiction,
both types of texts approaching the more informal end of the stylistic
continuum.

Table 10. Be versus have in various text types (I).

be have Total

Fiction 195 402 (67%) 597

Diary/journal 96 217 (69%) 313
(Private) letter 110 111 (50%) 221
Drama 145 162 (53%) 307
Science 59 73 (55%) 132
Sermon 65 93 (59%) 158

Chi-square = 149.3, df = 8, Ρ < 0.001

To follow the process of change within each text type, the breakdown fig-
ures for the different corpora are given in Table 11 (the figures obtained
for the ARCHER Corpus are given by century).
42 Merja Kytö

Table II. Be versus have in various text types (II).

be have Total

Fiction Late ME (1350-1500) 24 16 (40%) 40

EModE (1500-1710) 38 16 (30%) 54
ARCHER 1 (1650-1700) 18 17 (49%) 35
COPC (1680-1780) 15 14 (48%) 29
ARCHER2 (1700-1800) 71 72 (50%) 143
ARCHER3 (1800-1900) 23 153 (87%) 176
ARCHER4 (1900-1990) 6 114 (95%) 120

Diary EModE (1500-1710) 29 17 (37%) 46

Journal ARCHER 1 (1650-1700) 21 6 (22%) 27

ARCHER2 (1700-1800) 24 53 (69%) 77
ARCHER3 (1800-1900) 15 67 (82%) 82
ARCHER4 (1900-1990) 7 74 (91%) 81

Private letter Late ME (1350-1500) 18 2 (10%) 20

EModE (1500-1710) 40 7 (15%) 47

Letter ARCHER 1 (1650-1700) 15 6 (29%) 21

COPC (1680-1780) 7 14 (67%) 21
ARCHER2 (1700-1800) 17 24 (59%) 41
ARCHER3 (1800-1900) 11 27 (71%) 38
ARCHER4 (1900-1990) 2 31 (94%) 33

Drama Late ME (1350-1500) 28 2 (7%) 30

EModE (1500-1710) 25 5 (17%) 30

ARCHER 1 (1650-1700) 19 4 (17%) 23
ARCHER2 (1700-1800) 44 17 (28%) 61
ARCHER3 (1800-1900) 23 78 (77%) 101
ARCHER4 (1900-1990) 6 56 (90%) 62
 Be/have + past participle 43

Table 11. Continued.

be have Total

Science Late ME (1350-1500) 2 0 2

EModE (1500-1710) 6 3 9
ARCHER 1 (1650-1700) 16 12 (43%) 28
COPC (1680-1780) 8 7 (47%) 15
ARCHER2 (1700-1800) 23 16 (41%) 39
ARCHER3 (1800-1900) 2 20 (91%) 22
ARCHER4 (1900-1990) 2 15 (88%) 17

Sermon Late ME (1350-1500) 28 9 (24%) 37

EModE (1500-1710) 12 5 (29%) 17
ARCHER 1 (1650-1700) 5 4 9
ARCHER2 (1700-1800) 13 13 (50%) 26
ARCHER3 (1800-1900) 5 31 (86%) 36
ARCHER4 (1900-1990) 2 31 (94%) 33

Chi-square = 484.4, df = 6, Ρ < 0.001

While the nineteenth century is, as expected, the period when the final
rise of have takes place in most text types shown above (with relative fig-
ures all 70% or above recorded for ARCHER3), journals show relatively
high figures as early as the eighteenth century (ARCHER2). In journals the
rise of have is most marked from ARCHER 1 to ARCHER2 (47%), while
in fiction, drama and science a corresponding j u m p takes place only be-
tween ARCHER2 to ARCHER3. Further, journals and letters are the types
of text in which have is used in nearly or over 60% of the instances. All
in all, the evidence is scantier for letters and sermons. To highlight the
crucial stages of development, further breakdown figures will be given to
the subperiods distinguished for the eighteenth and nineteenth centuries in
the ARCHER Corpus (see Table 12; percentages are given for the better
represented categories only).
44 Merja Kytö

Table 12. Be versus have in various text types (III).

be have Total

Fiction ARCHER2a 1700-1750) 33 20 (38%) 53

ARCHER2b 1750-1800) 38 52 (58%) 90
ARCHER3a 1800-1850) 8 65 (89%) 73
ARCHER3b 1850-1900) 15 88 (85%) 103

Journal ARCHER2a 1700-1750) 13 9 (41%) 22

ARCHER2b 1750-1800) 11 44 (80%) 55
ARCHER3a 1800-1850) 9 22 (71%) 31
ARCHER3b 1850-1900) 6 45 (88%) 51

Letter ARCHER2a 1700-1750) 6 5 (45%) 11

ARCHER2b 1750-1800) 11 19 (63%) 30
ARCHER3a 1800-1850) 8 5 (38%) 13
ARCHER3b 1850-1900) 3 22 (88%) 25

Drama ARCHER2a 1700-1750) 18 4 (18%) 22

ARCHER2b 1750-1800) 26 13 (33%) 39
ARCHER3a 1800-1850) 13 22 (63%) 35
ARCHER3b 1850-1900) 10 56 (85%) 66

Science ARCHER2a 1700-1750) 9 10 (53%) 19

ARCHER2b 1750-1800) 14 7 (33%) 21
ARCHER3a 1800-1850) 1 10 (91%) 11
ARCHER3b 1850-1900) 1 10 (91%) 11

Sermon ARCHER2a 1700-1750) 1 1 2

ARCHER2b 1750-1800) 12 12 (50%) 24
ARCHER3a 1800-1850) 1 1 2
ARCHER3b 1850-1900) 4 30 (88%) 34

Chi-square = 359.1, df = 12, Ρ < 0.001

The figures show that the rise in the use of have in journals takes place
in the latter half of the eighteenth century (for reasons, see below). The
results point to meaningful differences in the process of change along the
 Be/have + past participle 45

textual continuum over the centuries (the influence of the action versus
process uses does not add much to the picture shown in Table 7; for
exact figures, see the table given in Appendix 2).

7.2.3. The influence of come and go on textual distributions

The occurrence of come and go, two of the most frequent verbs in my
data, influences the figures obtained for the textual distributions of be and
have. By way of introduction, Table 13 below presents the figures for
come and go contrasted with the other verbs in the data considered.

Table 13. Come and go versus other (action/process) verbs.

Come and go Other

be have Total be have Total

Late ME (1350-1500) 131 32 (20%) 163 116 46 (28%) 162

EModE (1500-1710) 198 48 (20%) 246 155 83 (35%) 238
ARCHER 1 (1650-1700) 39 12 (24%) 51 55 37 (40%) 92
COPC (1680-1780) 41 33 (45%) 74 93 159 (63%) 252
ARCHER2 (1700-1800) 76 37 (33%) 113 116 158 (58%) 274
ARCHER3 (1800-1900) 57 126 (69%) 183 22 250 (92%) 272
ARCHER4 (1900-1990) 16 111 (87%) 127 9 210 (96%) 219

Total 558 399 (42%) 957 566 943 (62%) 1509

Chi-•square = 258.0 Chi-square = 384.7

df = 6, Ρ < 0.001 df = 6, Ρ < 0.001

The rise of have is clearly less rapid with come and go than with other
verbs: the tendency is evident from the Late Middle English period on,
and the two verbs are equally numerous as the other verbs as late as the
twentienth century. One reason which accounts for the popularity of be
with come and go lies in the Stative type of uses mentioned above (cf.
note 8 and further discussion in section 8.5.2 below). The overall figures
are influenced by how characteristic the occurrence of these verbs is in a
text type (see Table 14 below). As shown above, the rise of have with
verbs other than come and go can be seen to take place earlier in fiction
46 Merja Kytö

and journal writing than in the other text types studied across the corpora;
in these text types have is, similarly, used to a greater extent than be with
verbs other than come and go. In letters and drama, be is quite persistent
with come and go, even during the nineteenth century, when have has
already superseded be with other verbs.

Table 14. Come and go versus other verbs in fiction, diary/journal, (private) let-
ter, drama, science and sermon.

Come and go Other

be have Total be have Total

Fiction Late ME (1350-1500) 14 6 (30%) 20 10 10 (50%) 20

EModE (1500-1710) 18 8 (31%) 26 20 8 (29%) 28
ARCHERl (1650-1700) 9 5 (36%) 14 9 12 (57%) 21
COPC (1680-1780) 9 4 (31%) 13 6 10 (63%) 16
ARCHER2 (1700-1800) 34 13 (28%) 47 37 59 (61%) 96
ARCHER3 (1800-1900) 17 57 (77%) 74 6 96 (94%) 102
ARCHER4 (1900-1990) 4 33 (89%) 37 2 81 (98%) 83

Diary EModE (1500-1710) 23 7 (23%) 30 6 10 (63%) 16

Journal ARCHERl (1650-1700) 9 0 9 12 6 (33%) 18

ARCHER2 (1700-1800) 8 10 (56%) 18 16 43 (73%) 59
ARCHER3 (1800-1900) 10 30 (75%) 40 5 37 (88%) 42
ARCHER4 (1900-1990) 4 23 (85%) 27 3 51 (94%) 54

Private Late ME (1350-1500) 12 2 (14%) 14 6 0 6

letter EModE (1500-1710) 19 3 (14%) 22 21 4 (16%) 25

Letter ARCHERl (1650-1700) 4 3 7 11 3 (21%) 14

COPC (1680-1780) 3 4 7 4 10 (71%) 14
ARCHER2 (1700-1800) 6 4 10 11 20 (65%) 31
ARCHER3 (1800-1900) 8 9 (53%) 17 3 18 (86%) 21
ARCHER4 (1900-1990) 2 15 (88%) 17 0 16 (100%) 16
 Be/have + past participle 47

Table 14, Continued.

Come and go Other

be have Total be have Total

Drama Late ME (1350-1500) 24 0 24 4 2 6

EModE (1500-1710) 18 2 (10%) 20 7 3 10

ARCHER 1 (1650-1700) 13 3 (19%) 16 6 1 7
ARCHER2 (1700-1800) 20 6 (23%) 26 24 11 (31%) 35
ARCHER3 (1800-1900) 19 22 (54%) 41 4 56 (93%) 60
ARCHER4 (1900-1990) 6 30 (83%) 36 0 26 (100%) 26

Late ME (1350-1500) 0 0 0 2 0 2
EModE (1500-1710) 0 0 0 6 3 9
ARCHER1 (1650-1700) 1 0 1 15 12 (44%) 27
COPC (1680-1780) 2 2 4 6 5 (45%) 11
ARCHER2 (1700-1800) 0 2 2 23 14 (38%) 37
ARCHER3 (1800-1900) 0 3 3 2 17 (89%) 19
ARCHER4 (1900-1990) 0 0 0 2 15 (88%) 17

Late ME (1350-1500) 12 3 (20%) 15 16 6 (27%) 22

EModE (1500-1710) 5 4 9 7 1 8
ARCHER1 (1650-1700) 3 1 4 2 3 5
ARCHER2 (1700-1800) 8 2 10 5 11 (69%) 16
ARCHER3 (1800-1900) 3 5 8 2 26 (93%) 28
ARCHER4 (1900-1990) 0 10 10 2 21 (91%) 23

7.2.4. The corpus-specific text types

Table 15 presents the results obtained for the genres that are less con-
sistently represented in the corpora studied. The scope of these results is
rather limited in view of the full span of diachrony studied, but the fig-
ures offer interesting glimpses at the development of text types other than
those illustrated above (Tables 10, 11 and 12). In the text types covering
more than one period, the overall lines of development point, as expected,
to the rise of have.
48 Merja Kytö

Table 15. Be versus have in various text types (IV).

be have Total

Bible Late ME (1350-1500) 23 11 (32%) 34

EModE (1500-1710) 86 9 (9%) 95

Biogr. etc. Late ME (1350-1500) 2 0 2

EModE (1500-1710) 32 12 (27%) 44
COPC (1680-1780) 6 7 (54%) 13

Education EModE (1500-1710) 8 10 (56%) 18

COPC (1680-1780) 3 1 4

Handbooks Late ME (1350-1500) 27 1 (4%) 28

EModE (1500-1710) 15 4 (21%) 19

History Late ME (1350-1500) 29 6 (17%) 35

EModE (1500-1710) 20 16 (44%) 36
COPC (1680-1780) 4 2 6

Other Late ME (1350-1500) 22 15 (41%) 37

EModE (1500-1710) 13 10 (43%) 23
COPC (1680-1780) 14 23 (62%) 37

COPCA COPC (1680-1780) 67 118 (64%) 185

Relig. treat. Late ME (1350-1500) 30 13 (30%) 43

Romance Late ME (1350-1500) 14 3 (18%) 17

Travel EModE (1500-1710) 19 9 (32%) 28

COPC (1680-1780) 10 6 (38%) 16

Trial EModE (1500-1710) 10 8 (44%) 18

Chi-square = 80.48, df = 6, Ρ < 0.001

 Be/have + past participle 49

7.3. The level of formality

As seen above, some types of text reflecting the more informal level of
usage tend to promote the use of the rising variant form. For an overall
picture, the use of be and have was studied in texts labelled as "inform-
al", i.e. comedy, diary, journal, private letter, and letter, as against those
labelled as "formal", i.e. law, documents, official letter, and sermon (for
the situational criteria applied when defining the level of formality, see
Rissanen—Kytö—Palander-Collin 1993: 9-10). However, when pooled
together in this way, the figures obtained for the "informal" texts do not
offer systematic evidence for considering this factor to be of significance
in the development of the be/have paradigm.

7.4. The relationship of the text to spoken language

The closer a text comes to a record of spoken language, the better one
could expect it to reflect ongoing linguistic change. In the classification
adopted for the present study, "speech-based" texts have supposedly been
taken down in real speech situations, while "scripts" include texts intend-
ed to be spoken. The only speech-based texts in the corpora studied are
the trials included in the Early Modern English section of the Helsinki
Corpus, while the scripts included in this study consist of drama and ser-
mons.
Given the relatively limited textual basis, it is hardly surprising that
practically no evidence was found to support the hypothesis that a closer
relationship of the text to spoken language would have promoted the use
of have. Of the 18 examples recorded in the Early Modern English sec-
tion of the Helsinki Corpus, ten occur with be and eight with have. Simi-
larly, the figures obtained for the use of have in "script" texts are lower
than those obtained for "written" texts over the centuries included in this
survey.

7.5. Oral versus other genres

Such factors as text type, level of formality and the influence of the
relationship of a text to spoken language could also be combined to de-
scribe the position a text holds in the continuum, extending from the more
50 Μ er ja Kytö

oral to the more literate genres (see the Introduction to this volume). The
texts defined as "oral" in this study include drama, fiction and (private)
letters, diaries and trials. No evidence was found in this study which
pointed to the role of the orality factor in the development of the be/have
paradigm.

7.6. Men and women writers

The data available for assessing the impact of the author's sex on the
choice of the forms is scarce from the earlier periods. Women writers are
represented to some extent in the Early Modern English section of the
Helsinki Corpus and more systematically so in the ARCHER Corpus. A
number of women writers can also be found among the letter writers in
Part Β of the Century of Prose Corpus, but as there are only a handful of
examples, this corpus will be omitted from this section of the presenta-
tion. The same holds for some less represented text types in the Helsinki
Corpus (e.g. official letters). For meaningful results, only those types of
text that can be found as produced by men and women authors in the
Helsinki Corpus or the ARCHER Corpus will be considered, that is, fic-
tion, journal, letter, drama, science, and sermon (the texts by anonymous
authors will be excluded from the analysis).
In the Late Middle English data women writers tend to use the rising
form have to a greater extent than men writers. In the Early Modern Eng-
lish period the figures obtained for both men and women authors remain
more or less even (see Table 16).

Table 16. The use of be/have by men and women writers in the Helsinki Cor-
pus: fiction, diary, private letter, religious treatise, travel.

be have Total

ME (1350-1500) Women 17 11 (39%) 28

Men 50 20 (29%) 70

EModE (1500-1710) Women 26 9 (26%) 35

Men 80 33 (29%) 113
 Be/have + past participle 51

This is the case with the results obtained for the first subperiod in the
ARCHER Corpus (1650-1700) as well. However, from the 1700s on, up
until the 1900s, women writers remain more conservative and systemati-
cally prefer the form be (see Table 17).

Table 17. The use of be/have by men and women writers in the ARCHER Cor-
pus: fiction, journal, letter, drama, science, sermon.

be have Total

ARCHER la (1650--1700) Women 14 8 (36%) 22

Men 75 34 (31%) 109
ARCHER2a (1700--1750) Women 24 9 (27%) 33
Men 50 34 (40%) 84
ARCHER2b (1750--1800) Women 29 28 (49%) 57
Men 83 118 (59%) 201
ARCHER3a (1800--1850) Women 14 31 (69%) 45
Men 26 94 (78%) 120
ARCHER3b (1850--1900) Women 5 15 (75%) 20
Men 34 234 (87%) 268
ARCHER4a (1900--1950) Women 1 12 (92%) 13
Men 9 70 (89%) 79
ARCHER4b (1950--1990) Women 4 69 (95%) 73
Men 11 170 (94%) 181

Weighing the influence of the extralinguistic factors

By way of summary, of the main extralinguistic factors thought to have

influenced the development of the be/have paradigm, text type and the
author's sex seem to have been of most potent influence across the cen-
turies (the role of foreign influence will be discussed under the treatment
of linguistic factors in section 8.5.1, concentrating on the loan words used
as main verbs). I shall now turn to the role of a number of essential lin-
guistic factors; for the fundamental axes (stative versus mutative, and ac-
tion versus process uses), see section 6 above.
52 Merja Kytö

8. Linguistic factors

8.1. Three verb constructions

As a variant form necessarily occurs in one of the four verbal construc-
tions considered in the present study — present perfect, past perfect, per-
fect infinitive and the -ing construction — the influence of this structural
context is intrinsic. This means that the other linguistic factors to be dealt
with are, in some sense, optional contextual co-occurrence features rather
than independent factors. They are of interest not only for their own value
but also vis-ä-vis the verb construction found in the context. The possible
combined effect of these contextual factors will be taken into considera-
tion in the discussion which follows.

8. J. 1. Perfect versus past perfect

The past perfect, which highlights the perfectivity of action, paved the
way for the rise of have; for corpus examples, see (9a) and (9b).

(9) a. Nothing material happened during our journey to Mosco, where

we arrived on the thirtieth of July, and joined the ambassador,
who had arrived there two days before, having passed us on the
road. (COPC_B/Travel/John Bell/1719)

b. By this I would be understood to mean, the additional articles of

news collected out of the daily historians, from all which, I nev-
er could conceive, what kind of advantage can redound to a ra-
tional creature, who can neither receive instruction or entertain-
ment, by reading that Mr Suchaone died at his country house,
when perhaps the gentleman is in perfect good health; and if
Squire Rentroll is arrived in town with a grand retinue, I appre-
hend it in no way interesting to any man breathing, except his
tailor. (COPCB/Periodical/Gray 's Inn Journal! 1753)

From early on, the use of have is more common in past perfect than in
present perfect constructions (see Table 18); the present perfect remains
one of the last strongholds of be, up until this century.
 Be/have + past participle 53

Table 18. Present perfect versus past perfect constructions.

Present perfect Past perfect

be have Total be have Total

Late ME (1350-1500) 154 26 (14%) 180 93 26 (22%) 119

EModE (1500-1710) 166 37 (18%) 203 158 52 (25%) 210
ARCHER 1 (1650-1700) 43 14 (25%) 57 36 21 (37%) 57
COPC (1680-1780) 73 72 (50%) 145 45 77 (63%) 122
ARCHER2 (1700-1800) 110 63 (36%) 173 62 95 (61%) 157
ARCHER3 (1800-1900) 48 166 (78%) 214 27 142 (84%) 169
ARCHER4 (1900-1990) 11 158 (93%) 169 11 124 (92%) 135

Total 605 536 (47%) 1141 432 537 (55%) 969

Chi-:square = 390.7 Chi-:square = 275.2

df = 6, Ρ < 0.001 df == 6, Ρ < 0.001

Again, the rise of have is more noticeable with action verbs than with
process verbs (for evidence, see Appendix 3).

8.1.2. Perfect infinitive

The perfect infinitive, like the past perfect, emphasizes perfectivity and
thus promotes the use of have; for corpus examples, see (10a) and (10b).

(10) a. Jay called upon me at half-past three in his gig, to take me to

dinner at his father's, Hartham Park. Company at dinner: the
Hawkins; Lady Frances W. was to have come with them, but, to
my somewhat disappointment, she had been called away to Lon-
don the day before; (ARCHER3a/Journal/Thomas Moore)

b. The parliament writs were delivered yesterday. Mr Skelton is

still in the Tower, and is in more danger than was at first appre-
hended. We hear no further of the Dutch fleet. Captain Ouseley
is said to be come to town to give his reasons for tossing the
Mayor of Scarborough in a blanket. (COPC_B/Letters/1688)
54 Merja Kytö

The use of be was rare as early as the Late Middle English period (see
Table 19).

Table 19. Perfect infinitive versus other constructions.

Perfect infinitive Other

be have Total be have Total

Late ME (1350-1500) 0 26 (100%) 26 247 52 (17%) 299

EModE (1500-1710) 9 40 (82%) 49 344 91 (21%) 435
ARCHER1 (1650-1700) 6 6 (50%) 12 88 42 (32%) 119
COPC (1680-1780) 4 31 (89%) 35 130 161 (55%) 291
ARCHER2 (1700-1800) 6 17 (74%) 23 186 179 (49%) 365
ARCHER3 (1800-1900) 4 55 (93%) 183 75 321 (81%) 396
ARCHER4 (1900-1990) 2 30 (94%) 32 23 291 (93%) 314

Total 31 205 (87%) 236 1093 1137 (51%) 2230

Chi-•square = 26.29 Chi·-square = 674.2

df =: 6, Ρ < 0.001 df == 6, Ρ < 0.001

Among the various types of perfect infinitive constructions, the presence

of a modal auxiliary has been a further Äave-triggering factor (see ex-
amples (11a) and ( l i b ) , and Table 20).

(11) a. And trewely, most dred and souueraign lord, gladder ne moor
confortable tithinges might neuer haue come, nor in better tyme,
for to satisfie and refresshe J>e feruent desir of your poure
lieges, J)at haue loong thrusted aftur knowlech of your prosper-
ite, than were your sayd gracious lettres, ... (London Letters
(LLETT) 79)

b. ... for my Affairs are a little dubious at present, perhaps I may

be gone in half an Hour, perhaps I may be your Guest till the
best part of that be spent; (Farquhar, The Beaux Stratagem 8)
 Be/have + past participle 55

Table 20. Types of perfect infinitive constructions (the breakdown figures are
given for the instances with modal auxiliaries only).

Perfect infinitive + modal auxiliary

be have Total

Late ME (1350-1500) 0 22 (100%) 22

EModE (1500-1710) 3 30 (91%) 33
ARCHER 1 (1650-1700) 5 6 (55%) 11
COPC (1680-1780) 0 18 (100%) 18
ARCHER2 (1700-1800) 4 13 (76%) 17
ARCHER3 (1800-1900) 4 47 (92%) 51
ARCHER4 (1900-1990) 2 24 (92%) 26

Total 18 160 (90%) 178

Chi-square = 23.48, df = 6, Ρ < 0.001

Perfect infinitive + other than modal auxiliary

Total 13 45 (78%) 58

8.1.3. -Ing constructions

As -ing constructions express the prolongation and durativity of action,

one could expect them to favour have with mutative intransitives (Ry-
den—Brorström 1987: 193-194; see examples (12a) and (12b); see, also,
the discussion on durative contexts, below). In my data the Aave-promot-
ing influence of the -ing constructions is less evident (see Table 21).

(12) a. Saturday, September 20th Having fallen into a sound sleep

while drifting with the tide into New York, I was awoke about
the first dawn of the morning, ... (ARCHER2b/Journal/William
Strickland)
56 Merja Kytö

b. ... not only deprived of the favor of God, but also of His image;
of all virtue, righteousness, and true holiness, and sunk partly
into the image of the devil, in pride, malice, and all other dia-
bolical tempers; partly into the image of the brute, being fallen
under the dominion of brutal passions and groveling appetites.
(ARCHER2b/Sermon/John Wesley)

Table 21. -Ing constructions.

-ing Other

be have Total be have Total

Late ME (1350-1500) 0 0 0 247 78 (24%) 325

EModE (1500-1710) 20 2 (9%) 22 333 129 (28%) 462
ARCHER1 (1650-1700) 9 8 (47%) 17 85 41 (33%) 126
COPC (1680-1780) 12 12 (50%) 24 122 180 (60%) 302
ARCHER2 (1700-1800) 14 20 (59%) 34 178 175 (50%) 353
ARCHER3 (1800-1900) 0 13 (100%) 13 79 363 (82%) 442
ARCHER4 (1900-1990) 1 9 10 24 312 (93%) 336

Total 56 64 (53%) 120 1068 1278 (54%) 2346

Chi·•square = 34.87 Chi--square = 620 .0

df = 6, Ρ < 0.001 d f := 6, Ρ < 0.001

Given that the past perfect, perfect infinitive and the -ing constructions all
tend to promote the use of have, the influence of the co-occurrence fea-
tures might be of most interest in instances of the present perfect, the last
stronghold of the receding form be. This co-occurrence effect of con-
textual factors will be taken into consideration in the discussion which
follows.

8.2. Durative and iterative contexts; conditional / optative uses

As shown above, the semantic distinction between action and process

uses has had a profound influence on the long-term development of the
 Be/have + past participle 57

be/have paradigm. Other contexts found to have been of importance are

durative (see example (13a)), iterative (13b), conditional ((11a) above)
and optative ((13c)) contexts, all prone to promote the use of have (the
examples cited illustrate the early uses, in particular); for extensive dis-
cussion, see Ryden—Brorström (1987).

(13) a. There were in hominum memoria 3. clothiers at one tyme, thus

nam id, Style, Kent and Chapman, by whom the toun of Bath
then florishid. Syns the death of them it hath sumwhat de-
cayed. (Leland, The Itinerary of John Leland I 143)

b. On whom [= six grete dogges] Reynart the theef had grete

enuye by cause they were so sure that he cowde none gete of
them / how wel oftymes hath this fel theef goon rounde aboute
this wal / and hath leyde for vs in suche wyse that the dogges
haue be sette on hym and haue hunted hym away (Caxton, The
History of Reynard the Fox 11)

c. God wolde ye had nat come here; but ye com never in felyship
of worshipfful folke for to do good, but allwayes grete harme.
(Malory, Morte Darthur 50)

Though the relevant examples found for the above categories are in the
minority if compared with the total of examples studied, these contexts
form a strong co-occurrence pattern with the use of have (see Table 22;
as different semantic distinctions may overlap in the same context, dura-
tive contexts are contrasted against non-durative, iterative against non-
iterative and conditional / optative against non-conditional / non-optative
contexts).
Restricting the set of examples included at this stage of the analysis to
constructions with the present perfect does not change the relative figures
obtained for durative and iterative contexts to any great extent. More evi-
dent is the co-occurrence of the verb construction as a have-promoting
factor in the conditional / optative category: apart from two examples
found in IF clauses, the six examples all occur in instances of the past
perfect or perfect infinitive (with a modal auxiliary).
58 Merja Kytö

Table 22. Durative, iterative and conditional / optative contexts.

be have Total

Durative 24 66 (73%) 90
Other 1100 1276 (54%) 2376

Chi-square = 13.47, df = 1, Ρ <0.001

Iterative 13 46 (78%) 59
Other 1111 1296 (54%) 2407

Chi-square = 13.51, df = 1, Ρ < 0.001

Conditional / optative 8 171 (96%) 179

Other 1116 1171 (51%) 2287

Chi-square = 131.6, df = 2, Ρ < 0.001

8.3. Negation

The number of instances found in negative contexts remains very small

compared with the affirmative contexts. The overall figures suggest that
have occurs in negative contexts from the eighteenth century on more fre-
quently, relatively speaking, than be if compared with the use of the
forms in affirmative contexts (see Table 23).
This tendency holds for the perfect constructions found in the ARCHER
Corpus, again with figures 20% lower for the use of have in the nine-
teenth century data (the examples with negation found in the pre-1700
data in this corpus remain too few for further conclusions). Convergent
results were obtained for the data included in the Century of Prose Cor-
pus.
For an early instance (a. 1470) of the construction found with have +
not, see (13c); notice the co-occurrence pattern with the optative use and
past perfect.
 Be/have + past participle 59

Table 23. Negation.

Negation Affirmative

be have Total be have Total

Late ME (1350-1500) 4 7 11 243 71 (23%) 314

EModE (1500-1710) 28 13 (32%) 41 325 118 (27%) 443
ARCHER1 (1650-1700) 6 4 10 88 45 (34%) 133
COPC (1680-1780) 7 16 (70%) 23 127 176 (58%) 303
ARCHER2 (1700-1800) 6 22 (79%) 28 186 173 (48%) 359
ARCHER3 (1800-1900) 2 27 (93%) 29 77 349 (82%) 426
ARCHER4 (1900-1990) 0 22 (100%) 22 25 299 (92%) 324

Total 53 111 (68%) 164 1071 1231 (53%) 2302

Chi-•square = 48.47 Chi-square = 610.5

df =: 6, Ρ < 0.001 df = 6, Ρ < 0.001

8.4. Complements in the context

The presence of a complement has been claimed to add to the notion of

perfectivity and thus to promote the use of have (for a discussion, see the
criteria for selecting examples given in section 4 above). For the purposes
of this study, the definition given for a "complement" has been rather
flexible. Two clear categories emerge, the instances of process or like
uses with an adjective complement in the context (as in (14a) and (14b)),
and instances with a noun complement (see (15a) and (15b)).

(14) a. On Sir John Perrott's Arivall at Waterford he mett there with

the Lord Cheife Justice, that worthy Gentleman Sir William
Drewry, who was newly come off of a Journey and was fallen
Sicke; soe that being very feeble in his Bedd, he did Knight Sir
William Pelham, who succeeded hym in Goverment, Sir Tho.
Perrott the Sonne of Sir John Perrott, with Sir William Gorges
and one other; (Perrott(?), The History of That Most Eminent
Statesman, Sir John Perrott 112)
60 Merja Kytö

b. At present therefore an unconstrained carriage, and a certain

openness of behavior, are the height of good breeding. The fash-
ionable world has grown free and easy; our manners sit more
loose upon us: nothing is so modish as an agreeable negligence.
(COPC A/Joseph Addison/1711)

(15) a. Lord Fop. ... People of that Rank being under those Decorums,
they can seldom give you a langer View, than will just serve to
shoot 'em flying. So that the Course of my other Pleasures is
not very much interrupted by my Amours.
Lov. But your Lordship now is become a Pillar of the State;
you must attend the weighty Affairs of the Nation. (Vanbrugh,
The Relapse of Virtue in Danger 138)

b. A man, who has contracted a custom of eating fruit by the use

of pears or peaches, will satisfy himself with melons, where he
cannot find his favorite fruit; as one, who has become a drunk-
ard by the use of red wines, will be carried almost with the
same violence to white, if presented to him. (COPC A/David
Hume/1739)

The third category is a mixed bag, covering instances with complements

indicating time, place, manner and so forth, see (16a) and (16b); for com-
parison, instances with no complements are given in (17a) and (17b).

(16) a. Throckmorton: No, I will neuer accuse my selfe unjustly; but in

as muche as I am come hither to bee tried, I pray you let me
haue the Law favourably. (State Trials, The Trial of Throckmor-
ton I 67.C2)

b. "Where dwellest thou?" quoth this parson. "Alas! sir," saithe

this roge, "I haue smal dwelling, and haue com out of my way;
and I should now," saith he, "go to any towne nowe at this time
of night, they woulde set me in the stockes and punishe me."
(Harman, A Caveat or Warening for Commen Cursetors 39)
 Be/have + past participle 61

(17) a. When they were come, and had setled themselves to attentione,
Doctor Heath, Archbishop of Yorke, and Lord Chauncelor of
England ... declared to them the assured advertisement of the
death of the Queene: (Hayward, Annals of the First Four Years
of the Reign of Queen Elizabeth 2)

b. For he [= Erne] eteth no more than ones a day / he lyueth as a

recluse / he chastiseth his body and wereth a sherte of heer / hit
is more than a yere that he hath eten no flesshe / as I yesterday
herd saye of them that cam fro hym he hath lefte and geuen
ouer his Castel maleperduys (Caxton, The History of Reynard
the Fox 10)

The overall figures obtained for the use of the variant forms confirm that
a complement in the context supports, from early on and up until the
present-day, the use of have (see Table 24).

Table 24. The presence of a complement in the context.

Complement No complement

be have Total be have Total

Late ME (1350-1500) 179 69 (28%) 248 68 9 (12%) 77

EModE (1500-1710) 286 114 (29%) 400 67 17 (20%) 84
ARCHER1 (1650-1700) 72 45 (38%) 117 22 4 (15%) 26
COPC (1680-1780) 121 175 (59%) 296 13 17 (57%) 30
ARCHER2 (1700-1800) 135 165 (55%) 300 57 30 (34%) 87
ARCHER3 (1800-1900) 44 323 (88%) 367 35 53 (60%) 88
ARCHER4 (1900-1990) 14 273 (95%) 287 11 48 (81%) 59

Total 851 1164 (58%) 2015 273 178 (39%) 451

Chi-:square = 552.4 Chi·-square = 108.0

df = 6, Ρ < 0.001 df == 6, Ρ < 0.001
62 Merja Kytö

Eliminating the possible influence of the /zave-promoting verb construc-

tions (past perfect, perfect infinitive and the -ing construction), the same
holds for most of the data found in perfect constructions. An exception to
the rule is the Early Modern English section of the Helsinki Corpus, in
which the percentages for the use of have remain more or less the same
irrespective of the complement context (only a handful of examples are
found in the Century of Prose Corpus; see Table 25).

Table 25. The presence of a complement in the context in perfect constructions.

Complement No complement

Perfect be have Total be have Total

Late ME (1350-1500) 109 22 (17%) 131 45 4 (8%) 49

EModE (1500-1710) 137 31 (18%) 168 29 6 (17%) 35
ARCHER 1 (1650-1700) 32 12 (27%) 44 11 2 (15%) 13
COPC (1680-1780) 70 67 (49%) 137 3 5 8
ARCHER2 (1700-1800) 76 55 (42%) 131 34 8 (19%) 42
ARCHER3 (1800-1900) 30 135 (82%) 165 18 31 (63%) 49
ARCHER4 (1900-1990) 5 134 (96%) 139 6 24 (80%) 30

Total 459 456 (50%) 915 146 80 (35%) 226

Chi-square = 323.7 Chi-square = 73.49

df = 6, Ρ <0.001 df == 6, Ρ <0.001

Compared with the overall figures obtained for the other categories dis-
tinguished, the Aave-promoting influence of nouns indicating distance,
measure and like concepts has been particularly strong; the role of adjec-
tive complements remains less clear (see Table 26).
In the light of the results obtained for the perfect constructions, ad-
jective complements have had little connection with the rise of have (of
the 71 examples with be or have, 42 or 60% have be)·, in this category
the figures obtained for noun complements are not much more convincing
 Be/have + past participle 63

(of the 80 examples, 43 occur with have·, owing to the breakdown of ex-
amples across the centuries, the amount of the data representative of the
earlier periods remains scarce).

Table 26. Some types of complements.

Adjective Distance, measure etc.

be have Total be have Total

Late ME (1350-1500) 5 0 5 1 6 (46%) 13

EModE (1500-1710) 8 0 8 16 12 (44%) 28
ARCHER1 (1650-1700) 3 1 4 5 8 (62%) 13
COPC (1680-1780) 26 5 (16%) 31 10 18 (64%) 28
ARCHER2 (1700-1800) 15 16 (52%) 31 11 25 (69%) 36
ARCHER3 (1800-1900) 4 27 (87%) 31 2 32 (94%) 34
ARCHER4 (1900-1990) 0 27 (100%) 27 0 25 (100%) 25

Total 61 76 (55%) 137 51 126 (71%) 177

Chi-square = 71.54 Chi--square = 35.06

df = 6, Ρ < 0.001 df == 6, Ρ < 0.001

Other

Late ME (1350-1500) 167 63 (27%) 230

EModE (1500-1710) 262 102 (28%) 364
ARCHER 1 (1650-1700) 64 36 (36%) 100
COPC (1680-1780) 85 152 (64%) 237
ARCHER2 (1700-1800) 109 124 (53%) 233
ARCHER3 (1800-1900) 38 264 (87%) 302
ARCHER4 (1900-1990) 14 221 (94%) 235

Total 739 962 (57%) 1701

Chi -square = 475.5

df == 6, Ρ < 0.001
64 Merja Kytö

8.5. Main verb

8.5.1. Loan verbs

The use of loan words as main verbs is a possible factor promoting the
rise of have, as they could be thought to add to the resources of language
by attracting the use of a rising rather than a receding form. Evidence
supporting this hypothesis was found in the data studied. For Table 27
three categories of verbs were distinguished, the early (or the pre-1400
century) and the later (or the post-1400) Latinate and Romance loans, the
remaining set of examples being assigned to the category defined as
"other". The influence of the loan word factor is evident during the eight-
eenth and nineteenth centuries.
Again, eliminating the possible influence of /zave-promoting verb con-
structions, the corresponding figures are lower than those given in Table
27 by some 10% for the use of have with the early loans up until the
1800s (for the eighteenth century the difference is even greater, around
20%). Of the later loans, no more than 26 examples are found in perfect
constructions, but 20 of these occur with have.

Table 27. Loan words.

Early loans Later loans

be have Total be have Total

Late ME (1350-1500) 60 17 (22%) 77 0 0 0

EModE (1500-1710) 81 38 (32%) 119 0 5 5
ARCHER 1 (1650-1700) 27 13 (33%) 40 3 2 5
COPC (1680-1780) 20 71 (78%) 91 3 11 (79%) 14
ARCHER2 (1700-1800) 43 79 (65%) 122 4 9 (69%) 13
ARCHER3 (1800-1900) 7 104 (94%) 111 1 13 (93%) 14
ARCHER4 (1900-1990) 4 65 (94%) 69 1 18 (95%) 19

Total 242 387 (62%) 629 12 58 (83%) 70

Chi-square = 199.5 Chi-square = 12.25

df = 6, Ρ < 0.001 df = 6, Ρ < 0.05
 Be/have + past participle 65

Table 27. Continued.

Other

be have Total

Late ME (1350-1500) 187 61 (25%) 248

EModE (1500-1710) 272 88 (24%) 360
ARCHER 1 (1650-1700) 64 34 (35%) 98
COPC (1680-1780) 111 110 (50%) 221
ARCHER2 (1700-1800) 145 107 (42%) 252
ARCHER3 (1800-1900) 71 259 (78%) 330
ARCHER4 (1900-1990) 20 238 (92%) 258

Total 870 897 (51%) 1767

Chi-square = 464.0
df = 6, Ρ < 0.001

8.5.2. Frequent main verbs

The development of the be/have paradigm with very frequent main verbs
is of special interest. The four verbs represented in the material studied
with more than 100 occurrences are come, go, pass and fall. Of these
pass and fall follow the overall pattern for the rise of have, pass ahead of
the general rate of change with have in over 90% of the examples from
the 1700s on (see Table 28; cf. action verbs in Table 7). For an early cor-
pus example of the occurrence of fall with have (from 1477-1484), see
(18):

(18) And for as moche as syth the accomplyssheme[n]te of this sayd

booke made by the sayd Ranulph ended the yere of oure lord a
/ M/CCC/lvij / many thynges haue fallen whiche ben requysyte
to be added to this werke / by cause mennes wyttes in this tyme
ben oblyuyous and lyghtly forgeten many thynges dygne to be
putte in memorye / (Caxton, The Prologues and Epilogues 68)
66 Merja Kytö

Table 28. Pass and fall as main verbs.

Pass Fall

be have Total be have Total

ME (1350-1500) 22 6 (21%) 28 17 5 (23%) 22

EModE (1500-1710) 18 7 (28%) 25 17 7 (29%) 24
ARCHER 1 (1650-1700) 0 7 7 0 7 7
COPC (1680-1780) 1 28 (97%) 29 9 17 (65%) 26
ARCHER2a (1700-1750) 2 22 (92%) 24 1 2 3
ARCHER2b (1750-1800) 2 19 (90%) 21 3 12 (80%) 15
ARCHER3a (1800-1850) 1 16 (94%) 17 1 9 10
ARCHER3b (1850-1900) 0 15 (100%) 15 1 13 (93%) 14
ARCHER4a (1900-1950) 0 3 3 0 5 5
ARCHER4b (1950-1990) 1 8 9 0 5 5

Total 47 131 (74%) 178 49 82 (63%) 131

Chi-square = 94.42 Chi-square = 47.07

df = 6, Ρ < 0.001 df = 6, Ρ < 0.001

The figures obtained for the perfect constructions do not change the pic-
ture given by the overall figures (with fall and pass the breakdown fig-
ures remain too low for further conclusions).
As pointed out above, come and go favour the use of be far longer than
do the other verbs (see note 8 and Table 13). For come the final turn-over
period starts in the 1800s, for go a century later (see Table 29). In all the
examples with go from the final subperiod of the ARCHER Corpus, the
emphasis is on state, see examples (19a) and (19b):

(19) a. I now know that except for the restored and restorable build-
ings along the river it [= the old city] is gone, irretrievably.
(ARCHER4b/JournaI/Ralph Kirkpatrick)

b. He led us to his third-floor room, ... and handed one portion to

Gretchen, with the command, 'Allow the rice paper to dissolve
in your mouth, then swallow it.' And he was gone.
(ARCHER4b/Fiction/James A. Michener)
 Be/have + past participle 67

For a corpus example from the early eighteenth century (1707), see ( l i b )
above.

Table 29. Come and go as main verbs.

Come Go

be have Total be have Total

ME (1350-1500) 79 21 (21%) 100 52 11 (17%) 63

EModE (1500-1710) 131 23 (15%) 154 67 25 (27%) 92
ARCHER 1 (1650-1700) 18 7 (28%) 25 21 5 (19%) 26
COPC (1680-1780) 21 13 (38%) 34 20 20 (50%) 40
ARCHER2a (1700-1750) 14 4 (22%) 18 19 4 (17%) 23
ARCHER2b (1750-1800) 15 13 (46%) 28 28 16 (36%) 44
ARCHER3a (1800-1850) 7 22 (76%) 29 18 10 (36%) 28
ARCHER3b (1850-1900) 2 56 (97%) 58 30 38 (56%) 68
ARCHER4a (1900-1950) 3 16 (84%) 19 5 12 (71%) 17
ARCHER4b (1950-1990) 0 47 (100%) 47 8 36 (82%) 44

Total 290 222 (43%) 512 268 177 (40%) 445

Chi-square = 223.4 Chi-square = 71.03

df = 6, Ρ < 0.001 df = 6, Ρ < 0.001

Weighing the influence of linguistic factors

By way of summary, the more powerful linguistic factors seen to have in-
fluenced the development of the be/have paradigm include various seman-
tic contexts (notably action/process uses, and durative, iterative and condi-
tional or optative contexts), the verb construction (notably tense and per-
fect infinitive), and the presence of an object-like or other complement.
For further testing of the linguistic and extralinguistic factors studied,
I shall now turn to logistic regression analysis, a method specifically suit-
ed to the study of the significance that the simultaneous influence and
interaction of factors may have in unevenly distributed data.
68 Merja Kytö

9. Factor interaction: a generalized linear model analysis

9.1. Methodological considerations

The frequency tables given above highlight the processes of change in

many respects, but the comparison of marginals does not necessarily tell
the whole truth about the combined influence of different factors, espe-
cially when the data analysed is unevenly distributed. One of the more
powerful methods of dealing with the simultaneous influence of factors
and factor interaction is logistic regression analysis.
Within logistic regression analysis, the impact of various factors on the
use (or non-use) of the variant forms can be assessed by testing how sta-
tistical models based on various combinations of factors account for vari-
ation in the data. Though the statistical format of the data available for
the present study did not lend itself to more complex modelling pro-
cedures, the significance of the individual factors and their interaction
could be studied. The analysis was carried out with the help of the Glim
program using a logit link (McCullagh—Neider 1983; Francis—Green—
Payne 1993). 19 First I calculated a model excluding the impact of both
linguistic and extralinguistic factors. I then calculated the simultaneous
influence of factors, without and with interaction between them (see be-
low). The model fitted by the program was E(log(p/(l-p))), in which Ε
was the mean, and ρ the probability with which a variant form may (or
may not) occur in a cell.
Owing to the breakdown of the data, no more than two factors could be
combined in each model calculated. The subperiod division was one of
the two factors included in each model throughout the analysis, as this
factor was, as expected, the one which scored extremely highly in statis-
tical significance when fitted to the model which excluded the impact of
both linguistic and extralinguistic factors. The other central factors con-
sidered were text type, the level of formality, the relationship of a text to
spoken language, orality, author's sex, tense, perfect infinitive, object
complement, action/process uses, main verb, loan words as main verbs,
negation and the -ing construction.
 Be/have + past participle 69

9.2. Evaluating the simultaneous influence and interaction of

factors

Two kinds of models were calculated. First the simultaneous influence of

two factors was studied without interaction between the factors and then
with interaction between them. In the former instance we observed how
the mere presence of the factors in the model accounts for the occurrence
of be or have; in the second we observed how the factors act against each
other. For instance, when calculating the model for the simultaneous in-
fluence of the text type factor and time without interaction between the
two factors, we assumed that the influence of text type does not change
from one subperiod to another; when calculating the model for the simul-
taneous influence of text type and time with interaction, we assumed that
the influence of text type may change across the subperiods.
In Table 30, I present the levels of significance for the simultaneous
influence of two factors without interaction (13 models, with time and
one other factor in each), and the simultaneous influence of the factors
with interaction (13 models, respectively).

Table 30. The models describing the simultaneous influence of two factors: (1)
without interaction and (2) with interaction. The levels of significance:
****** = highly significant (p < 0.001); *** = significant (p < 0.01);
* = nearly significant (p < 0.05); - = non-significant.

Simultaneous influence Without With

interaction interaction
Time
****** ******
text type
****** ***
relationship to spoken language
*** ******
level of formality
* ***
orality
****** *
author's sex
****** ******
action/process verb
****** ******
main verb
****** ***
loan word as a main verb
****** ******
tense (perfect / past perfect)
****** ******
perfect infinitive
***
-ing construction -

******
negation -
****** ******
object-like complement
70 Merja Kytö

With a few exceptions (the level of formality, orality and the -ing con-
struction), the simultaneous influence of time combined with one of the
other factors without interaction proved to be highly significant. The re-
sults obtained for the simultaneous influence of the two factors with inter-
action vary to a great extent, pointing to changes in the influence of these
factors across the diachrony (the figures of lesser significance obtained
for factors such as the author's sex may reflect biases in the data: the ma-
terial from later periods includes a greater number of texts by women
writers than the material from the earlier periods).
Finally, along with time, the other highly significant factors proved to
be the text type, action/process uses, main verb, tense, perfect infinitive
and the presence of an object-like or other complement.

10. Concluding remarks

In this study the stages of the be/have variation with mutative intransi-
tives are traced back in history from the Late Middle English period to
present-day usage. A number of linguistic and extralinguistic factors can
be seen to have influenced the choice of the form over the centuries. In
the Late Middle English period, the use of have increases gradually, gains
in momentum in the late 1700s and supersedes the use of be in the early
1800s.
Among the more potent extralinguistic factors influencing the process
of change are chronology and text type; among the more powerful lin-
guistic factors are action/process uses, durative, iterative and conditional
contexts, main verb, tense, perfect infinitive and the presence of an object
complement.
Evaluating the coverage and depth of the present study, indeterminate
examples excluded from the analysis would merit further discussion in
another context. The influence of prescriptive grammarians on the final
turn-over stages of the paradigm would, similarly, be worth further atten-
tion (for a discussion, see Ryden—Brorström 1987: 206-211). Finally,
still more could be learnt of the dynamics of the be/have paradigm by
turning to such large-scale issues as the establishment of the verb be in
passive constructions, the transitivization of many originally intransitive
verbs and the increasing use of many past participles as adjectives or
nouns.
 Be/have + past participle 71

Appendix 1.

The verbs studied.

Verb HC/ME HC/EModE COPC ARCHER Total

1350-1420 1500-1710 1680-1780 1650-1990

abate 5 _ 1 6
abide 5 1 - - 6
advance - - 3 8 11
alight - - - 3 3
allay - 1 - - 1
alter - 2 2 6 10
amble - - - 1 1
amend 1 - 1 2
amount - 1 - 1 2
appear 1 3 20 15 39
arise 1 2 3 7 13
arrive - 3 13 46 62
ascend 2 - - 2
aslaken 1 - - - 1
augment - - - 2 2
awake 1 - 1
become 7 10 36 112 165
befall 1 - - 6 7
begin 2 6 8 26 42
break 1 4 1 6 12
brew 1 - - - 1
burn - - 1
burst 2 - - - 2
bust — - - 1 1
cease 3 1 1 13 18
chance - 1 - 3 4
change 2 5 1 20 28
clear - - - 1 1
climb 1 1 2
come 100 154 34 224 512
congeal - - - 2 2
convert 1 1 2
cool 4 - - 4
creep 1 - 2 1 4
72 Merja Kytö

Appendix 1. Continued.

Verb HC/ME HC/EModE COPC ARCHER Total

1350-1420 1500-1710 1680-1780 1650-1990

decay _ 2 _ _ 2
decline - - 1 1 2
decrease 1 - - 1 2
degenerate - - 1 - 1
degrade - - - 3 3
delay - 1 - - 1
depart 6 15 4 7 32
descend 2 - 1 6 9
deteriorate - - - 1 1
digress - - 1 - 1
diminish - - - 3 3
disappear - - - 18 18
distil - 1 - - 1
diverge - - - 1 1
do 8 1 15 24
drift - - 2 2
drive 1 1 - 1 3
drivel - - - 1 1
drop - - - 4 4
dry 1 4 - 2 7
dwell 5 4 1 - 10
ebb - 1 - 1 2
elapse - - 2 9 11
embark - - - 2 2
emerge - - - 4 4
enlarge 1 - - 1
enter 8 12 5 18 43
err 1 1 - - 2
escape - 6 3 13 22
evaporate - - 1 1 2
exile 1 - - - 1
expand - - - 1 1
expire 2 5 2 4 13
extend - 1 - 2 3
fade - - - 1 1
fail 1 - 8 5 14
fall 22 24 26 59 131
 Be/have + past participle 73

Appendix 1. Continued.

Verb HC/ME HC/EModE COPC ARCHER Total

1350-1420 1500-1710 1680-1780 1650-1990

fare _ 1 1
ferment - 1 1
flee 5 1 2 15 23
float 1 - - - 1
flow - - - 3 3
fly - 2 - 4 6
form 1 - - - 1
freeze - 1 - 17 18
gender 1 - - - 1
get 1 9 7 46 63
go 63 92 40 250 445
grow 1 14 25 50 90
hap - 1 - - 1
happen 1 5 27 54 87
harden - - 2 2
hasten - - 1 1
heal 1 - - - 1
improve - - 2 2
increase - 1 1 12 14
inhabit - 1 - - 1
intrude - 1 - - 1
issue - 1 - 1
itinerate - - - 1 1
jump - - - 4 4
keep - 1 - - 1
land - 1 2 5 8
languish - - - 1 1
last - - - 1 1
lean - 1 1
leap - 1 1
leave 1 1 - 8 10
lengthen - - - 1 1
lie 4 9 2 8 23
lunge - - - 1 1
march - 2 2 4
melt 1 1 2 4
mend 1 - - 1
miscarry - 1 2 1 4
74 Merja Kytö

Appendix 1. Continued.

Verb HC/ME HC/EModE COPC ARCHER Total

1350-1420 1500-1710 1680-1780 1650-1990

mount 1 _ „ 1
move - - - 10 10
nigh 2 - - - 2
obtain - 1 - - 1
occur - - - 2 2
part 2 - 2 5 9
pass 28 25 29 96 178
perch - - - 1 1
perish - 2 2 4 8
pirouette - - - 1 1
proceed - 6 3 6 15
progress - — — 2 2
putrefy - - 4 - 4
reach - - - 3 3
recover - 3 1 4 8
reduce - - - 1 1
relapse - - 1 - 1
remain - 3 4 5 12
rest 1 4 1 - 6
retire - - 2 10 12
retract - 1 - - 1
retreat - - - 2 2
return - 11 6 39 56
revive - - 1 - 1
ride 4 1 1 1 7
rise 3 6 6 11 26
roast 1 - - - 1
roll - - 2 2
row 1 2 - - 3
rub — - - 1 1
run 5 5 6 23 39
rush - - - 1 1
rust - - - 1 1
sail 1 - 2 8 11
secede - - - 1 1
set - - 4 5 9
shrink — - - 2 2
 Be/have + past participle 75

Appendix 1. Continued.

Verb HC/ME HC/EModE COPC ARCHER 'Intal

1350-1420 1500-1710 1680-1780 1650-1990

sink 3 7 8 18
sit 3 4 1 2 10
slacken - - 1 - 1
slip - 1 2 2 5
soften - - - 2 2
split - - 1 2 3
spring - 1 5 5 11
stand 6 5 1 7 19
start - - - 5 5
stay - 6 2 10 18
steal - - - 1 1
step - 1 - 1 2
stir - - 1 1 2
stop - - - 4 4
stray - - 2 2
subside - - 1 1
succeed - - 2 - 2
succumb - - - 1 1
surface - - - 1 1
swage 1 - - - 1
swell - - - 1 1
swim - - - 1 1
tarry 1 1 - - 2
thaw - - - 1 1
traffic - - - 2 2
transpire - - 1 2 3
travel - 2 5 6 13
trespass 5 - - - 5
trip - - - 1 1
turn 9 8 4 15 36
vanish - - 3 10 13
wade - 1 2 - 3
wake - 1 - - 1
walk 3 8 1 10 22
wander - - - 3 3
waste 5 2 - 7
wax 4 2 — 6
76 Merja Kytö

Appendix 1. Continued.

Verb HC/ME HC/EModE COPC ARCHER Total

1350-1420 1500-1710 1680-1780 1650-1990

weaken — — — 1 1
wear - 1 - - 1
weigh — 1 - - 1
wend 4 - - 1 5
widen - 1 - - 1
win 2 - - 1 3
wither — - - 3 3
youth 1 - - - 1

Total 361 552 409 1550 2872

Appendix 2.

Action versus process verbs in some types of text studied (fiction, diary/journal,
(private) letter, drama, science and sermon):

Action verbs Process verbs

be have Total be have Total

Fiction LME (1350-1500) 20 16 (44%) 36 4 0 4

EModE (1500-1710) 36 15 (29%) 51 2 1 3
ARCHER1 (1650-1700) 14 16 (53%) 30 4 1 5
COPC (1680-1780) 13 14 (52%) 27 2 0 2
ARCHER2 (1700-1800) 52 62 (54%) 114 19 10 (34%) 29
ARCHER3 (1800-1900) 22123 (85%) 145 1 30 (97%) 31
ARCHER4 (1900-1990) 5 88 (95%) 93 1 26 (96%) 27

Diary EModE (1500-1710) 28 16 (36%) 44 1 1 2

Journal ARCHER 1 (1650-1700) 19 6 (24%) 25 2 0 2

ARCHER2 (1700-1800) 14 50 (78%) 64 10 3 (23%) 13
ARCHER3 (1800-1900) 14 60 (81%) 74 1 7 8
ARCHER4 (1900-1990) 5 58 (92%) 63 2 16 (89%) 18
 Be/have + past participle 77

Appendix 2. Continued.

Action verbs Process verbs

be have Total be have rotal

Private LME (1350-1500) 16 2 (11%) 18 2 0 2

letter EModE (1500-1710) 33 7 (18%) 40 1 0 7

Letter ARCHER 1 (1650-1700) 11 6 (35%) 17 4 0 4

COPC (1680-1780) 6 13 (68%) 19 1 1 2
ARCHER2 (1700-1800) 13 23 (64%) 36 4 1 5
ARCHER3 (1800-1900) 8 23 (74%) 31 3 4 7
ARCHER4 (1900-1990) 2 25 (93%) 27 0 6 6

Drama LME (1350-1500) 27 2 (7%) 29 1 0 1

EModE (1500-1710) 23 4 (15%) 27 2 1 3

ARCHER 1 (1650-1700) 16 4 (20%) 20 3 0 3
ARCHER2 (1700-1800) 27 11 (29%) 38 17 6 (26%) 23
ARCHER3 (1800-1900) 19 67 (78%) 86 4 11 (73%) 15
ARCHER4 (1900-1990) 6 46 (88%) 52 0 10 10

Science LME (1350-1500) 0 0 0 2 0 2

EModE (1500-1710) 3 1 4 3 2 5
ARCHER 1 (1650-1700) 4 10 (71%) 14 12 2 (14%) 14
COPC (1680-1780) 4 3 7 4 4 8
ARCHER2 (1700-1800) 4 13 (76%) 17 19 3 (14%) 22
ARCHER3 (1800-1900) 1 14 (93%) 15 1 6 7
ARCHER4 (1900-1990) 0 5 5 2 10 (83%) 12

Sermon LME (1350-1500) 20 9 (31%) 29 8 0 8

EModE (1500-1710) 9 4 (31%) 13 3 1 4
ARCHER 1 (1650-1700) 5 4 9 0 0 0
ARCHER2 (1700-1800) 12 10 (45%) 22 1 3 4
ARCHER3 (1800-1900) 3 25 (89%) 28 2 6 8
ARCHER4 (1900-1990) 2 20 (91%) 22 0 11 11
78 Merja Kytö

Appendix 3.

Action versus process verbs in present perfect and past perfect constructions.

Action verbs Present perfect Past perfect

be have Total be have Total

LME (1350-1500) 120 25 (17%) 145 83 26 (24%) 109

EModE (1500-1710) 135 33 (20%) 168 138 50 (27%) 188
ARCHER1 (1650-1700) 32 12 (27%) 44 23 21 (48%) 44
COPC (1680-1780) 37 61 (62%) 98 30 69 (70%) 99
ARCHER2 (1700-1800) 73 52 (42%) 125 37 85 (70%) 122
ARCHER3 (1800-1900) 39 144 (79%) 183 24 111 (82%) 135
ARCHER4 (1900-1990) 9 112 (93%) 121 8 95 (92%) 103

Total 445 439 (50%) 884 343 457 (57%) 800

Process verbs Present perfect Past perfect

be have Total be have Total

ME (1350-1500) 34 1 (3%) 35 10 0 10
EModE (1500-1710) 31 4 (11%) 35 20 2 (9%) 22
ARCHER 1 (1650-1700) 11 2 (15%) 13 13 0 13
COPC (1680-1780) 36 11 (23%) 47 15 8 (35%) 23
ARCHER2 (1700-1800) 37 11 (23%) 48 25 10 (29%) 35
ARCHER3 (1800-1900) 9 22 (71%) 31 3 31 (91%) 34
ARCHER4 (1900-1990) 2 46 (96%) 48 3 29 (91%) 32

Total 160 97 (38%) 257 89 80 (47%) 169

Notes

1. For further details and references to literature, see Ryden—Brorström

(1987: 18-20).
2. For principles of diachronic variation analysis, see Weinreich—Labov—
Herzog (1968); Romaine (1982); Rissanen (1986). An exhaustive list of
literature on the be/have variation can be found in Ryden—Brorström
(1987); see also note 4 below. For discussion on systemic and pragmatic
 Be/have + past participle 79

aspects of the perfect construction, see Rainer (1989) and McCoard (1978);
for the use of be and have in sequences of auxiliaries in the course of the
history of English, see Warner (1995, 1997).
The WordCruncher data retrieval program was used to search for examples
with be and have in the corpora studied. Additional searches were carried
out on individual verbs to retrieve the more irregular spellings (e.g. the
form a for have·, for lists of verbs, see, e.g. Friden 1948; Söderlind 1951:
48 ff.; Visser 1952, 1963-1973; Rydön—Brorström 1987).
For classifications and principles of classification, see e.g. Jespersen (1931:
29-46); Frid6n (1948: 43 ff.); Visser (1963-1973: 2044-2084).
The sources for the examples drawn from the Helsinki Corpus are listed in
the Bibliography at the end of this volume; see also Kytö (1996: 133-163).
The examples cited from the Century of Prose Corpus and the ARCHER
Corpus are based on the information found in the database.
The examples drawn from the ARCHER Corpus are referred to by the sub-
section of the corpus, text type and the author's name. The examples drawn
from the Century of Prose Corpus are referred to by the author's name and1
or by text type, and the year given in the database (see Mili6 1990).
The OED has been consulted when weighing the status of possibly ambigu-
ous adjectival or adverbial past participles. For example, the past participle
was regarded as indeterminate in the following examples by John Locke
(all examples are from his Directions Concerning Education, p. 54):
For when the children are grown up and these ill habits with them ...
And that when they are grown up they should looke on them as their
best and their only sure freinds ...
Cf. in participial constructions when the auxiliary is omitted:
... unlesse you have a minde to make your children when grown up
weary of you ...
... we must looke upon our children when grown up to be like to our
selves, with the same passions, the same desires.
As pointed out by Denison (1993: 344), Quirk et al. refer to these uses as
"pseudo-passive constructions" in which "the participle is active rather than
passive in meaning", e.g. 'By the time she got there, her friend was gone'
(1985: 170, note [a]). In Present-day English, perfective constructions are
"almost synonymous" with pseudo-passive constructions and have super-
seded these with most intransitive verbs, compare 'Mary has come' with
(the archaic) ' M a r y is come' (Quirk et al. 1985: 170, note [a]).
In these examples the element of action is emphasized and the use of have
promoted (for discussion and references, see Ryd6n—Brorström 1987: 23).
In some studies examples such as (4b) have been regarded as "transitive"
and excluded from analysis by terming the complement "object", or "quasi-
object"; in others they have been regarded as "intransitive" (Nakamura
80 Merja Kytö

1987: 29), or as surface structures of locative or instrumental cases (e.g.

Ryd6n—Brorström 1987: 22-24).
11. According to Ryden—Brorström (1987), this distinction is, above all, as-
pectual (for problems related to the use of the term "aspect" as a grammati-
cal category, see, e.g. Zandvoort 1962: 19-20; Mitchell 1985: I, §§865-
873). Discussions on the role of other factors can be found in, e.g., Friden
(1948); Johannisson (1958); Zimmermann (1973); Kakietek (1987 [1976]);
Ryd6n—Brorström (1987); Ryd6n (1991).
12. I have already discussed some aspects of the data drawn from the Early
Modern English period in a pilot study (Kytö 1994). The figures in the pre-
sent study may differ in minor details from those already given in the pilot
study, mainly owing to revisions in the criteria for including examples in
the study.
13. For the purposes of statistical analysis, the examples were coded and the
tables produced by means of the SURVO program (see Mustonen 1992;
Korhonen 1993; Sadeniemi 1991).
14. The references given by Ryden—Brorström (1987: 198, note 16) include
Barber (1976: 261), Fridin (1948: 115, 117), Kisbye (1971: 130) and
Poutsma (1904-1916: II, 2: 216).
15. In the Helsinki Corpus the text type definitions have been assigned to indi-
vidual texts on the basis of extralinguistic criteria (see Rissanen—Kytö—
Palander-Collin 1993: 10-14); the same principle holds for the other cor-
pora studied.
16. The category of journals in the ARCHER Corpus include diaries and person-
al travel accounts, compiled from day to day or within a broader time per-
spective.
17. For the most part the letters included in the ARCHER Corpus are representa-
tive of correspondence engaged in between more or less close friends at the
informal end of the stylistic continuum (though more formal letters also
occur, usually by writers representative of lower ranks than the addressee).
The figures given for the Century of Prose Corpus stand for the category
definition "letters/memoirs" (no more than 21 instances were recorded for
this category).
18. Examples found in stage directions have been excluded from the study.
19. For a helpful discussion on logistic regression analysis from the linguist's
point of view, see Sankoff (1988). I am grateful to Ms Leena Sadeniemi at
the Computing Centre of the University of Helsinki for useful comments on
the method and the Glim program.
 Be/have + past participle 81

References

Barber, Charles
1976 Early Modem English. London: Andrö Deutsch.
Biber et al. = Biber, Doug—Edward Finegan—Dwight Atkinson—Ann Beck—
Dennis Burges—Jena Burges
1994a "The design and analysis of the ARCHER Corpus: A progress report
[A Representative Corpus of Historical English Registers]", in: Merja
Kytö—Matti Rissanen—Susan Wright (eds.), 3-6.
Biber et al. = Biber, Doug—Edward Finegan—Dwight Atkinson
1994b "ARCHER and its challenges: Compiling and exploring a representa-
tive corpus of historical English registers", in: Udo Fries—Gunnel
Tottie—Peter Schneider (eds.), Creating and using English language
corpora. Papers from the Fourteenth International Conference on
English Language Research on Computerized Corpora, Zurich 1993.
Amsterdam—Atlanta, GA: Rodopi, 1-13.
Denison, David
1993 English historical syntax: Verbal constructions. London—New York:
Longman.
Elsness, Johan
1989 "The English present perfect: Has it seen its best days?", in: Leiv
Egil Breivik—Arnoldus Hille—Stig Johansson (eds.), Essays on Eng-
lish language in honour of Bertil Sundby. Oslo: Novus Forlag, 95-
106.
1991 The perfect and the preterite. The expression of past time in con-
temporary and earlier English. PhD thesis, University of Oslo.
Filppula, Markku
1994 "From Anglo-Irish to Hiberno-English: Divergence and convergence
in the Irish dialects of English", in: Wolfgang Viereck (ed.), ZDL-
Beiheft 76: Verhandlungen des Internationalen Dialektologen-kon-
gresses Bamberg 1990. (Band 3.) Stuttgart: Franz Steiner Verlag,
180-196.
Francis, Brian—Mick Green—Clive Payne
1993 The Glim system. Oxford: Clarendon Press.
Frid6n, Georg
1948 Studies on the tenses of the English verb from Chaucer to Shake-
speare, with special reference to the late sixteenth century. (Essays
and Studies on English Language and Literature 2.) Uppsala: Alm-
qvist and Wiksell.
Helsinki Corpus
1991 The Helsinki Corpus of English Texts. Helsinki: Department of Eng-
lish, University of Helsinki.
82 Merja Kytö

Jespersen, Otto
1931 Α Modem English grammar in historical principles. Part IV: Syntax;
Third Volume: Time and tense. Heidelberg: Carl Winters Universi-
tätsbuchhandlung.
Johannisson, Türe
1958 "On the be and have constructions with mutative verbs", Studia Lin-
guistica 12: 106-118.
Kakietek, Piotr
1987 "The perfect auxiliaries in the language of Shakespeare", in: Vivian
Salmon—Edwina Burness (eds.), A reader in the language of Shake-
spearean drama. (Benjamins Paperbacks 7.) Amsterdam—Phila-
delphia: John Benjamins. [Reprinted from Studia Anglica Posnanien-
sia 8 (1976): 45-53.]
Kallen, Jeffrey L.
1989 "Tense and aspect categories in Irish English", English World-Wide
10: 1-39.
Kisbye, Torben
1971- An historical outline of English syntax. Aarhus: Akademisk Boghan-
1972 del.
Korhonen, Markku
1993 SURVO 84C peruskurssi. [SURVO 84C Basic Course.] (Guides of
Computing Centre 20.) Helsinki: Computing Centre, University of
Helsinki.
Kytö, Merja
1994 "BE vs. HAVE with intransitives in Early Modern English", in: Fran-
cisco Fernändez—Miguel Fuster—Juan Jose Calvo (eds.), English
historical linguistics 1992. (Current Issues in Linguistic Theory 113.)
Amsterdam—Philadelphia: John Benjamins, 179-190.
Kytö Merja (comp.)
1996 Manual to the diachronic part of the Helsinki Corpus of English
Texts. Coding conventions and lists of source texts. (3rd edition.)
Helsinki: Department of English, University of Helsinki.
Kytö, Merja—Matti Rissanen—Susan Wright (eds.)
1994 Corpora across the centuries. Proceedings of the First International
Colloquium on English Diachronic Corpora, St Catherine's College
Cambridge, 25-27 March 1993. (Language and Computers: Studies
in Practical Linguistics 11.) Amsterdam—Atlanta, GA: Rodopi.
McCoard, Robert W.
1978 The English perfect: Tense-choice and pragmatic inferences. (North-
Holland Linguistic Series 38.) Amsterdam—New York: North-Hol-
land Publishing Company.
 Be/have + past participle 83

McCullagh, P. S.—J. A. Neider

1983 Generalized linear models. (2nd edition.) London: Chapman and
Hall.
Melchers, Gunnel
1992 '"Du's no heard da last o'dis' — on the use of be as a perfective
auxiliary in Shetland dialect", in: Matti Rissanen—Ossi Ihalainen—
Terttu Nevalainen—Irma Taavitsainen (eds.), History of Englishes.
New methods and interpretations in historical linguistics. (Topics in
English Linguistics 10.) Berlin—New York: Mouton de Gruyter,
602-610.
Milid, Louis T.
1990 "A new historical corpus", ICAME Journal 14: 26-39.
1994 "Is a grapheme a lexeme? And other problems of definition in the
Century of Prose Corpus", in: Merja Kytö—Matti Rissanen—Susan
Wright (eds.), 65-71.
Mitchell, Bruce
1985 Old English syntax. I. Oxford: Clarendon Press.
Mustonen, Seppo
1992 SURVO. An integrated environment for statistical computing and re-
lated areas. Helsinki: Survo Systems Ltd.
Nakamura, Fujio
1987 "Auxiliary have in Samuel Pepys's 'Diary', with special reference to
mutative verbs: (1) A statistical study", Studies in Modern English 3:
23-51.
OED = The Oxford English dictionary
1961 Vols. 1-12 with supplement and bibliography. Ed. by James A. H.
Murray—H. Bradley—W. A. Craigie—C. T. Onions. Oxford: Claren-
don Press.
1989 Second edition.
Poutsma, Hendrik
1904- A grammar of Late Modern English. Groningen: P. Noordhoff.
1929
Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik
1985 A comprehensive grammar of the English language. London—New
York: Longman.
Rainer, Eva Maria
1989 Das Perfekt im Spätmittel- und Frühneuenglischen: eine Frequenz-
und Funktionalanalyse anhand von Brieftexten. (Innsbrucker Beiträge
zur Kulturwissenschaft. Anglistische Reihe 2.) Innsbruck: Amce.
Rissanen, Matti
1986 "Variation and the study of English historical syntax", in: David San-
koff, Diversity and diachrony. (Current Issues in Linguistic Theory
53.) Amsterdam—Philadelphia: John Benjamins, 97-109.
84 Merja Kytö

Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.)

1993 Early English in the computer age: Explorations through the Helsinki
Corpus. (Topics in Linguistics 11.) Berlin—New York: Mouton de
Gruyter.
Romaine, Suzanne
1982 Socio-historical linguistics, its status and methodology. (Cambridge
Studies in Linguistics 34.) Cambridge: Cambridge University Press.
Ryd£n, Mats
1991 "The be/have variation with intransitives in its crucial phases", in:
Dieter Kastovsky (ed.), Historical English syntax. (Topics in English
Linguistics 2.) Berlin—New York: Mouton de Gruyter, 343-354.
Ryd£n, Mats—Sverker Brorström
1987 The be/have variation with intransitives in English, with special ref-
erence to the Late Modern period. (Acta Universitatis Stockholmien-
sis/Stockholm Studies in English 70.) Stockholm: Almqvist and Wik-
sell International.
Sadeniemi, Leena
1991 Kieliaineiston käsittelyä Survolla. Versio 1.1. [Applying Survo to lin-
guistic data. Version 1.1.] (Guides of Computing Centre 18.) Hel-
sinki: Computing Centre, University of Helsinki.
Sankoff, David
1988 "Variable rules", in: Ulrich Ammon—Norbert Dittmar—Klaus J.
Mattheier (eds.), Socio!inguistics / Soziolinguistik. An international
handbook of the science of language and society / Ein internationales
Handbuch zur Wissenschaft von Sprache und Gesellschaft (Hand-
bücher zur Sprach- und Kommunikationswissenschaft, eds. Hugo
Steger—Herbert Ernst Wiegand 3.2.) Berlin—New York: Walter de
Gruyter, 984-997.
Söderlind, Johannes
1951 Verb syntax in John Dryden's prose. I—II. (Essays and Studies on
English Language and Literature 10.) Uppsala: Almqvist and Wiksell.
Visser, F. Th.
1952 A syntax of the English language of St Thomas More. Part II: The
verb. Louvain: Librairie Universitaire.
1963- An historical syntax of the English language. I—III. Leiden: E. J.
1973 Brill.
Warner, Anthony R.
1995 "Predicting the progressive passive: Parametric change within a lexi-
calist framework", Language 71: 533-557.
1997 "Extending the paradigm: An interpretation of the historical develop-
ment of auxiliary sequences in English", English Studies 78/2: 162-
189.
 Be/have + past participle 85

Weinreich, Uriel—William Labov—Marvin I. Herzog

1968 "Empirical foundations for a theory of language change", in: W. P.
Lehmann—Yakov Malkiel (eds.), Directions for historical linguistics:
A symposium. Austin—London: University of Texas Press, 95-195.
Zandvoort, R. W.
1962 "Is 'aspect' an English verbal category?", in: Charles L. Barber—
Frank Behre—Urban Ohlander—Yngve Olsson—Svante Stubelius—
Johannes Söderlind—R. W. Zandvoort (eds.), Contributions to Eng-
lish syntax and philology. (Gothenburg Studies in English 14.) Göte-
borg—Stockholm—Uppsala: Almqvist and Wiksell, 1-20.
Zimmermann, Ruediger
1973 "Structural change in the English auxiliary system: On the replace-
ment of be by have", Folia Linguistica 6: 107-117.
On the forms and functions of the verb be
from Old to Modern English

Matti Kilpiö

1. Introduction

This paper examines two topics which are both connected with the verb
be but are really two separate, distinct areas of study: developments in the
morphology of be and developments in the functional load of be from
Old English to Early Modern English.
Anyone embarking on a diachronic survey of the development of be
will soon find out how kaleidoscopically the areas of main interest shift
when we move from one main period of English into another. Thus in
Old English the rich morphology of the verb, with its two parallel present
tense paradigms (see 2.1 below) and the syntactic and semantic uses to
which the coexisting variant forms are put (see Kilpiö 1992 and 1993)
seem to provide the most interesting and challenging area of study. In
Middle English other areas of interest come to the foreground:

(1) The geographical variation seen in the forms of be. This is con-
siderable, before the rise of Standard English particularly;
(2) The rise of new complex verb phrases, made possible by the in-
creasing occurrence of participial forms of be such as beyng,
ibe, etc.
(3) The increase in the functional load of be as a result of the de-
velopment of various auxiliary uses of the verb.

Finally, in Early Modern English, due to standardization processes char-

acteristic of that period, the geographical variation of be becomes difficult
to follow. Developments mentioned under (2) and (3) above, however,
gain in importance.
88 Matti Kilpiö

The present study concentrates on (1) and, particularly, (3) above, as

the latter is an area which has interesting points of contact with Merja
Kytö's chapter in this volume on the rivalry of be and have as temporal
and aspectual markers with the past participle of intransitive verbs.
The data used in the present study is drawn from the Helsinki Corpus
of English Texts. It covers the subperiods OE1-4 (OE1: -850, OE2: 850-
950, OE3: 950-1050, OE4: 1050-1150), ME1 (1150-1250), ME3 (1350-
1420) and EModEl (1500-1570) and is divided between different forms
of be as follows:

(a) Section 1.2 dealing with the morphology of be takes into con-
sideration all present tense forms, 18,257 instances in all, both
indicative and subjunctive. 1 In sections 2 and 3 dealing with the
functions of be and with developments in its functional load, a
selection of data containing every tenth instance2 of the total of
present tense forms is used.

(b) For ME3 and EModEl, past tense forms, both indicative and sub-
junctive, a selection of data comprising every eighth instance of
all past having been analysed;

(c) For ME3 and EModEl, a selection of data consisting of every

tenth instance of non-finite forms, infinitives, past participles
and -ing forms was analysed.

The reason why present tense forms of be dominate the discussion is

partly practical, dictated by the large number of instances of such a high-
frequency word as be. Concentration on the present rather than the past
tense also provides a greater range of topics for detailed study, given the
morphological richness of forms in the present system.

2. An overview of the morphology of the present tense

forms of be from OE to EModE

By and large, the long-range diachronic developments in the inventory of

forms in the present tense show an apparent diversification of forms when
 The verb be 89

we move from OE to ME, and a subsequent process of simplification and

regularization when EModE is reached.

2.1. Old English

The state of things in OE can be seen in Table 1 on p. 90.

The most striking feature of the OE pres. ind. and subj. forms is the
existence of two coexisting paradigms, one with initial b-, the other with-
out. The variation in the pres. ind. is used for semantic and syntactic pur-
poses; thus

(a) 6-forms are more common in actional passives, non-Z>-forms in statal.

In this respect, compare (1) with (2):

(1) forÖsem se yfla willa bid tostenced swas^asr rec beforan fyre
CAlfred's Boethius 117)
'because the evil will is dispersed like smoke before fire'

(2) & Jjasr eac Beorhte seo cwen bebyrged is

(Bede's Ecclesiastical History 110)
'and Queen Bertha is also buried there'

(b) ό-forms rather than ηοη-ό-forms are used with the future; conversely,
deictic locatives or temporals linking the state or action to the present
moment or situation are more common with non-6-forms than ό-forms.
Example (3) combines both the futuric use of 6-forms and the deictic
temporal nu with a non-A-form:

(3) & f>aet his good & sio his gesaelö him nahwonan utane ne com,
ac wass symle on him selfum, & nu is, & a bid.
(Alfred's Boethius 89)
'and that his good and his happiness did not come to him from
anywhere outside, but had always been in him, is now, and will
ever be'
90 Matti Kilpiö

Table 1. Old English: Distribution of pres. ind. and subj. forms.

Pres. ind. sing. 1st pers. 2nd pers. 3rd pers

eom/ beo eart bist is/ bip
wese wesep

OE1 — _ _ _ 7 3
OE2 41 21 23 8 521 423
OE3 180 38 146 26 2035 953
OE4 12 1 12 5 506 320

OE1-4 233 60 181 39 3069 1699

Pres. ind. plural 1st pers. 2nd pers. 3rd pers

sind(on)/ beop sind(on) beop sind(on)/ beop
earon earon earon

OE1 _ 4 1
OE2 4 9 4 189 178
OE3 48 24 38 21 473 329
OE4 4 3 2 - 101 41

OE1-OE4 56 36 44 21 767 549

Total of 6-forms: 2404 = 35.6%; total of non-A-forms: 4350 = 64.4%

Pres. subj. sing. plural

sie beo Total sien beon Total

OE1 21 21 2 2
OE2 355 11 366 89 4 93
OE3 510 96 606 151 49 200
OE4 96 17 113 7 10 17

OE1-4 982 124 1106 249 63 312

Perc. of A-forms of the total of: pres. subj. sg. forms pres. subj. pi. forms
OE1 0.0% 0.0%
OE2 3.0% 4.3%
OE3 15.8% 24.5%
OE4 15.0% 58.8%
OE1-4 11.2% 20.2%
 The verb be 91

(c) ό-forms alone are met in sentences in which the verb can be given an
iterative interpretation, see (4); ό-forms are more common than ηοη-ό-
forms when the verb can be given a durative reading, as in (5);

(4) Swa bid öis eorölice lif oft yögiende swa swa sae, & öeah bid
swiöe acorenlic {Alfred's Cura Pastoralis 409)
'Thus this earthly life often fluctuates like the sea, and yet it is
very worthy of choice'

(5) hi wilniaÖ öaes anes goodes |)e a bid, ^ast is God.

{Alfred's Boethius 94)
'they desire the only good that exists for ever, that is, God'

(d) ό-forms are significantly more frequent than non-6-forms in generic

sentences (Kilpiö 1992 and 1993), see (6).

(6) DeaÖ bid sella

eorla gehwylcum {)onne edwitlif. {Beowulf 2890-2891)
'Death is better to each warrior than a life of dishonour'

For a more detailed discussion of developments in the proportional

share of ό-forms and non-6-forms in the pres. ind., see Kilpiö (1993: 9 9 -
100).
When we turn to the present subjunctive, what strikes the eye is the rise
in the proportionate share of beo, beon in comparison to sie, sien. The
periods crucial for this change seem to be OE2 (850-950) and OE3 (950-
1050), when the percentages of 6-forms rise from 3% to 15.8% in the sg.
and from 4.3% to 24.5% in the pi. The rise continues in OE4 in the plu-
ral. Campbell's treatment of the two pres. subj. paradigms (1959: 351) is
unsatisfactory on two counts: first, what he says about the dialectal dis-
tribution of beo and sie does not hold good as far the gloss to the Gospel
of St. Matthew in the Rushworth Gospels is concerned (see Kilpiö 1989:
73); further, he passes over the diachronic developments very briefly by
only referring to the generalization of beo in 1WS. The rise of beo/beon
during the OE period is confirmed not only by the evidence provided by
the Helsinki Corpus but by the study of the transmission of individual
texts. Thus, for example, the 11th century revision of Gregory's Dia-
logues found in MS Η quite systematically replaces the pi. form syn of
MS C by beon in the pi. and occasionally sy/si by beo in the sg. (See Ap-
pendix).
92 Matti Kilpiö

2.2. Middle English, subperiod ME1 (1150-1250)

Table 2. ME1 (1150-1250): Distribution of pres. ind. and subj. forms.

Pres. ind. 1st pers. 2nd pers. 3rd pers.

am beo art bist is bip

MEX/1:
Singular
Peri Didaxeon (SO) - - - 50 72
Bodley Horn. (SO) 1 - 2 85 38
Holy Roodtr. (SO) 1 - 5 3 23 5
Vesp. Horn. (KO) 6 - 3 3 42 33
Lamb. Horn. (WMO) - - 3 150 11
Trinity Horn. (EMO) 2 - 2 6 77 2

Totals 10 0 15 12 427 161

Plural sind(en) aren beop ben

(+beo we/ge)

Peri Didaxeon (SO) 11 — 28

Bodley Horn. (SO) - - 47
Holy Roodtr. (SO) - - 12
Vesp. Horn. (KO) 11 - 6
Lamb. Horn. (WMO) - - 41
Trinity Horn. (EMO) - 1 3 27

Totals 22 1 137 27

ME1:

Singular 1st pers. 2nd pers. 3rd pers.

am beo art bist is bip

Sawles Warde (WMO) 3 - - - 76 2

Ancrene W. (WMO) 7 - 3 222 3
Hali M. (WMO) 1 - 7 183 8
Layam. Brut (WMO) 14 2 10 1 49 7
 The verb be 93

Table 2. Continued.

Singular 1st pers. 2nd pers. 3rd pers.

am beo art bist is bip

/Catherine (WMO) 4 3 — 43 —

Margarete (WMO) 6 16 - 42 1
Juliane (WMO) 21 2 13 - 51 2
Vices and V. (EMO) 12 - 11 - 162 —

The Pet. Chron. (EMO) - - - - 4

The Ormulum (EMO) 2 1 2 - 93 1

Totals 70 5 65 1 925 24

Plural sind(en) aren beop ben

(+beo we/ge)

Sawles Warde (WMO) — 2 22 —

Ancrene W. (WMO) - - 45 -

Hali M. (WMO) - - 44 -

Layam. Brut (WMO) 1 - 10 -

Katherine (WMO) - - 10 -

Margarete (WMO) - 1 9 -

Juliane (WMO) - - 8 -

Vices and V. (EMO) - - 39 9

The Pet. Chron. (EMO) - - - -

The Ormulum (EMO) 10 - -

Totals 11 3 187 9

Pres. subj.
Sing, si (sy etc.) beo (be, bie etc.) Total of sg. forms
35 191 226
Plural sy(n) beon (bien etc.) Total of pi. forms
3 60 63

Percentage of ό-forms of the total of

Pres. subj. sg. forms: 84.5%
Pres. subj. pi. forms: 95.2%
94 Matti Kilpiö

The texts in Table 2 have been listed in two groups, MEX/1 and ME1.3
The manuscripts of all the texts date from 1150-1250, but those texts
marked as MEX/1 can be assumed, with a varying degree of probability,
to go back to Old English originals; X before the slash (/) refers to the
date of the original version as having the value 'unknown' (see Kytö
1996: 48). With texts representing ME1, the date of composition and
manuscript both fall within the time limits 1150-1250. 4
With the rapid disappearance in ME of the coexistence of the two paral-
lel present tense paradigms of OE, the semantic-syntactic distinction be-
tween the two is also lost. As Forsström (1948: 215-220) points out, the
general tendency in ME was towards a present indicative paradigm with
am, art, is in the singular and bep, ben or are(n), depending on the dia-
lect, in the plural.
In the light of Table 2, the following observations can be made. In the
indicative, the actual coexistence of the b- and non-6-forms is restricted
to the singular forms of texts representing MEX/1: even in this group most
texts have a ό-form as the exclusive plural form so that variation in the
plural between non-Z»- and i-forms is only seen in Peri Didaxeon, Ves-
pasian Homilies and, barely, in Trinity Homilies. Expressed in frequen-
cies/1,000 words, the frequency of ηοη-ό-forms in texts representing
MEX/1 is 11.6 in the singular and 1.3 in the plural; that of ό-forms 4.7 in
the singular and 4.2 in the plural.
In texts of which the original was written in the ME1 period, the b-
forms in the singular are in a clear minority even in texts of West Mid-
land origin and are not met at all in the three texts representing the East
Midland dialect. In the pres. ind. pi. the coexistence of b- and non-6-
forms is restricted to Sawles Warde, Layamon's Brut and Margarete. The
rapid disappearance of Z>-forms in the singular and their diametrically op-
posite generalization in the plural in ME1 texts can be seen in the follow-
ing frequencies/1,000 words: non-6-forms 14.7 in the singular, 0.3 in the
plural; 6-forms 0.4 in the singular, 2.7 in the plural.
Subjunctive forms with initial s- are practically restricted to the MEX/1
period. Thus of the 35 pres. subj. sg. forms with initial s-, 32 occur in
MEX/1 texts {Peri Didaxeon 17, Vespasian Homilies 15) but only three in
ME1 texts (The Ormulum 2, Layamon's Brut 1); all the three pres. subj.
pi. forms with initial s- are met in MEX/1 texts {Peri Didaxeon 2, Ves-
pasian Homilies 1).
 The verb be 95

To sum up, of the forms inherited from OE, losses are suffered both by
Z>-forms in the pres. ind. sg. (beo, bist, bip) and by forms with initial s-
(ind. pi. sind(en), subj. sg./pl. sy/syn). As a result, the most typical present
indicative paradigm of to be in ME1 is am, art, is in the sg. and bep (or
less commonly, be(n)) in the pi. The ind. pi. form aren, a development of
OE earon, is represented by four instances only. In the present subjunc-
tive the typical paradigm is sg. be (beo, bie, bo), pi. ben (beort, bien, bon,
byri).
The losses suffered both by ό-forms in the pres. ind. sg. and by s-forms
in the ind. pi. and subj. sg./pl. are something that could be expected from
OE developments. In Kilpiö (1993) I point out that there is, generally
speaking, a decline in the relative share of ό-forms in the present indica-
tive singular (43.5% in OE2, 30.1% in OE3 and 38.1% in OE4; for a dis-
cussion of the unexpected rise in OE4 and its probable connection with
the composition of the texts in OE4, see Kilpiö 1993: 100). In subperiod
ME1, the corresponding percentage of ό-forms is 28.7% in texts represent-
ing MEX/1 and 2.8% in texts representing ME1. As for the losses suffered
by J-forms, it has already been seen that from OE1 to OE4 the share of s-
forms in the present subjunctive declines steadily and the development is
particularly marked in the plural. Assuming that the gradual loss of s-
forms began with present subjunctive forms, both singular and plural, it
could be hypothetized that, having gradually lost the support of the sub-
junctive s-forms, the present indicative plural sind(on), the last stronghold
of the Λ-forms, was under pressure of being replaced by other forms. On
the analogy of the present subjunctive, the most likely candidate was a b-
form, and, in fact, by far the most common present indicative plural form
in ME1 is bep.

2.3. Middle English, subperiod ME3, with special reference to

LALME

The texts in Table 3a and 3b have been arranged in two groups.5 The
first, smaller group (21,450 words), marked ME2/3, consists of texts with
originals dating to period ME2 (1250-1350), but with manuscripts written
in ME3. The second, larger group (162,780 words) consists of texts com-
posed in ME3 and preserved in manuscripts written in the same period.
96 Matti Kilpiö

I I (Ν CM I I I I I I I I I
δ
•ο

I —I I I I I I I I I I I I I
-5

I I οο I I I I I oo I I I I
oo
υ
ft
m vi m
Ό οο h ^·-'n N H h fΝi iΝn Ή
OΝiOi»»

I I — I I I I I I I I I I I I

λ. I I "Τ I I I I I I I I I I I I
Ι-ι
υ
Ο.
Ό« _«t (Ν ΟΟ rf I<
r1 η η I η Ν I
es NI fN

I Im I I I I I I I I I I

I I I I I I I I I I I I I I

1>
α.
δ I 1-Η —
ι » <N
β

Τ
>-) t/3 z—s
co '—
w
<υ Ζ
as
•ο
.9 Ο 52 οk.
χλ ύ) [Tl »3 eα aε
2 .9 Ο
Ρη 00 S ^ it: ο Η
 The verb be 97

ι ι ι ι ι ι ι ι ι ι ι ι — ι
δ
-ο

I I I I I I I I I I I I I I I I ©
-ο

Ι Ι Ι Ι Ι Ι Ι Ι Ι Ι Ι Ι Ι Ο Ι - Τ } · ON
(Ν Ό (Ν
tN
υ
α.
Τt-l3 0 0 V i e r - - « ' H O 1 0 0 N t ^ ' - ι ι/-> © «Ο -η Γ<1 Ο
γί VO
Ο
IN
I I I I I I I I I I I I I — I I <

I I I I I I I I I I I ο I ΓΛ
υ
Ρ.
Ό
<ν 5u
β CO I I I I I I 00 ι ι ι ι ι ι ι ι VO CQ
VO
<N
Ο Ι/Ί
Η
I I I I I I I I I I I I I —— I <N r<"> fN
00
Tt

I I I I I I I I I I I I I m I I m <N m
<N
Μ
Ο
α
δ 0 0 — | | Ι | | < Ν « Ν | | | | | « 0 | ο i-H r-
β

hJ S c
0 0 _ι
-Ο
1 ws J O W j hJ ω «U
Ö, Ο S •O
ω >-ι s s Ζ *a
•2 Sv ^ w w c ω w Β "Π R ob
κ •S ^ U3 CO
α α
ε « ο§
tu 0
<u Ή
δ ο R Ο
j» <3
. δ ® u CO
I 'S J ^
60
g S fe ·2 -ν I &5 CO
"rt

ο Ο 8
w a 3 >-1 3 05 Η OH
98 Matti Kilpiö

Table 3b. ME3 (1350-1420): Distribution of pres. ind. and subj. forms; plural.

Pres. ind. plural are(n) er(e) is es bep be(n)

ME2/3:
Aelred of R. (SL) 30 6
Handl. Synne (SL) 14 - - - 3
Cursor M. (NL) 7 7 - 2

Totals 21 7 0 2 33 6

ME3:
Treat. Horses (SO) - - - 17
Wycliffe NT (SL) - - - - - 33
Purvey Prol. (SL) - - - - 15
Trevisa Polychr. (SL) - - - - 3 -
Brut (WML) 1 - - - 1 3
Chaucer Boeth. {EMO) - - 1 - - 115
Chaucer CT (EML) 1 59
Chaucer Astr. (EMO) - - - - - 45
Cloud of Unkn. (EMO) 2 - 1 - - 83
Wycliff s. WSERM (EML) 3 56
Wycliff. s. WESERM (EML) - - - - - 49
Equal, of Plan. (EMO) - - - - 12
Conf. A mant is (EMO) - - - - - 6
Lett. Henry V (EMO) _ _ _ _ _ 2
Lett. Henry V (EML) _ _ _ _ _ 3
Judgements (EML) _ _ _ _ _ _
Phlebotomy (EMO) 19 - 1 - 2 1
London Letters (EML) 3 - - - - -
Mandeville (EML) - 1 - - 76
Wycliffe OT (EML) _ _ _ _ _ 26
Petitions (EML) 5 20
Proclamations (EML) 5 — — — — —
Returns (EMO) - - - - 1 13
Testaments (EMO) _ _ _ _ _ 3
Usk Appeal (EMO) _ _ _ _ _ 2
 The verb be 99

Table 3b. Continued.

Pres. ind. plural are(n) er(e) is es bep be(n)

Bened. Rule (NL) 1 6 - 2 - -

N. Horn. Cycle (NL) - 19 - 1 - -

Pricke of Consc. (NL) — 29 — 2 - —

Totals 37 54 4 5 27 622

Pres. subj. pi. be, ben 1 2 3 Total

11 14 101 126

The most salient developments to be gleaned from Tables 3a and 3b are:

(1) The dominant pres. ind. sg. forms are 1 sg. am 2 sg. art 3 sg. is.
Nearly all singular forms with initial b- have disappeared, there is only
one of byp (Handlyng Synne) and three instances of the Northern bes(e)
{Cursor Mundi, The Northern Homily Cycle). At least all the three bes(e)
instances refer to the future; this is probably also true about the byp in-
stance in Handlyng Synne. In texts representing the Northern dialects {The
Benedictine Rule, Cursor Mundi, The Northern Homily Cycle, The Pricke
of Conscience) the predominant pres. ind. 3 sg. form is es. This form is
also occasionally met in the pres.ind. 1 and 2 sg. The eight pres. ind. 3
sg. es forms in Chaucer's Boethius are examples of the limited occurrence
of es as a variant of is in the East Midland dialect (see Forsström 1948:
98). Forsström (1948: 189) favours Wright's hypothesis that es is related
to ON es and is thus of Scandinavian origin.

(2) Outside texts representing the Southern dialect the form bep in the
pres. ind. pi. recedes and gives way on the one hand to be(n), the com-
monest plural form in texts of East Midland origin, and on the other to
are(n), er(e). Of the two last-mentioned forms, are is often said to reflect
Scandinavian influence; on the other hand, Samuels (1985: 274) is of the
100 Matti Kilpiö

opinion that only er(e), encountered in the Great Scandinavian Belt, is un-
questionably of Scandinavian origin. The form es as the pres. ind. 3 pi. is
marginal, there being seven instances scattered among four Northern texts
{The Benedictine Rule (2x), Cursor Mundi (2x), The Northern Homily
Cycle (lx), The Pricke of Conscience (2x)). The same can be said about
the pres. ind. 3 pi. is: the four instances are divided between Chaucer's
Boethius, The Cloud of Unknowing, Mandeville's Travels and Phlebo-
tomy, all of them texts representing the East Midland dialect.
Forms of be from ME3 can profitably be compared with the larger ma-
terial provided by A Linguistic Atlas of Late Middle English {LALME
1986). Among the items for which the Linguistic Profiles of LALME were
collected there are seven for the verb be: 17 ARE, 18 WERE NOR, 19 IS,
20 ART SOU, 21 WAS, 77 BE ppl NOR, 183 NE + BE SOU. The abbrevia-
tions NOR and SOU here refer to the two partly overlapping parts of the
survey covering areas north and south of the Wash (see LALME 1986, vol.
3: x).
Of the items listed above, nos. 17 and 19 provide interesting material
for comparison.
Under Item 17 ARE LALME lists 120 variants. Of these, the HC ME3
material has only seventeen, but the variants in the HC seem to cover all
the main morphological types of late Middle English. One variant, repre-
sented by seven instances in the HC but not in LALME is the pres.ind. pi.
es. As pointed out above, all these instances occur in northern texts.
For Item 19 IS there are 33 variant forms in LALME. The HC ME3 ma-
terial has five; once again the main types are included in the HC material.
One HC spelling, bese, once attested in The Northern Homily Cycle, is
not in the LALME material.
Generally speaking it can be said that the dialect areas attached in the
HC to individual texts and, as a logical extension, to morphological fea-
tures like forms of be met in these texts, seems to be in harmony with the
localization of these features in LALME. This is according to expectations
as an effort has been made to give HC texts the LALME localization where
possible (Nevanlinna—Pahta—Peitsara—Taavitsainen 1993: 46).
 The verb be 101

2.4. Early Modern English, subperiod EModEl

Table 4. EModEl: Distribution of pres. ind. and subj. forms.

Present indicative

Sg. 1 (h)am(e) 2 art(e) 3 is PI. be(en) ar(e) is

181 57 2116 423 317 26

The present subjunctive forms are sg. be(e) (297 instances) and pi. be(e) (102
instances) or been (1 instance).

With EModEl the inventory of forms is reduced. More than anything else
this must be attributed to the process of standardization. As Nevalainen—
Raumolin-Brunberg (1993: 66-67) point out, one of the guiding principles
in compiling the Early Modern British English part of the Helsinki Cor-
pus has been the standardness of texts, the application of which principle,
although not strictly followed throughout, has resulted in a corpus which
more than anything else represents the rising standard. This probably ex-
plains why the pres. ind. 1 sg be and 2 sg. beest mentioned in Görlach
(1978: 87) are not found in the EModEl and the latter definitely not in
any of the periods EModEl-3.
When compared to the state of things in ME3, the most notable new
development is the rise of ar(e) to the status of the most common pres.
ind. pi. form. There is an interesting difference between the plural forms
be(en) and ar(e) in their distribution between the different persons:

Table 5. EModEl: Breakdown of ind. pi. be(en) and ar(e) between different
persons.

1st pers.pl. 2nd pers.pl. 3rd pers.pl.

be(en) 10 30 383
ar(e) 9 58 250
102 Matti Kilpiö

The form be(en) is used almost exclusively in the 3rd person (in 91% of
the total of be(en) instances), whereas ar(e) is used with a 3rd pers. pi.
subject only in 79% of all the ar(e) instances, and is conspicuously com-
mon in the 2nd pers. pi. when compared with be(en). It would be possible
to speculate that the replacement of be(en) by ar(e) may have mainly pro-
ceeded through the 2nd person plural, where the phonetic, and indeed ety-
mological, link between 2nd pers. sg. art and pi. are may have been con-
ducive to the change. That the 2nd pers. pi. seems to be the 'channel'
through which ar(e) was then gaining ground could perhaps be explained
by the use of the polite 2nd pers. pi. pronoun (and the plural verb form)
for the singular. According to this line of argument, the plural verb form
chosen with ye/you would be more likely to be are rather than be(en), as
the verb form used with the singular 2nd pers. subject thou is art, phone-
tically and etymologically related to are. The ultimate disappearance of
be(en) in the indicative plural of course means a reduction in the func-
tional load of the be(en) form since in EModEl the homonymous be still
does service as a form for the infinitive, pres. ind. pi, pres. subj., the im-
perative and the past participle. With its disappearance from the indicative
paradigm, not yet entirely completed even in EModE3, the distribution of
ö-forms becomes 'tidier' in that after be ceases to be a pres. ind. form,
the only high-frequency use left to it is its use as the infinitive.
A thing that strikes the eye in the plural indicative is the fairly large
number of instances in which is occurs with a plural subject; for an ex-
ample see (9) below. This exemplifies the greater freedom EModE writers
had in matters of congruence; neither is the possibility of dialectal influ-
ence excluded.

3. The functions of be

3.1. The three main functions

The status of be throughout the history of English is threefold: in de-

scending order of frequency, the verb is used as a lexical verb in copular
uses, e.g.
 The verb be 103

(7) Bicause I am ignoraunt, I would learne,... {State Trials, The

Trial of Sir Nicholas Throckmorton I 72.C1)

(8) Tythes are tributes or rewardes to nedye soules. (Fitzherbert,

The Book of Husbandry 37)

(9) and all my brethern and sisters is in good health (William

Plumpton, Plumpton Correspondence 176)

(10) And doest thou think that such thynges as suffisaunce, and
power be, are to be dispysed (Colville, Boethius' Consolation of
Philosophy 68)

as an auxiliary in active and passive constructions, e.g.

(11) I am sure he is gone a weye with thy capo~s. (A Hundred Mery

Talys 146)

(12) Therefore power, is desired, for it is thoughte also to be good.

(Colville, Boethius' Consolation of Philosophy 76)

(13) Martha assone as she hearde that Iesus was comynge, went and
met him (Tyndale, The New Testament John 11 20)

and finally as a lexical verb in non-copular uses, most notably in exist-

ential sentences like (14) and (15):

(14) But certes by nature, there is but one God (Colville, Boethius'
Consolation of Philosophy 75)

(15) And next vnto this panicle there is another pannicle called Pia
mater, or meeke mother (Vicary, The Anatomie of the Bodie of
Man 29)

(16) 'And hastow wel knowen the causes,' quod sehe, 'whi it esT
(Chaucer, Boethius' De Consolatione Philosophiae 429.C1)
'And have you properly grasped the causes,' she said, 'why it is
the case?'
104 Matti Kilpiö

3.2. Copular and auxiliary uses causing problems of definition

The statistically marginal use of be as a lexical, non-copular verb remains

clearly defined throughout the long diachrony. This is not, however, the
case with all types of copular and auxiliary uses. In the following areas
there have been changes in the category of verb which be represents, or,
at least, scholars have not been in total agreement as to whether the status
of be has changed during the centuries:

(a) Be as a tense auxiliary used in forming perfect and past perfect forms
of intransitive verbs. — This use, attested from OE onwards, is exemp-
lified by (17):

(17) ...ίοφοη hi ne beod of godum mode cumene

{Gregory's Dialogues MS C 76)
'because they have not come from a good mind'

In his discussion of constructions of this type, Mitchell (1985: §§734-

740) does not take a stand as to whether we have here a copular construc-
tion with the past participle cumene predicatively used or whether beon
here has already acquired auxiliary status, forming a verb phrase with the
past participle.
A similarly ambivalent attitude towards the OE constructions is adopted
by Mustanoja (1960: 499). From his discussion it can be gathered, how-
ever, that he regards similar constructions in ME as instances of a verb
phrase where be functions as an auxiliary. The decision adopted in the
present study is to treat even OE constructions of this type as verb phrases
in which beon is a tense auxiliary or tense marker (for the subsequent de-
velopment, see Kytö in this volume).

(b) Combinations of be and a present participle of a lexical verb. — This

use, represented by progressive forms like (13) above, partly originates in
OE copular constructions in which the present participle of a lexical verb
is used as a subject complement (for theories on the development of the
ModE periphrasis, see Mustanoja 1960: 586-589 and CHEL 2: 251-256).
Here is an example of the OE periphrasis when it is most purely verbal:
 The verb be 105

(18) Jiast is Jaonne J>aet we sceolon jjone geleafan & J>aet ondgit mid
godum dasdum gefyllan, Jjonne beo we urum Haelende fylgende
{The Blickling Homilies 23)
That means then that we must fulfil the faith and the under-
standing with good deeds: then are we following our Saviour.'

What is problematic with many of the OE examples is, however, that it is

very difficult to decide the exact status of the construction. In fact, many
of them, unlike (18) above, can be more plausibly interpreted as copular
constructions in which the present participle functions as subject comple-
ment, while in others both the interpretation as a copular construction and
as a verb phrase offer themselves as equally plausible alternatives. The
OE 'expanded forms', as Nickel 1966 calls them, have therefore been
classified as copular constructions in the present study, but combinations
of be + present participle from ME onwards have been listed under the
auxiliary uses of be.

(c) The be to construction. — This construction, met in English from OE

onwards is, from the point of view of PresE, a modal idiom expressing
futurity, with varied connotations of 'compulsion' etc.6 (Quirk et al.
1985: 143, 217, 236). Although Quirk et al. regard this construction as a
marginal auxiliary, the be to construction occurring in the present data has
been systematically treated as a copular construction with the infinitive as
subject complement.

(d) The be about to construction. — A ME development (for a discussion

see Mustanoja 1960: 353-354), this construction expresses near futurity.
Quirk et al. include this construction among the marginal auxiliaries; in
the present study it has been treated as a copular construction just like the
be to construction.

3.3. Developments in the functions of be

In his discussion of the use of be as a tense auxiliary, Mustanoja makes

the following remark: "It is perhaps not without significance that while be
is becoming an auxiliary par excellence of the passive voice, it is losing
ground as an auxiliary of the perfect and pluperfect tenses" (Mustanoja
1960: 501). There is indeed good reason to assume that developments in
106 Matti Kilpiö

different parts of the verbal system are not separate and autonomous but
interdependent.

3.3.1. Chronological trends in the relative share of the main functions of

be: a survey of present tense forms from OE to EModEl

Table 6 presents an overview of the relative share of the three main uses
of the verb be from Old to Early Modern English in the present tense,
indicative and subjunctive. 7

Table 6. Relative share of the three main functions of be in OE, ME1, ME3
and EModEl in the select corpus (every 10th instance): present indica-
tive and subjunctive.

OE ME1 ME3 EModEl

Ν % Ν % Ν % Ν %

Copular uses 527 65% 176 68% 276 64% 235 64%
Auxiliary uses 231 28% 59 23% 119 28% 103 28%
Main verb, non-copular 56 7% 22 9% 34 8% 31 8%

Total instances 814 257 429 369

The overall impression gained from the statistics in Table 6 is one of

great stability in the relative share of the different functions of be
throughout the periods studied. For a discussion of the implications of
this, see section 3.3.

Table 7. Types of auxiliary uses.

OE ME1 ME3 EModEl

Ν % Ν % Ν % Ν %

Passive auxiliary 202 88% 52 88% 95 80% 93 90%

Pass, or tense auxiliary 19 8% 1 2% 7 6% 4 4%
Tense auxiliary 10 4% 6 10% 12 10% 6 6%
Progressive (ME-) - — 0 0% 5 4% 0 0%

Totals 231 59 119 103

 The verb be 107

Table 7 gives the breakdown of the auxiliary uses of be in the periods

studied, throughout which the use of be as a passive auxiliary is the most
important. With rare constructions like the progressive8 the method of
sampling adopted here clearly involves a random factor. As the corre-
sponding eight OE examples have been classified as copular construc-
tions, only ME3 contains examples of the progressive (see, however,
Table 9 below for EModEl instances of the progressive).

Table 8. Quasi-auxiliary uses of be included in the copular instances.

OE ME1 ME3 EModEl

Be to 19 3 3 7
Be about to — 1

Percentage of be to constructions of all copular constructions:

OE 3.6%, ME1 1.8%, ME3 1.1%, EModEl 3%.

As can be seen, neither the be to construction, illustrated above by (10),

nor the be about to construction are common in any of our periods. The
rather high OE percentage is evidently due to the commonness in OE of
the (particularly deontic) construction of the type seen in (19):

(19) Nu ge habbaö gehyred anrasdlice hwaet eow to donne is and

hwaet eow to forgone is. (^lfric, Letter to Wulfsige 34)
'Now you have heard definitely what you are to do and what
you are to abstain from.'

The rise of be to in EModEl after being rather dormant in the two ME

periods is in accordance with Mustanoja (1960: 524), who says that the
"construction is comparatively infrequent in OE and early ME, but be-
comes more common in later ME and early Mod.E."

3.3.2. The relative share of the main functions in ME3 and EModEl in
finite past tense forms and in non-finite forms

As stated above on p. 106 in connection with the statistics presented in

Table 6, the overall impression gained from a survey of the relative share
of the three main functions of be is one of great stability. Particularly
108 Matti Kilpiö

with regard to the transition from Middle English to Early Modern Eng-
lish this runs counter the expectations that the relative share of auxiliary
uses at the expense of the remaining two uses would rise. It is for this
reason that I here supplement the information provided by Table 6 by
considering the main functions of be in past tense forms and in non-finite
forms of be for the last two subperiods, ME3 and EModEl.

Table 9. Relative share of the three main functions of be in a select corpus

(every 8th instance)" of past tense forms of be in ME3 and EModEl
and the breakdown of the auxiliary uses between different types of
auxiliaries.

Function of be ME3 Ν % EModEl Ν %

Copular 120 54% 126 54%

Auxiliary 82 37% 91 39%
Main verb, non-copular 19 9% 16 7%

Totals 221 100% 233 100%

Auxiliary uses
- Passive auxiliary 71 81
- Passive or tense auxiliary 4 3
- Tense auxiliary 4 4
- Progressive auxiliary 3 3

Total auxiliaries 82 91

" Every 8th, not every 10th instance was analysed here. The solution adopted
was purely practical since the structure of the WordCruncher program makes the
selection of every 8th example speedier and more mechanical than 10th.

There are immediately obvious features of the relative shares of the main
functions of be in the light of Tables 6 and 9. The first is that with past
tense forms, auxiliary uses are relatively more common and copular uses
correspondingly less common than with the present tense forms in sub-
periods ME3 and EModEl. The difference is of the same magnitude in
both subperiods studied as appears from the following juxtaposition of
percentages. The figures before the slash ( / ) give the percentage met in
the present, the one after the slash the percentage met in the past tense:
 The verb be 109

ME3: copular 64% / 54%; auxiliary 28% / 37%; main verb, non-
copular 8% / 9%.
EModEl: copular 64% / 54%; auxiliary 28% / 39%; main verb, non-
copular 8% / 7%.

There is no obvious explanation for the difference between the present

and past tense of be with regard to the relative frequency of the main
uses of the verb.
The second noticeable thing that emerges from Table 9 is that when we
move from ME3 to EModEl there are no big changes in the relative pro-
portions of the main uses; thus the addition of the past tense to the survey
does not change the picture of relative stability gained from a study of the
present tense forms.
Table 9 also shows the distribution of past tense forms of be between
different auxiliary uses. The figures can be compared with those in Table
7; they show a similar kind of breakdown where the use of be as a pas-
sive auxiliary is preponderant (87% of the ME3 and 89% of the EModEl
auxiliary instances of was, were, etc. represent the passive auxiliary).
The data in Table 10 consists of those non-finite forms of be, infini-
tives, past participles and -ing forms in ME3 and EModEl which are used
in verb phrases so as to represent one of the three main functions of the
verb be. Thus, to give a couple of examples, in (20) the infinitive has the
function of a non-copular main verb, in (21) the past participle has the
function of a passive auxiliary and in (22) the -ing form, a present parti-
ciple, functions as a copula:

(20) Lat be soche falsheed; {The Cloud of Unknowing 23)

'Let such falsehood be'

(21) But [in] this thing hath ben discoveryd to the [that] thow
seydest that thow wistest not a litel herbyforn
(Chaucer, Boethius' De Consolatione Philosophiae 436.C2)
'But in this thing has been revealed to you what you said that
you did not know a little before this time'

(22) ye mylner beyng w'yn asked who was ther (A Hundred Mery
Talys 36)
'the miller, being within, asked who was there'
110 Matti Kilpiö

Table 10. Relative share of the main functions of be in a select corpus (every
10th instance) of nonfinite forms of be.

ME3 Ν % EModEl Ν %

Infinitive
Copular 43 40% 42 37%
Auxiliary 57 52% 66 58%
Main verb, non-copular 9 8% 6 5%

109 100% 114 100%

Past participle
Copular 7 54% 12 60%
Auxiliary 5 38% 7 35%
Main verb, non-copular 1 8% 1 5%

13 100% 20 100%
-ing form
Copular 2 100% 13 68%
Auxiliary 0 0% 6 32%
Main verb, non-copular 0 0% 0 0%

2 100% 19 100%

All non-finite forms

Copular 52 42% 67 44%
Auxiliary 62 50% 79 52%
Main verb, non-copular 10 8% 7 4%

Total forms analysed 124 100% 153 100%

Compared with the division of finite forms of be between the different

functions of the verb set out in Tables 6 and 9 above, the breakdown of
nonfinite forms seen in Table 10 again presents a different kind of pic-
ture. As only infinitives are represented by a large enough number of in-
stances to enable us to make reliable statistical comparisons between the
two periods examined, the main focus will be on this non-finite form. It
is worth noticing that with infinitives, both in ME3 and in EModEl, the
auxiliary uses are the most common function of the infinitive (52% of the
infinitive instances in ME3, 58% in EModEl). In the infinitive there is also
a rise in the percentage of auxiliary uses when we move from ME3 to
 The verb be 111

EModEl. By the same token, the relative shares of copular and non-
copular main verb uses decrease in EModEl compared to ME3.
It is interesting to note that the finiteness or non-finiteness of the form
of be affects its distribution between the copular and auxiliary uses. This
is understandable in view of the general tendency to increase three-verb
groups in 16th century English and the natural avoidance of non-finite
copulas of the type illustrated by example (22).9
Of the auxiliary instances, the majority represent the passive auxiliary
both in ME3 and EModE. Thus of the 57 infinitives used as auxiliaries in
ME3 56 are passive auxiliaries and one is a tense auxiliary; all five past
participles with auxiliary function represent the passive auxiliary. In sub-
period EModEl, all the non-finite forms with auxiliary function represent
the passive auxiliary. This confirms the picture gained from finite forms
of be functioning as an auxiliary.
One feature in Table 10 that points the way to future developments is
the great increase in the number of occurrences of being used either as a
gerund or as a present participle when we move from subperiod ME3 to
EModE3.10 This expansion naturally paves the way to the enrichment of
the morphology of the progressive.

3.4. Linguistic factors

As was seen above, the general impression from the statistics of Table 6
is one of great stability, the relative shares of copular, auxiliary and non-
copular main verb uses remaining almost the same over some 800 years.
As far as the copular and non-copular main verb uses are concerned, this
should cause no great surprise since in any language there must be means
to express relations of being and of existence available and it could be
hypothetized that the need for these means may remain fairly constant
over long periods of time. It has also been seen that the addition of past
tense forms of be to the survey for the last two periods studied did not
alter the picture of essential stability.
With the auxiliary uses the situation seems to be more complex. In the
analysis of the three main functions of be one of my working hypotheses
was that the relative share of auxiliary uses would grow the closer we got
to PresE. The hypothesis was based mainly on the assumption that the re-
lative share of passive constructions in comparison with copular uses and
non-copular main verb uses would grow in EModEl through the loss of
the ME indefinite agent me(n) (cf. Görlach 1976: 110); however, the
112 Matti Kilpiö

statistics in Tables 6 and 9 (finite present and past tense forms of be)
show no real move in this direction. When non-finite forms are consider-
ed (see Table 10 above), some changes can be observed, notably in the
infinitive. With this form the auxiliary uses are the most important func-
tion of be and there is a perceptible rise in the percentage of auxiliary
uses when we move from ME3 to EModEl.
Still, it may well be that EModEl is a period which is too early for us
to note a considerable rise in the relative share of auxiliary uses. The pro-
gressive form, which uses be as its auxiliary, is rare throughout the
EModE period (Görlach 1976: 109; note also the very low number of pro-
gressive forms among the coded EModEl auxiliary instances, comprising
only three instances among the past tense forms, see Table 9). A notice-
able rise in the relative share of the auxiliary uses of be is to be expected
only when the progressive form really sets in and is also established in
compound tenses.
Unlike the use of be as the auxiliary of the progressive, its use as a
tense auxiliary (see example (11) above) is a receding feature. Although
have had been gaining ground at the expense of be with mutative intransi-
tives since the late Middle English period, the decisive steps in the pro-
cess where have superseded be were taken in the 18th and 19th centuries
(see Kytö in this volume). It is not possible to show any specific link
between the be/have variation with intransitives and the small changes in
the distribution of the different uses of be seen in Tables 6, 7 and 9. One
possible conclusion to be drawn is that considering the preponderance
throughout the periods studied (1) of the copular uses of be and (2) of the
passive auxiliary uses of be compared to all the remaining auxiliary uses,
it becomes evident that the use of be as a tense auxiliary is marginal and
one which thus was perhaps more liable to be superseded by the have
perfect than not. In considering the reasons why have ousted be, Olga
Fischer (CHEL 2: 261-262) mentions

(a) the greater functional load of be (passive, progressive and perfect

auxiliary) compared to have (only one auxiliary use, that of the perfect
auxiliary in ME) as well as ambiguities arising from the double function
of be, as passive and perfect auxiliary and from the confusion arising
from the occasional writing of the -ing of the progressive as -en and

(b) the rise of be to the status of the passive auxiliary par excellence after
the loss of wurthen.
 The verb be 113

All these seem to be valid considerations; to the functional load of be

should probably be added its most important use throughout the periods
I have considered, namely its various copular uses. It could be argued that
the only use of be comparable to the lexical verb have is its use as a
lexical, non-copular verb.

3.5. Extralinguistic factors

Although, as the discussion in section 3.1 above shows, the relative shares
of the three main uses of be show remarkable stability over a long time
span, there are small changes which are at variance with attempts to
interpret long-term changes as unidirectional. A case in point is the devel-
opment seen in the copular and auxiliary uses in the long diachrony. As
already pointed out, the relative share of copular and auxiliary uses
remains steady from OE to EModEl except for a slight rise of copular and
a corresponding decline of auxiliary uses seen in ME1. With due al-
lowance for the low number of instances analysed, probably too low for
definite conclusions, it is still possible to hypothesize as to why this
should be the case. The use of be as a passive auxiliary is by far the most
important auxiliary use of this verb in all the periods considered. It would
seem that the very fact that so much of OE represents the category of
translation or even glossing — in the number of running words, the
combined share of these two categories rises to c. 41% of all the OE texts
in the HC — would elicit from OE writers more passive forms and thus,
indirectly, would increase the relative share of the auxiliary category than
would be the case if they were writing an original text independent of
Latin. Compared to the OE texts in the HC, ME1 writings do not contain
much translated material. The only exception that I am aware of is
Vespasian Homily 3, a late version of the Diets of Cato included in OE4.
This text represents only c. 2% of the total number of words in ME1.
Thus it can be assumed that native tendencies in the preference for the
active over the passive voice can more easily come to the fore in the ME1
subcorpus, thus making the proportionate share of auxiliary uses of be
less. This kind of reasoning makes the tacit assumption that if the selec-
tion of texts found in the Old English part of the HC relied as little on
translation as does that of ME1, the share of copular constructions would
be equally high or higher than that attested for ME1.
In ME3, the share of translations of the total of running words again
rises to c. 22% and in EModEl to c. 16%. Whether the rise in the
114 Matti Kilpiö

percentage of auxiliary uses attested in ME3 is due to the extralinguistic

factor of Latin originals influencing English translations or whether it is
rather due to purely linguistic factors is difficult to gauge. The main thing
to note here is that once the relative proportions of copular and auxiliary
uses attested in ME3 are reached they largely remain at the level of ME3
through EModEl if the development in the infinitive from ME3 to EModEl
is ignored.

4. Conclusions

In this study I have examined two distinct areas connected with the verb
be: developments in its morphology and developments in its functional
load from Old English to Early Modern English.
On a superficial level of analysis, the morphology of be seems to
undergo a development in which there is an increasing diversification of
forms as Middle English is reached and a subsequent process of sim-
plification and regularization as Early Modern English is reached. This is
an oversimplification, as it is quite evident that if the Old English lan-
guage had been handed down to us in a more balanced way, without the
overpowering dominance of the late West Saxon standard in its transmis-
sion, and if we indeed had texts from all Old English dialects from all the
periods, the picture gained of the morphology of be would be more
varied. At the other end of the time span of this study, the EModEl sub-
period suddenly shows a very small number of forms for be, although the
distance in time between the end of ME3, 1420, and the beginning of
EModEl, 1500, is only eighty years. Once again it would be wrong to
conclude that the proliferation of forms seen in ME3 had suddenly come
to an end. It is only that the wealth of dialectal variation still present in
the language in the Early Modern English period is not caught by the
Helsinki Corpus, and quite intentionally so, as one of the principles in
compiling the Early Modern British English part was the standardness of
texts.
All this does not mean that there are no changes in the development of
be during the periods studied here. The most important change when we
move from Old English to Middle English is the rapid loss of the two co-
existing present paradigms, the one with initial b- and the one without ini-
tial b-, both in the indicative and subjunctive. This led to different kinds
 The verb be 115

of suppletive paradigms combining b- and ηοη-ό-forms in various ways.

Some Old English forms, e.g. δ-forms in the singular, and ί-forms in
general, quickly became vestigial and all but disappeared after the transi-
tion from Old to Middle English. In Middle English Scandinavian influ-
ence can also be seen in some present tense forms of be. Some of the
functions carried by the (none too clear) distinction between OE b- and
non-6-forms had to be taken over by other expressions, a case in point
being the replacement of the futuric use of the /»-forms by the periphrastic
shal or wil construction.
Another important development, pointing the way to a 'tidier' system
than that prevailing in Middle English, is the reduction of forms seen in
the rising standard of Early Modern English. With one exception, that of
plural be(en), forms with initial b- have disappeared from the present in-
dicative; even be(en) is in the process of being ousted by ar(e).
The second main topic of this study was developments in the functional
load of be. Here a survey of present tense forms, indicative and subjunc-
tive, from OE to EModEl, showed that the relative share of each of the
three main functions of be, copular, auxiliaiy, and non-copular main verb
uses remains very stable over some 800 years. Past tense and non-finite
forms were subjected to a similar survey for the last two periods consid-
ered here, ME3 and EModEl. Like present tense forms, past tense forms
showed relative stability in the proportional share of the functions of be
in the transition from ME3 to EModEl. The non-finite forms, particularly
the infinitive, yielded the most interesting results. First of all, the infini-
tive be is an auxiliary in over 50% of all the infinitive occurrences in
both ME3 and EModE; secondly, the relative share of the auxiliary uses of
the infinitive increases with the transition from ME3 to EModEl. Thus, as
pointed out in section 3.3.2, the finiteness or non-finiteness of the form of
be affects its distribution between the copular and auxiliary uses, the
slight rise in the auxiliary uses in EModEl evidently reflecting the Early
Modern English tendency to expand the verbal group.
Within the auxiliary category, the predominance of the use of be as a
passive auxiliary in all the periods considered is the most noticeable fea-
ture. This purely quantitative consideration may partly explain why the
other complex forms with be as the auxiliary and the past participle of the
main verb, the compound tenses of mutative intransitives, gradually re-
place be by have.
116 Matti Kilpiö

Appendix

A comparison between pres. subj. sg. sie/beo and pres. subj. pi. sien/beon
in those parts of Waerferth's translation of Gregory's Dialogues for which
there exist two parallel versions, the original represented by MS C and the
revision represented by MS H. The page and line references are to Hecht
1900. The left-hand column gives the forms of MS C, the right-hand
column those of MS H. If no comment is added it is to be understood that
the forms correspond.

SINGULAR:

C has 1st pers. sg., Η has 3rd pers. sg.:

GD(C) GD(H)
109.23 sy [direct speech] 109.20 wasre (pret. subj.)
[indirect speech]
109.24 sy [direct speech] 109.21 waere (pret. subj.)
[indirect speech]
2nd pers. sg.:
GD(C) GD(H)
7.10 sy 7.9 eart (ind.)

3rd pers. sg.:

GD(C) GD(H)
5.32 sy 5.29 sy
8.6 byTp (ind.) 8.5 sy
8.23 sy 8.22 si
20.8 sy 20.7-8 Q)u...wite]
20.14 sy 20.14 is (ind.)
20.16 sy 20.16 si
23.18 sy 23.17 sy
25.20 sy 25.21 beo
28.3 sy 28.3 si
33.8 sy 33.8 sig
35.8 sy 35.9 sy
35.11 sy [no real corresp. with:] 35.12 sy
46.9 sy 46.8 beo
60.7 sy 60.7 sy
C: no counterpart for 61.7 si
 The verb be 117

C: no counterpart for 61.9 si

63.5 [we sceolon witan] 63.6 si
65.25 sy 65.25 si
75.32 sy 75.31 sig
76.29 sy [in an active construction] 76.28 sig [in a passive constr.]
C: no counterpart for 76.13 beo
C: no counterpart for 90.21 sy
102.19 sy 102.21 sy
105.31 sy 105.31 sy
107.19 sy 107.17 is (ind.)
128.20 sy 128.16-17 ne gewuröe j^set
138.3 sy -138.2 sy [roughly corresp.]
146.2 sy 146.2 sy
146.36 sy 146.36-147.1 hi...syndon [very
distant, or in fact no real
correspondence]
147.2 sy 147.3 hi syndon [free relationship]
157.2 sy 157.2 sy
172.31 sy 172.28 si

PLURAL:

1st pers. pi.:

41.5 syn 41.5 beon

3rd pers. pi.:

40.1 syn 40.1 beon
61.4 syn 61.4 beon
61.5 syn 61.5 [fremian]
61.9 syn 61.9 beon
76.31 syn 76.30 beon
87.4 syn 87.4 beon
94.22 sy 94.21 beon
102.29 syn 102.31 beon
102.31 syn 102.33 beon
137.2 syn H: no counterpart
138.32 syn 138.27 syndon (indie.)
139.31 syn 139.30-31 []>u...wite]
151.27 syn 151.24 beon
118 Matti Kilpiö

Notes

1. The (evidently) rare instances of his representing is have not been disam-
biguated for OE, ME3 and EModEl.
2. With the exception of sie forms from OE3 and OE4; of these, every 50th in-
stance was coded.
3. For the abbreviated titles of texts used in Tables 2 and 3, see Kytö (1996:
167—230). The final letter in the Helsinki Corpus dialect codings is L if
LALME is the source of the definition; otherwise the coding ends in O. Thus
EML, EMO = 'East Midland', WML, WMO = 'West Midland', NL, NO =
'Northern', SL, SO = 'Southern'.
4. The numbers of instances given in Table 2 represent absolute frequencies.
As the total numbers of words of texts representing MEX/1 is 39,050 and
that of ME1 texts 72,010, the numbers of instances for the two groups of
texts are not directly comparable.
5. For the abbreviated titles for texts and the dialect codings, see note 3
above.
6. Historically speaking, this construction is an amalgamation of two construc-
tions, the OE impersonal type expressing obligation or necessity (see ex-
ample (19), a type which changes into a 'personal' variant in ME, and the
OE type he is to cumenne 'he is to come' with personal subject; for a dis-
cussion, see CHEL 2: 336-337).
7. Imperative forms have been left out of account here, as they are basically
only met in copular, not auxiliary or non-copular main verb uses.
8. The progressive remains rare through the Early Modern English period, see
e.g. Görlach (1978: 108-109).
9. For a discussion of Early Modern English developments in the verbal
group, see Blake (1983: 81-103).
10. This impression is confirmed if we consider all the appropriate -ing forms,
both those coded and those left outside the analysis: ME3 contains no more
than 11 instances, EModEl as many as 183. Forms in -ing which have been
excluded from the discussion include (1) being as a non-verbal noun not
functioning in any of the main uses of be; (2) being in the set phrase 'for
the time being'.
 The verb be 119

References

Blake, N. F.
1983 Shakespeare's language. An introduction. London—Basingstoke:
Macmillan.
CHEL 2 = The Cambridge history of the English language. Volume II:
1993 1066-1476. Cambridge: Cambridge University Press.
Forsström, Gösta
1948 The verb 'to be' in Middle English: A survey of the forms. Lund:
Gleerup.
Görlach, Manfred
1978 Einführung ins Frühneuenglische. (Uni-Taschenbücher 820.) Hei-
delberg: Quelle and Meyer.
HC = Helsinki Corpus
1991 The Helsinki Corpus of English Texts. Helsinki: Department of
English, University of Helsinki.
Hecht, Hans (ed.)
1900 Bischofs Wcerferth von Worcester Ubersetzung der Dialoge Gregors
des Grossen. (Bibliothek der angelsächsischen Prosa 5.) Leipzig:
Wigand.
Kilpiö, Matti
1989 Passive constructions in Old English translations from Latin with
special reference to the OE Bede and the Pastoral Care. (Memoires
de la Soci6te N6ophilologique de Helsinki XLIX.) Helsinki: Soci6te
N6ophilologique.
1992 Dictionary of Old English: BEON. With attested spellings by Robert
Millar using material assembled by Haruko Momma. Toronto: Pon-
tifical Institute of Mediaeval Studies.
1993 "Syntactic and semantic properties of the present indicative forms
of the verb 'to be' in Old English", in: Matti Rissanen et al. (eds.),
97-116.
Kytö, Merja (comp.)
1996 Manual to the diachronic part of the Helsinki Corpus of English
Texts: Coding conventions and lists of source texts. (3rd edition.)
Helsinki: Department of English, University of Helsinki.
LALME = Mcintosh, Angus—M. L. Samuels—Michael Benskin—Margaret
Laing—Keith Williamson
1986 A linguistic atlas of late mediaeval English. 4 vols. Aberdeen:
Aberdeen University Press.
Mitchell, Bruce
1985 Old English syntax. I—II. Oxford: Clarendon Press.
120 Matti Kilpiö

Mustanoja, Tauno F.
1960 A Middle English syntax. Part I: Parts of speech. (M6moires de la
Societe N6ophilologique de Helsinki 22.) Helsinki: Societ6 Νέο-
philologique.
Nevalainen, Terttu—Helena Raumolin-Brunberg
1993 "Early Modern British English", in: Matti Rissanen et al. (eds.), 53-
73.
Nevanlinna, Saara—Päivi Pahta—Kirsti Peitsara—Irma Taavitsainen
1993 "Middle English", in: Matti Rissanen et al. (eds.), 33-51.
Nickel, Gerhard
1966 Die expanded Form im Altenglischen. Vorkommen, Funktion und
Herkunft der Umschreibung 'beon/wesan'+Partizip Präsens. Neu-
münster: Wachholz.
Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik
1985 A comprehensive grammar of the English language. London—New
York: Longman.
Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.)
1993 Early English in the computer age: Explorations through the Hel-
sinki Corpus. Berlin—New York: Mouton de Gruyter.
Samuels, M. L.
1985 "The Great Scandinavian Belt", in: Roger Eaton—Olga Fischer—
Willem Koopman—Frederike van der Leek (eds.), Papers from the
4th International Conference on English Historical Linguistics,
Amsterdam, 10-13 April, 1985 (Amsterdam Studies in the Theory
and History of Linguistic Science: Series IV — Current Issues in
Linguistic Theory 41.) Amsterdam—Philadelphia: Benjamins, 269-
281.
Re-phrasing in Early English: The use of
expository apposition with an explicit marker
from 1350 to 1710

Päivi Pahta and Saara Nevanlinna

1. Introduction

1.1. The aim of the study

The aim of our study is to examine the structure and use of nonrestrictive
expository apposition in Late Middle English and Early Modern English
texts in the Helsinki Corpus (HC). Originally we were interested in the
use of appositional constructions in different types of texts and in the fac-
tors motivating re-phrasing in the form of apposition, but since apposition
as such has not been researched in any detail in this period, we felt it ne-
cessary to discuss the structure of apposition as well. We approach the
problem by reflecting the present-day situation onto Late Middle English
and Early Modern English in order to see to what extent the outline of
modern usage was already established in those days. The focus of our in-
terest has been to trace the development of appositional constructions
containing an explicit marker during the centuries framing the transition
period between Middle English and Early Modern English (1350-1710).
We have also mapped the general situation in Old English and Early Mid-
dle English, which still very much reflects the Old English usage (see
Nevanlinna—Pahta forthcoming).
122 Päivi Pahta and Saara Nevanlinna

1.2. Earlier research on apposition

Although apposition is discussed in most works on contemporary English

grammar and linguistics, it has generally been done perfunctorily. 1 The
semantics of apposition is touched upon, e.g., in Halliday—Hasan (1976),
Lyons (1979), Bolinger (1979) and Jackendoff (1985). The syntactic char-
acteristics of apposition are discussed, e.g., in Matthews (1981), while
Huddleston (1971) and Quirk et al. (1972 and 1985) discuss fairly thor-
oughly both semantic and syntactic features of appositional constructions.
There are also a number of special studies, which include those by Hoc-
kett (1955), Sopher (1971), Burton-Roberts (1975) and Meyer (1987,
1989, 1991). Appositional reformulations as a stylistic device have been
discussed by Blakemore (1993). The most detailed discussions of the top-
ic so far are by Acuna (1996) and the corpus-based study of apposition in
contemporary English by Meyer (1992).2
Language historians have generally shown less interest in apposition
(cf. Strang 1970). Denison (1993) refers several times to interesting struc-
tures involving apposition (see General Index p. 526), but since he con-
centrates on verbal constructions, apposition remains only marginal to his
theme. Some aspects of apposition in Old English, either verse or prose,
have been discussed in a number of special studies, for instance, in Pel-
tola (1960), Robinson (1985) and Blockley (1989), as well as in the com-
prehensive Old English Syntax by Mitchell (1985). The use of repetitive
word pairs in Old and Middle English is discussed in Koskenniemi (1968)
and (1983), and the use of the parenthetic as who say and its variants in
Old and Middle English in Nevanlinna (1974) and (1993). Some apposi-
tional constructions used in the presentation of terminology in Middle
English medical prose are discussed in Wallner (1987), Jones (1989) and
Norri (1992).

1.3. Apposition as a grammatical relation

Apposition has been defined in a number of ways. The treatment of ap-

position is summarized in A Dictionary of Linguistics and Phonetics by
David Crystal (1985: 20), who states that "apposition(al)" is "a traditional
term retained in some models of GRAMMATICAL description for a se-
quence of units which are CONSTITUENTS at the same grammatical level,
and which have an identity or similarity of REFERENCE". He goes on to
add that there are "many theoretical and methodological problems in de-
 Expository apposition 123

fining the notion of apposition, because of the existence of several con-

structions which satisfy only some of these criteria, and where other SE-
MANTIC or syntactic issues are involved, as in titles and other desig-
nations".3
The narrowest type of definition is given by Fries (1952: 187), who
argues that the appositives must be coreferential, juxtaposed noun phrases.
Curme (1931) admits as appositions some other constructions, as well,
such as appositive genitives (p. 84, the vice of intemperance), appositive
adjectives (p. 93, the room above), and predicate appositives (p. 30, he
came home sick). In most instances these are juxtaposed. In another work
(1970) he recognizes an apposition in the form of an explanatory remark
comprising a whole dependent or independent clause. He enumerates co-
ordinative conjunctions connecting words, phrases or sentences and intro-
ducing an explanation, such as namely, to wit, that is (to say), for ex-
ample, etc., but he does not yet recognize them as apposition markers.
Jespersen (1961) has a fairly broad definition of apposition including
certain uses of reflexive pronouns (vol. VII, p. 172, you yourself must set
some tasks) and participles (vol. V, p. 406, he sat smoking). He also in-
cludes among the appositional constructions some clauses which are in
apposition with noun phrases (vol. Ill, p. 27, their idea that priests are in-
fallible).
Sopher (1971: 412) points out that the head group (the first appositive)
and the appositional group (the second appositive) may belong to differ-
ent grammatical categories, and that the morphological or structural fea-
tures are not relevant, since apposition is a syntactic category. He further
postulates that apposition is a grammatical category distinct from both
coordination and subordination, and that the appositional groups cannot
be linked by and or any other coordinating conjunction. Sopher presents
both notional and formal criteria. Both appositional groups have a single
referent, and are notionally interchangeable, so that the second apposition-
al group may replace the first without altering the syntactic structure of
the sentence. Where the introduction of an apposition marker is not pos-
sible, notional equivalence and the single referent may be tested by the al-
ternative criteria, namely the test of "interchangeability" or "replaceabil-
ity", and the "and" test.
Huddleston (1971: 251-5) argues that in transformational grammar
there is a tendency to use the term "appositive" in the sense of "nonre-
strictive", though traditional grammar uses "apposition" for both restric-
tive and nonrestrictive constructions. He distinguishes three main types of
apposition: (i) Characterizing apposition derived from an underlying rela-
124 Päivi Pahta and Saara Nevanlinna

tive. No apposition marker is used in this type, (ii) Identifying or equative

apposition, in which the second (subordinate) member may be a depend-
ent or independent clause. The optional markers that is, viz., namely, etc.,
can be used, (iii) Narrowing apposition, which includes exemplification
and particularization. (See also Burton-Roberts 1975 and Matthews 1981.)
Quirk et al. (1985: §17.65 ff.) treat apposition primarily as a relation
between noun phrases, "constituents of the same level", although they
also refer to appositive postmodification by /Aa/-clauses (§17.26, the news
that his team had won), appositive nonfinite clauses (§17.35, the appeal
to give blood), and appositive prepositional phrases (§ 17.47, the news of
the team's victory). According to Quirk et al. (1985: §17.65) the apposi-
tives are either identical in reference or the reference of one is included
in the reference of the other. Apposition may be full or partial, strict or
weak, and restrictive or nonrestrictive, and appositional constructions are
combinations of these various categories. In full apposition the second
appositive is omissible without the sentence becoming unintelligible,
whereas this is not so in partial apposition. In strict apposition the two ap-
positives belong to the same syntactic class, whereas in weak apposition
they belong to different syntactic classes. In restrictive apposition the
second appositive is necessary for the identification of the first. In nonre-
strictive apposition, the appositives are in separate information units, pro-
viding relatively independent information. The first appositive normally
acts as the defined expression and the second has a defining role. In Pres-
ent-day English, the defining role of the second appositive is reflected by
the fact that it is marked as parenthetic by punctuation or intonation.
Quirk et al. (1985: §§17.74-17.87) also classify appositions into various
semantic classes ranging from different degrees of equivalence (appella-
tion, identification, designation and reformulation) to attribution, and in-
clusion (exemplification and particularization).
In recent discussions of apposition, both Acuna (1996) and Meyer
(1987: 101 and 1992: 3) argue that all earlier definitions of apposition,
taken individually, are inadequate or incomplete. Most definitions are too
narrow, whereas some admit as appositions constructions which should
not be included (e.g., Curme's category of appositive adjectives, see p.
123 above). Meyer points out that the difficulties which linguists and
grammarians have had in defining apposition are caused by the fact that
they regard it as primarily a syntactic relation. Meyer argues that apposi-
tion is not merely a syntactic relation; it must also be viewed as a seman-
tic and pragmatic relation, and should be defined in terms of those char-
acteristics which distinguish it from other grammatical relations — com-
 Expository apposition 125

plementation, modification, parataxis, peripheral elements, and coordina-

tion. In Apposition in contemporary English, M e y e r (1992) provides a
comprehensive analysis o f the syntactic, semantic and pragmatic charac-
teristics of apposition based on his study of appositions in three computer
corpora of English. 4 H e views apposition as a grammatical relation hav-
ing various realizations, which are described in Table 1, adopted from
Meyer (1992: 6).

Table 1. The linguistic characteristics of units in apposition (Meyer 1992: 6).

Syntactic Semantic Pragmatic

characteristics characteristics characteristics

Syntactic form: Semantic relations: Information

nominal apposition, coreference, cataphoric structure:
NP/clause apposition, reference, part/whole new or partially
appositions with relations, synonymy, new information
obligatory markers, hyponymy, attribution in second unit of
non-nominal apposition apposition
Semantic classes:
Syntactic function: identification, Functional
subject, object, appellation, potential:
complement, adverbial exemplification, tendency of some
particularization, appositions to
Linear structure: characterization, occur more
single/double/ paraphrase, commonly in
triple apposition, reorientation, self- some contexts
juxtaposed/ correction than others
unjuxtaposed
apposition Semantic integration:
restrictive/
Hierarchical nonrestrictive
structure:
binary/non-binary
apposition
126 Päivi Pahta and Saara Nevanlinna

In this study we have adopted the outline of Meyer's definition of apposi-

tion as a syntactic, semantic and pragmatic relation with some modi-
fications. We decided to adopt this rather broad definition of apposition
for two reasons. Working on historical data, we did not want our theoreti-
cal framework to impose preconceived restrictions on our findings. Hav-
ing worked through the data, we felt that from a functional point of view
we could not exclude non-nominal phrases, clauses or sentences from our
discussion. Syntactic variation in the form of the units in apposition did
not seem to affect the linguistic function of expository apposition, i.e.,
that of providing additional information.
Some of the syntactic and semantic relations that qualify as appositional
according to Meyer's definition remain outside the context of this study.
We only deal with nonrestrictive apposition. In restrictive apposition the
second appositive restricts the reference of the first, while in nonrestric-
tive apposition the second appositive provides relatively independent in-
formation about the first, explaining, illustrating or naming an object,
person or idea expressed in the first appositive to which it is notionally
equivalent and to which it is or may be linked by an explicit apposition
marker.5 We are also exclusively concerned with those nonrestrictive ap-
positional constructions where the appositives are linked by such markers.
We have further restricted this study to expository appositions which dis-
play what Quirk et al. call "a relation of equivalence" (see p. 124 above).
Expository apposition includes instances of coordinative apposition,
where the units of apposition are joined by and or or. Apposition resem-
bles coordination in that the two units are typically constituents of the
same level. The units of coordinative apposition are synonymous, corefer-
ential or related by cataphoric reference. In coordinative apposition no
reduction is implied, whereas, for instance, non-appositional (segregatory)
coordination can be regarded as implying the reduction of two or more
coordinated clauses (Quirk—Greenbaum 1973: §7.21, §9.41).
Table 2 shows the linguistic characteristics of the type of apposition to
be discussed in this study.
 Expository apposition 127

Table 2. The linguistic characteristics of nonrestrictive expository apposition

with an explicit marker discussed in this study.

Syntactic characteristics

Syntactic form:
nominal apposition,
NP/clause apposition,
non-nominal apposition

Syntactic function:

subject, object, complement, adverbial, verb

Semantic characteristics

Semantic relations:

coreference, cataphoric reference, synonymy

Semantic classes:

identification, appellation, characterization, paraphrase, revision

Pragmatic characteristics

Generic distribution:
tendency of some structures to occur more commonly in some contexts
than in others
1.4. Apposition markers

In Present-day English numerous expressions are available for marking

nonrestrictive apposition (see Quirk et al. 1985: §17.73). Markers of ap-
position are used to explicitly express the semantic relationship existing
between the appositives. They indicate whether the information supplied
by the second appositive is expository or exemplifying. 6 In an expository
relation, where the appositives are referentially or semantically equivalent,
128 Päivi Pahta and Saara Nevanlinna

i.e., they present the same truth value (Leech 1988; Allwood et al. 1977),
markers such as that is (to say), namely or in other words can be used in
Present-day English to introduce the second appositive. In an exemplify-
ing relation, where the units are referentially and semantically only par-
tially equivalent, i.e., the reference of one is included in the reference of
the other, markers such as for example, e.g. and say can be used in
Present-day English (Meyer 1992: 97).
Markers of apposition can be either obligatory or optional (see Meyer
1992: 25-26). The use of obligatory markers is governed by semantic and
syntactic reasons and their omission leads to an ungrammatical construc-
tion. An obligatory marker of apposition is required, e.g., in exemplifying
apposition, to unambiguously mark the part/whole relation which exists
between the appositives. In the use of optional markers of apposition,
pragmatic considerations are essential. The only marker of expository ap-
position which Meyer has classified as obligatory is or. The other ex-
pository markers included in his study (that is, that is to say, i.e., viz,
namely, and in other words) are classified as optional. The use of and as
an apposition marker is not discussed by Meyer at all, although and is in-
cluded in the list of apposition markers in Quirk et al. (1985: §17.73).
In Meyer's study of Present-day English usage, optional markers of ex-
pository apposition were fairly rare, occurring in less than three percent
of the appositional constructions in the corpora. The majority of the op-
tional markers in his study appeared in the learned genres of the corpora,
which suggests that they are indicators of formal style and would accord-
ingly be inappropriate in less formal Present-day English styles such as
fiction, press reportage or spontaneous conversation (Meyer 1992: 96-98).
One of the aims of this study is to see whether this could be true for
earlier periods of English as well.
The device of using a special marker to indicate apposition is found in
Old English, though appositional constructions without a marker are more
common.7 Apposition was not uncommon in Old English as a stylistic
device, modelled on Latin rhetoric (cf. Bloomfield 1965: 6). Robinson
(1985: 60) points out that in Beowulf apposition relating to a particular
subject serves as a means of emphasis. It is also a "retarding device" that
enables the reader "to consider an object as an action from more than one
perspective", as well as making the poet "move on swiftly and easily".8
The markers of nonrestrictive expository apposition that occurred in our
corpus are listed in Table 3. The list is somewhat different from the cor-
responding list of markers in contemporary English (cf. Quirk et al. 1985:
§17.73 and Meyer 1992: 26, 97). One of the commonest present-day
 Expository apposition 129

markers, in other words, did not occur in our data. The earliest example
of the phrase in the Century of Prose Corpus (1680-1780) is from 1690,9
which indicates that it appeared towards the end of the period we have
studied. On the other hand, the list contains some markers which are no
longer used, such as yea, variants of as who say, i. (an earlier variant of
i.e.), and a number of unique markers. Some of the unique markers were
detected quite accidentally, which suggests that other similar phrases
could perhaps be found in contemporary texts. Furthermore, particularly
in the Middle English sections of our corpus, the markers were mani-
fested in a variety of different spellings. In Table 3, the markers occurring
several times are given in their Present-day English form, whereas the
unique markers are given in the form in which they were found in the
data.

Table 3. The markers of expository apposition in the Late Middle English and
Early Modern English subperiods of the Helsinki Corpus.

Markers occurring several times Unique markers

and, & pat is to menynge

or (else/rather) that is to meane
(and) that is this to seyn as who seith
that/this/which/it is to say and it is this
(that is) to wit pat ben pese
yea (and) as moche to say as
as who say/saith that is as myche to say as
(and/as) namely which is to say
(that is) to understand by interpretation
nay which is to say
as being interpreted
scilicet, sc. vel
id est, i.e., i.
viz.

In Present-day English, some indicators of apposition can either precede

or follow the second appositive, though structures of the latter type are
very rare.10 In our data we found no examples of apposition where the
marker was placed after the second appositive.
Words used as apposition markers are not necessarily restricted to ap-
positional use only. The apposition markers and and or are also common-
ly used as explicit coordination markers. Gradual semantic change may
130 Päivi Pahta and Saara Nevanlinna

also cause changes in the use of an apposition marker, as with namely,

originally used as a marker of particularization. Its use in the modern
function dates from the period surveyed in this study.

2. Classifying appositional constructions

Appositional constructions can be classified in a number of ways which

focus on their various syntactic, semantic or pragmatic characteristics,
some of which we concentrate on here. Attention is primarily paid to the
use of the explicit markers and the semantic classes of expository apposi-
tion and their distribution across different types of text. First the syntactic
form and function of units in expository apposition will be briefly dis-
cussed.

2.1. Syntactic form and function of apposition

Defining apposition in the way we have done, a wide range of syntactic

constructions can be admitted as appositions, including phrases, clauses,
or sentences. Phrasal appositives may be, to list the central constructions,
noun phrases, verb phrases, prepositional phrases, predicative phrases, ad-
verb phrases or adjective phrases. Appositional clauses may be predi-
cates, dependent or independent clauses.
The appositions in the corpora consisted of two or, rarely, more apposi-
tives which had the following syntactic forms: nominal apposition (noun
phrases in apposition), as in example (1), a noun phrase in apposition
with a clause or sentence (2), or non-nominal apposition, in which the ap-
positional units are clauses, sentences or non-nominal phrases (3)—(4).

(1) And J requyre and byseche alle suche that iynde faute or errour
that of theyr charyte they correcte and amende hit.
(Caxton, The Prologues and Epilogues 78)
'And I require and beseech all those who find fault or error that
they should correct and amend it.'
 Expository apposition 131

(2) The office of a tutor is firste to knowe the nature of his pupil,
that is to say, wherto he is mooste inclined or disposed, and in
what thyng he setteth his most delectation or appetite. (Elyot,
The Boke Named the Gouernour 24)

(3) It is harde, ye, it is impossyble, that a man may have alle joye
in this worlde (In Die Innocencium 12)
'It is hard to believe, or rather, it is impossible, that a man may
enjoy everything in this world'

(4) And be it further enacted, That all Justices of the Peace ... find-
inge any such Oxen ... shall take and seize the same as forfeited,
and shall give and distribute the same to Prisoners and other
poore Folkes by theire discretion. (The Statutes of the Realm IV
1058)

Nonrestrictive appositional structures can also be incomplete in that

there is ellipsis in the second unit (cf. Quirk et al. 1985; Halliday—Hasan
1976; Matthews 1981). The term ellipsis is used here in the sense of
grammatical reduction (Quirk et al. 1985) or contraction (Matthews
1981). The parts implied can be inferred from the context; in other words,
they can be recovered from the first unit, where they are expressed in the
same form. The defective second unit may thus contain only what is ad-
ded to the first appositive, as in examples (5) and (6). In the examples the
parts omitted are indicated in square brackets by way of illustration.

(5) Laborande in a febre efimera, sc. [a febre efimera] of hotnesse,

if vertue and age suffre, mynusche he of cephalica of J>e rijt
arme in somere, in wynter of J)e lifte. (A Latin Technical Phle-
botomy 41)
'For treating ephemeral fever, i.e., (ephemeral fever) caused by
heat, if the condition and age of the patient allow, let blood of
the cephalic artery of his right arm in the summer and of his left
arm in the winter.'

(6) wine (that is to say [wine] of midle age) is hote in the seconde
degree (Turner, A New Boke of the Natures and Properties of
All Wines B3R)
132 Päivi Pahta and Saara Nevanlinna

In example (7) there are two appositions which both have ellipsis in the
second appositive. The second apposition is embedded in the second unit
of the first apposition.

(7) 'Honours thy fadyre and pi modyre.' That es, [honoure thy fa-
dyre and pi modyre] in twa thynges, fiat es, fin] bodyly and
gastely [thynges], (Rolle, Prose Treatises 11)
'Honour your father and your mother, that is, in two things, that
is, in bodily and spiritual.'

Units of apposition can appear in a variety of syntactic functions within

phrases, clauses, or sentences. In Present-day English, appositions are
most common in functions connected with the noun phrase (Meyer 1992:
34-35, Quirk et al. 1985: §17.65), and this seems to be the tendency in
our historical data as well. Appositional constructions are very common
in the function of subject (8) and object (9).

(8) Forsothe euer eithir was nakid, that is, Adam and his wif.
(Wycliffe, The Old Testament Genesis 2 25)

(9) Tak pentafoyloyn, id est quintfoyle, & welle it wele in water

(The 'Liber de Diversis Medicinis' in the Thornton Manuscript
13)
'Take pentafilon, that is, cinquefoil and boil it well in water'

In our data appositions were quite frequent in the function of the verb,
as in examples (10) and (11), whereas in Present-day English these are
very rare (Meyer 1992: 36).

(10) For the reformynge of the great Mischiefes and Inconveniences

that daylie growe and increase by reason of the pestering of
Houses with div~se Famylies... (The Statutes of the Realm IV
852)

(11) l>e splene ... is heled or leched aboute J)e lefte side, as Galien
saij). (The Cyrurgie of Guy de Chauliac 64)
'The spleen is healed or cured along the left side, as Galen
says.'
 Expository apposition 133

Appositions were also found as subject complement (12), object com-

plement (13), and adverbial (14).

(12) t>e sexte dedly synn es 'slewthe or slawenes'

(Gaytryge, Sermon 13)
'The sixth deadly sin is sloth or slowness'

(13) My deere harte, I cannot forgett the obligation that I am bound

in to the, that is to write to thee (Thomas Knyvett, The Knyvett
Letters 59)

(14) Whereupon sone after that is to wit, on the friday the thirtene
day of Iune many Lordes assembled in the tower (More, The
History of King Richard 111 46)

2.2. The semantic classes of apposition

Expository appositional constructions can be divided into various seman-

tic classes. This section shows the semantic classification used in our
study, for which we are obliged to Meyer (1987: 103 and 1992: 73-82),
and to Quirk et al. (1985: §§17.74-17.81). Quirk et al. (1985) distinguish
between four relations of equivalence in apposition (appellation, identifi-
cation, designation and reformulation). Their classification includes only
noun phrases in apposition, but it seems to be applicable to other ap-
positional constructions as well. The same semantic classes are also found
in Meyer (1987: 113). In his more extensive corpus-based study Meyer
(1992) made some adjustments to his earlier classification, redefined
some classes and added some classes to account for semantic relation-
ships not included in Quirk et al. (1985). He divides appositions into three
basic types according to the various ways in which the second unit of ap-
position conveys new information about the first unit. The new infor-
mation provided by the second unit may be more specific than the infor-
mation provided by the first unit (identification and appellation); it may
be less specific than the information given in the first unit (characteriza-
tion); or the two units may be equally specific (paraphrase, self-correc-
tion, and reorientation). We adopt Meyer's scheme for the semantic clas-
sification of appositions according to the specificity of the units. The in-
dividual classes that we have distinguished within the last of the major
semantic categories differ to some extent from those that Meyer derives
134 Päivi Pahta and Saara Nevanlinna

from his modern corpus material (see pp. 140-142 below). Assigning ap-
positional constructions to these classes is not unproblematic and in some
cases the interpretation is left to the discretion of the scholar.
Table 4 shows the semantic classes into which we have classified ap-
positions and the frequency of these classes in our data. In the following
sections we discuss the semantic classes in more detail and provide exam-
ples of different types of constructions belonging to these classes.

Table 4. The semantic classes of apposition.

SEMANTIC CLASS Ν %

Identification 228 5.9

Appellation 20 0.5
Characterization 8 0.2
Paraphrase 3478 89.9
Revision 135 3.5

Total 3869 100

2.2.1. Appositions in which the second appositive is more specific than

the first appositive

Identification (Meyer 1992: 74; Quirk et al. 1985: §17.77; identifying or

equative apposition in Huddleston 1971). The second appositive, which is
more specific than the first one, identifies what is referred to in the first
appositive. Two subtypes can be distinguished.
In the first subtype the appositives are coreferential, i.e., they refer to
the same "piece of reality" (Leech 1981: 156), as in examples (15) and
(16).

(15) who J)at is absent paie pe peyne, jDat is to witene, a pound of

wex (Returns 55)
'he who is absent shall pay the fine, that is, a pound of wax'

(16) It remaineth now, that wee consider the thing prescribed, name-
ly wherein we must bee built. (Hooker, Two Sermons upon Part
of S. Judes Epistle 41)
 Expository apposition 135

The first appositive can also be a pro-form (a plural personal pronoun, an

indefinite pronoun, a numeral, or the like) coreferring to the second appo-
sitive (17).

(17) It is to wete |)at in flebotomie 4 pyngis are principalli attendid:

sc., custome, tyme, age, & vertue (A Latin Technical Phlebo-
tomy 39)
'It is to be known that in phlebotomy four things are mainly to
be attended to, that is to say, custom, time, age and virtue'

The second subtype of idenfication contains appositions in which one

of the units has no referential value, i.e., it is a clause or a sentence
(Meyer 1992: 74-75). In this type of apposition the units are related by
cataphoric reference; the first unit, usually a noun phrase, refers cata-
phorically to the second. Examples (18) and (19) illustrate this kind of
identification.

(18) But saynte Austyne saythe a comfortable worde again, to them

that gyue theyr tythes truely, that is to saye: Decimae sunt
tributa egentium animarum: Tithes are tributes or rewardes to
nedye soules. (Fitzherbert, The Book of Husbandry 37)

(19) And if J>at hors schal be y-reden while his back is sore £)en jx>u
most use pe cure aforsaide j}at is to say. at euene take of pe
sadel. & wasche it with hot wyn or vryne. (A Late Middle Eng-
lish Treatise on Horses 127)
'And if the horse is ridden while his back is sore then you must
use the aforesaid cure, that is to say, in the evening take off the
saddle and wash it with hot wine or urine.'

The commonest present-day markers in the relation of identification are

namely, that is, that is to say (Meyer 1992: 74). In our data the markers
used are (and) that is, that is to say, (that is) to wit, id est, scilicet, as,
(as) namely, and viz., some of which have several spelling variants.

Appellation (Meyer 1992: 76; Quirk et al. 1985: §17.76). Appellation re-
sembles identification, but in addition to identifying the referent of the
first unit, the second unit also names it. Examples (20) and (21) illustrate
this kind of usage.
136 Päivi Pahta and Saara Nevanlinna

(20) Aftir these thingis Jhesus wente ouere the see of Galilee, that is
Tiberias. (Wycliffe and Purvey, The New Testament John 6 1)
'After these things Jesus went over the sea of Galilee, that is,
Tiberias.'

(21) And withyn shorte tyme after, I, standynge yn Cheapesyed,

sawe these iiij ryed ['ride'] throwe Chepe, (that is to saye,)
kynge Phyllype, quene Marye, cardynall Poole, and Steven Gar-
dynar (Mowntayne, Narratives of the Days of the Reformation
209)

The commonest apposition markers in Present-day English are that is,

that is to say, in other words (Meyer 1992: 76). In our samples only that
is and that is to say are found in Middle English and that is to say/meane,
(and) namely, and viz. in Early Modern English.

2.2.2. Appositions in which the second appositive is less specific than

the first appositive

Characterization (Meyer 1992: 78; designation in Quirk et al. 1985:

§17.79). This is the converse of identification and appellation; the second
appositive provides some general characteristic of the first. The apposi-
tives are coreferential noun phrases (or attributive, cf. Jackendoff 1985:
73). Characterization is illustrated in example (22).

(22) he Jjat lokuj) in Godus lawe, f>at is lawe of parfijt fredom, and
dwelluj) parfijtly in J?is lawe by al his lif (Wycliffite Sermons 27
1 590)
'He that observes God's law, that is, the law of perfect freedom,
and abides completely to this law all his life'

When that is is used as a marker, it is often difficult to distinguish the de-

monstrative that from the relative pronoun that, as in example (23).

(23) ye shullen do no thyng which may in any manere displese God,

that is youre creatour and makere. (Chaucer, The Tale of Meli-
bee 234.CI)
'You shall do nothing that can in any manner displease God,
that is, your creator and maker.'
 Expository apposition 137

In our corpus characterization is not a frequent class of apposition."

The markers that is and that is to say occurring in our data may also be
used in Present-day English (Meyer 1992: 78).

2.2.3. Appositions in which the second appositive is as specific as the

first appositive12

Paraphrase (Meyer 1992: 79; reformulation in Quirk et al. 1985: §17.80).

The units of this type of apposition are related by synonymy and can
represent lexical synonymy or be synonymous phrases, clauses or sen-
tences (cf. Burnley 1992: 462-463). Lexical appositive units display de-
grees of synonymy varying from absolute synonymy at one end of the
scale to speaker synonymy at the other (Cruse 1986: 265-268); this can
also be used for describing the synonymous relation in phrasal apposition.
In absolute synonymy the units have identical meanings in any con-
texts, their contextual relations are identical, as in example (24).

(24) J)ei sehe wer loth & not wylly to do swech thyngys
(Kempe, The Book of Margery Kempe 1 55)
'though she were reluctant and not willing to do such things'

The second unit is often added in order to provide a more familiar vari-
ant, as in the following examples, where the second unit provides a native
translation variant of a Latin term (25) or name (26).

(25) Jjre substaunces ben made of f)e chyle by decoccioun (i. se-
pinge) in the lyuer (The Cyrurgie of Guy de Chauliac 62)
'three substances are made of the fluid of the intestines by a
process of decoction, i.e., boiling in the liver'

(26) archebisshop of Dorobernya, £at is Canturbury (Trevisa, Poly-

chronicon Ranulphi Higden VI 217)
'Archbishop of Dorobernia, that is, Canterbury'

Sometimes the more technical or foreign variant is contained in the sec-

ond appositive.
138 Päivi Pahta and Saara Nevanlinna

(27) Provided alwey that this acte extend not to eny Shipp or Shipp-
es, havyng eny of the seid Wares or marchaundisez, constreyned
by tempast of Weder or enemyes to arrive in eny porte or place
within this youre Reame. {The Statutes of the Realm II 535)

One of the appositives may also be a regional variant.

(28) Frende, how comeste ]30u in here — oJ)ur hidur — and haste no
leueree of my weddynggus? {Middle English Sermons (MS Roy-
al) 17)
'Friend, how do you come here, or hither, and are not wearing
a gown for my wedding?'

In speaker synonymy the second unit does not serve as an exact para-
phrase of the first unit but rather as an explanation or clarification of the
meaning the speaker intended the first unit to have. Semantically, the two
units share sufficient salient characteristics to be regarded as roughly syn-
onymous by the speaker, though not absolutely so. Examples (29) and
(30) illustrate speaker synonymy in paraphrase.

(29) for God may more do J)an any man may penke or vnderstonde.
{The Book of Vices and Virtues 109)
'for God can do more than anyone can think or understand.'

(30) And this loue or appetyte that euerye thynge hathe to it selfe,
procedeth ... by naturall intencion. (Colville, Boethius 80)

As some of the preceding examples show, the class of paraphrase also

includes instances of coordinative apposition. The units are either syno-
nyms or near-synonyms and they are joined by and or or. In this capacity
the 'inclusive' or and and are variants of the coordinative apposition
marker. Wrenn points out (1973: 265) that the Old English oppe meant
both 'or' and 'and'. The addition of the second appositive enlarges the
conceptual field of the first appositive and thus strengthens the effect of
the single phrase, as in (31).13
 Expository apposition 139

(31) thilke cerkle that is innerest or most withinne ... it is like a

mervayle or miracle. (Chaucer, Boethius 452.C1-2)
'That circle that is innermost or most central ... it is like a mar-
vel or a miracle.'

Coordinative apposition as paraphrase was very common in Old and

Middle English as a stylistic device, along with other coordinative word
pairs (cf. Mitchell 1985:§ 1809). The two units often alliterated (Oakden
1930; Koskenniemi 1968). In Present-day English, a great many coordi-
native word pairs, especially nouns, have become idiomatic in colloquial
language, with a fixed collocation (see e.g. McMordie 1976: 84).
There are several examples where the same two appositives are joined
both by and and by or in the same text (cf. Norri 1992: 177):

(32) Lamentation aryseth of foure affections, eyther of a great feare,

or dreade, or of a great shame, or of some sorrowe, or els of
some hatred ... the people were in a greate feare and dreade to
be oppressed (Fisher, The English Works of John Fisher 1 397)

(33) all thynges that be vnder or subiect to destinie be also subiect

and vnder gods prouidence and ordynaunce. To the whyche
prouidence or ordynaunce destinye is subiecte it selfe (Elizabeth
I, Queen Elizabeth's Englishings of Boethius, Plutarch, &c 107)

In clausal or sentential apposition the principles of truth-functional se-

mantics can be applied to describe the synonymy between the appositives
since both units are synonymous if they express the same truth value
(Allwood et al. 1977: 61; Leech 1981: 74). This type of usage, exempli-
fied in (34) and (35), is frequently resorted to when there is a need to ex-
pound a difficult passage, especially in religious and philosophical writ-
ings. The apposition marker acts here as the linguistic correlate of the
identity-operator in mathematics and logic (cf. Lyons 1979, 2: 472). The
first appositive (i.e. the first sentence) is true for the interpretation that the
speaker has given to it in the second clause or sentence, which acts as the
second appositive. In apposition expressing truth-functional synonymy
there are several markers that had previously undergone grammaticaliza-
tion, e.g., yea and as who say (see Pahta—Nevanlinna forthcoming).
140 Päivi Pahta and Saara Nevanlinna

(34) The riches of the sinner is laid vp for the righteous·, that is, the
righteous shall inioy that which the wicked gathereth. (Smith,
Sermons D8V)

(35) "drede Jje not, maide Mari for pou hast fownden grace aneyns
almyjthy God." As hoo seif). "Pou arte gracious in pi-selfe and
also shall be to all Cristen pepull." (Middle English Sermons
(MS Royal) 259)
'Don't be afraid, virgin Mary, for you have found grace with al-
mighty God, in other words, you are virtuous in yourself and
will be so to all Christians.'

Clauses or sentences in apposition also contain instances where the

second appositive provides a translation of the clause or sentence appear-
ing as the first appositive, as in (36).

(36) Sanat, sanctificat, et ditat surgere mane. That is to say, Erly

rysyng maketh a man hole in body, holer in soule, and rycher in
goodes. (Fitzherbert, The Book of Husbandry 101)

Coordinative apposition can also express truth-functional synonymy; for

example, in clausal paraphrase, as in (37).

(37) to dwellyn wyth me & neuyr to departyn fro me (Kempe, The

Book of Margery Kempe 1 53)
'to stay with me and never to leave me'

According to Meyer (1992: 79), the markers used in present-day ap-

positions in the class of paraphrase are or, and that is, that is to say, and
in other words. In the diachronic data several markers are found. That is,
and, or, or else/rather, that/this is (for) to say, as who say/saith, yea
(and), id est, i.e., /., viz. and scilicet are all frequent, but there are also a
number of rare variants including as, (the which is) as moche to saye as,
pat is to vnderstande, pat is to menynge, and this to sayn as who saith.14

Revision (Quirk et al. 1985: §17.80;15 reorientation and self-correction

in Meyer 1992: 80-81). In revision the units of apposition are corefer-
ential, but not synonymous as in paraphrase. Two subtypes can be dis-
tinguished according to the nature of coreference between the units. In the
 Expository apposition 141

first type, which corresponds to what Meyer (1992) calls reorientation, the
two appositives are strictly coreferential and there is a close connection
between the meaning of the units and their referents in the external world.
The second appositive refocuses the reference of the first, that is to say,
it provides a different way of viewing it. Examples (38) and (39) illustrate
this kind of revision.

(38) Who is elles kepere of good or dryvere awey of yvel but God...?
(Chaucer, Boethius 453.C1)
'who else is a keeper of good or a dispeller of evil but God?'

(39) The tend commandement an pe laste es, Jjat we 3erne noghte J>e
wyefe of oure neghteboure. (Gaytryge, Sermon 6)
'The tenth and the last commandment is that we should not
desire the wife of our neighbour.'

The writer can use reorientation for reflecting what semantic features of
a notion are uppermost in his mind. A good example is Chaucer's illustra-
tion of the various sides of the profession of an advocate:

(40) thise oratours or advocattes\ the accusours or advocattes; the

deffendours or advocatz (Chaucer, Boethius 449.C1)
'these orators or advocates; accusers or advocates; defenders or
advocates'

In the second type of revision, which is equivalent to what Meyer

(1992) calls self-correction, the units are related by speaker coreference,
i.e., the speaker intends the units to have the same extralinguistic referent,
although they have distinct meanings. The second appositive corrects a
mistake in the first appositive or provides more accurate information and
is usually more emphatic than the first. In Present-day English this kind
of reformulation is generally analyzed as a performance error in im-
promptu speech, and can be appropriately called self-correction. In our
written historical data this type of revision occurs most often in construc-
tions where the second appositive provides an expression which is more
appropriate or precise than the first unit. Examples of this kind of usage
are provided in (41) and (42).
142 Päivi Pahta and Saara Nevanlinna

(41) Some men here present, ye & not a fewe. woll perauenture muse
why & to what entent I brynge in thys (Fitzjames, Sermo Die
Lüne ΒIV)
'Some here present, and not just a few, will perhaps wonder
why and for what purpose I bring this in'

(42) in the forepart of his head, between the two fore-leggs, [a flea]
has two small long jointed feelers, or rather smellers, MM,
which have four joints, and are hairy, like those of several other
creatures (Hooke, Micrographia 13.5 210)

In some Early Modern English examples the second unit reveals the
ironic attitude of the speaker or writer, as in examples (43) and (44).

(43) For, the scholer, is commonlie beat for the making, whe~ the
master were more worthie to be beat for the mending, or rather,
marring of the same (Ascham, The Scholemaster 182)

(44) When he saw the king's temper, and his way of managing, or
rather of spoiling, business, he grew very uneasy (Burnet, Bur-
net 's History of My Own Time 1 I 170)

In Present-day English, the markers of apposition in revision are or,

sometimes followed optionally by the conjuncts rather or better, and that
is (to say) (Meyer 1992: 81). In our diachronic data and, or, or else/
rather, that is (to say), and yea (and) are found in this relation.

3. Generic distribution of appositional constructions

Although the use of apposition is in most cases optional, in some contexts

appositions are more necessary than in others. The necessity of the sec-
ond appositive can depend on the linguistic context or, more often, on the
situational context, which means that the speaker or writer has to de-
termine whether the audience needs to be supplied with new information
about the first appositive. He may choose to add the second appositive to
ensure that his audience understands exactly what he means by a certain
phrase or clause, or he may just wish to provide some further information
about the subject of discussion (cf. Meyer 1987: 114-115). This makes
 Expository apposition 143

apposition an additive relation, the second appositive adding information

to the discourse (Halliday—Hasan 1976: 248-250). Appositions are most
needed and, according to Meyer's study of apposition in Present-day Eng-
lish, most frequent in genres characterized by a low degree of shared
knowledge (Meyer 1991: 179).16 Certain types of appositions are also
found more often in some text types than in others, their communicative
functions being better suited to some text types than to others (cf. Meyer
1992: 92). Furthermore, there may be variation in the frequency of ap-
position in texts within individual text types, as text types can be very
heterogeneous in their use of linguistic features and there may be varia-
tion even within narrowly defined groups of texts (cf. Biber 1988, Meur-
man-Solin 1993a and Taavitsainen 1993).

3.1. Expository appositions across text types

Although expository appositions with a specific marker were attested in

texts belonging to all text types represented in the HC, they were not
evenly distributed across the data. Tables 5 - 9 (see pp. 144-147) show the
frequencies of appositions in the five subperiods (ME3-EModE3) of the
HC included in our study.
Some general remarks can be made about the results presented in
Tables 5-9. Although there is temporal variation within all text types,
there seem to be some types of writing where expository appositions con-
taining an explicit marker are consistently rare throughout the period of
our study. In all texts classified as narrative the incidence of appositions
is low. This applies to both imaginative and non-imaginative types of
narration, i.e., fiction, romances, diaries, biographies, travelogues and his-
torical writings. In all texts belonging to these text types the incidence of
appositions is below 4 per 1000 words. Appositions with an explicit
marker are also infrequent in most texts which could be characterized as
speech-based, such as drama texts and proceedings of state trials. They
are also relatively rare in all correspondence, but in official letters they
are more frequent than in private letters in all the subperiods containing
both types of correspondence. It is also worth noting that the English
Bible texts included in our data contain very few appositional construc-
tions with an explicit marker.
144 Päivi Pahta and Saara Nevanlinna

Table 5. Frequency of appositions in text types in ME3 (1350-1420). Ν =

absolute numbers (for the abbreviations used, see pp. ix-x in this
volume).

Text category Text type Word count Ν Ν/1000

IS Science, medicine 3 580 61 17.04

STA Document 1 520 13 8.55
XX Philosophy 12 630 105 8.31
IR Religious treatise 37 080 133 3.59
XX Official letter 5 010 15 2.99
XX Document 13 870 41 2.96
IS Handbook, astronomy 13 040 38 2.91
IR Homily 7 280 12 1.65
NI Fiction 14 320 21 1.47
NN History 24 500 32 1.31
IR Rule 5 390 7 1.30
IS Handbook, medicine 6 110 7 1.15
NN Travelogue 5 530 6 1.08
XX Bible 20 910 17 0.81
IR Sermon 13 460 7 0.52

Total 184 230 515

Table 6. Frequency of appositions in text types in ME4 (1420-1500).

Text category Text type Word count Ν Ν/1000

IS Science, medicine 6 430 64 9.95

STA Law 11 240 110 9.79
IR Religious treatise 41 090 348 8.47
IR Sermon 25 010 170 6.80
IR Rule 1 810 12 6.63
XX Preface 6 200 35 5.65
XX Official letter 3 140 17 5.41
IS Handbook, other 11 560 53 4.58
XX Document 10 480 33 3.15
IS Handbook, medicine 5 670 16 2.82
NI Fiction 8 760 18 2.05
 Expository apposition 145

Table 6. Continued.

Text category Text type Word count Ν Ν/1000

XX Deposition 1 970 4 2.03

XX Bible 2 530 5 1.98
NN History 12 600 19 1.51
NN Biography, saints lives 3 820 5 1.31
XX Mystery play 20 100 19 0.95
XX Private letter 19 490 17 0.87
NI Romance 19 070 12 0.63
IS Handbook, astronomy 2 880 - -

Total 213 850 957

Table 7. Frequency of appositions in text types in EModEl (1500-1570).

Text category Text type Word count Ν Ν/1000

XX Philosophy 9 890 262 26.49

STA Law 11 790 303 25.70
EX Science, medicine 6 180 107 17.31
IS Handbook, other 10 000 116 11.60
IR Sermon 9 470 75 7.92
IS/EX Educational treatise 10 440 80 7.66
EX Science, other 6 700 41 6.12
XX Trial 15 970 59 3.69
NN History 11 090 22 1.98
NN Biography 5 440 9 1.65
NI Fiction 11 550 19 1.65
XX Official letter 6 300 10 1.59
NN Travelogue 14 100 22 1.56
XX Comedy 10 570 11 1.04
NN Autobiography 5 740 5 0.87
XX Private letter 10 640 9 0.85
NN Diary 13 060 8 0.61
XX Bible 21 230 13 0.61

Total 190 160 1 171

146 Päivi Pahta and Saara Nevanlinna

Table 8. Frequency of appositions in text types in EModE2 (1570-1640).

Text category Text type Word count Ν Ν/1000

STA Law 11 780 221 18.76

EX Science, medicine 6 620 75 11.33
EX Science, other 6 420 42 6.54
EX Educational treatise 11 220 65 5.79
NN Travelogue 14 780 58 3.92
IR Sermon 10 300 38 3.69
IS Handbook, other 12 290 44 3.58
NN Autobiography 4 120 13 3.16
NN History 10 090 26 2.58
NN Biography 4 810 7 1.46
XX Official letter 5 660 7 1.24
XX Trial 14 230 17 1.19
XX Comedy 11 810 13 1.10
XX Philosophy 6 880 6 0.87
XX Bible 22 190 15 0.68
NI Fiction 12 490 7 0.56
XX Private letter 11 590 5 0.43
NN Diary 12 520 2 0.16

Total 189 800 661

Table 9. Frequency of appositions in text types in EModE3 (1640-1710).

Text category Text type Word count Ν Ν/1000

STA Law 13 180 218 16.54

XX Philosophy 8 820 47 5.33
EX Science, other 11 280 60 5.32
IR Sermon 12 470 52 4.17
IS Handbook, other 11 370 37 3.25
NN Travelogue 10 470 31 2.96
IS Educational treatise 11 320 26 2.30
NI Fiction 12 040 25 2.08
XX Official letter 5 870 11 1.87
NN History 11 640 19 1.63
 Expository apposition 147

Table 9. Continued.

Text category Text type Word count Ν Ν/1000

NN Biography 6 170 9 1.46

XX Trial 13 760 9 0.65
NN Diary 11 210 7 0.62
XX Private letter 13 140 7 0.53
XX Comedy 12 740 5 0.39
NN Autobiography 5 560 2 0.36

Total 171 040 565

Throughout the period of our study the use of expository apposition with
a specific marker is frequent in all statutory writing included in the data.
It is also frequent in most writings that can be characterized as expository
or instructional, although there is more variation between individual texts
within these categories. In the field of secular instruction, appositions
containing a marker are frequent in scientific writings, particularly in the
medical treatises, whereas in handbooks they are generally less frequent.
They are also fairly common in educational treatises, and in some texts
belonging to religious instruction. Finally, they are frequently used in
most philosophical treatises included in our data. In the following sections
we attempt to look more deeply into those text types and individual texts
in which expository appositions with an explicit marker are relatively
most frequent.
The use of appositional markers in the data will be briefly discussed in
section 4 below, but it is useful at this point to distinguish between two
main types: appositions containing the native speech-based coordinators
and and or, and those containing other markers. And and or occur as
apposition markers in the majority of appositions registered (see p. 168).
In most instances they join two synonyms or near-synonyms in the class
of paraphrase.

3.1.1. Statutory writings

Samples of statutory writing are contained in all five subperiods included

in this study. English was reintroduced as the language of law towards the
end of the fourteenth century, so that only some statutory documents are
148 Päivi Pahta and Saara Nevanlinna

available from subperiod ME3, but the other four subperiods contain well
over 10,000 words of legislative writing.
The language of law contains a number of distinctive features which
have been commented on by several writers, including the use of archaic,
foreign and uncommon words; repetition of lexical items instead of refer-
ence devices; long, thinly-punctuated complex sentences with complicated
patterns of coordination and subordination; and the use of passive
voice.17 All these features contribute to the distinctive style of legal writ-
ing, which has been described as "a complex intricate, even bizarre style
of language" completely lacking "colour and humanity" (Maley 1987:
25). Legal language is also frequently described as formulaic, and each
species of legal transaction has its own collection of linguistic formulae,
which have in the course of time been thoroughly tested before the courts
and found to be adequate. The use of the established formulae contributes
to the extreme linguistic conservatism of present-day legal language
(Crystal—Davy 1974: 194; Hiltunen 1990).
According to Maley (1987: 27), the key to understanding legislative
language is the fact that it aims at controlling actions by words. The
binding force of the statute is attached to its verbal formulation, which
has to be extremely explicit and precise, because in a court of law "atten-
tion will be paid only to what [a document] appears to declare: any
intentions of the composer which fail to emerge clearly are not usually
considered in arriving at what the document means" (Crystal—Davy
1974: 193). Consequently, all legal documents must be written in a man-
ner which leaves no place for misinterpretation.18
The pursuit of certainty, explicitness and precision seems to be the
explanation for the frequency of appositions in statutory writing. In our
corpus, appositions in this category in each subperiod belong almost ex-
clusively to the semantic class of paraphrase (c. 98 per cent of the total),
vast majority of them consisting of synonymous or near-synonymous
words linked by and or or, as in example (45)

(45) Provided that yf ther be eny lyke Forfeyture co~mytted or done

by eny of the Qwenys sarvand~ beyng in her Cheker roule (The
Statutes of the Realm III 8)

It has been suggested that the source of the habit of using these word-
pairs lies in the multilingual origin of English law, particularly in the
period when English was replacing French as the language of law. It is
true that both Latin and French have influenced English legal language
 Expository apposition 149

heavily, and a large part of the synonymous or near-synonymous word-

pairs occurring in our data are of the type where a word of foreign origin
occurs in apposition to a native word for the same or a similar concept,
as in example (45) above. Further examples are provided by such pairs as
act and deed, goods and chattels, lie and extend, lose and forfeit, etc. In
the multilingual situation of medieval England there could be uncertainty
as to whether such synonyms meant exactly the same thing and the inclu-
sion of both alternatives contributed to the precision of the text and re-
duced opportunities for misinterpretation (Crystal—Davy 1974: 208). It
is not, however, in the fourteenth century that the custom of using syno-
nymous word-pairs appears for the first time in legal language. The same
habit is already found in Old English legal documents (cf. Hiltunen
1990). The native tradition is still continued in the Middle English and
Early Modern English legal texts (and even in Present-day legal English),
and our data includes many appositions consisting of native words, such
as by and with, have and hold, keep and hold, have and keep, or all and
every, as in (46).

(46) honeste men of the same Parishe being Housholders, shall have
full power and aucthoritie by vertue of this Acte to enter into
all and everie House (The Statutes of the Realm IV 853)

The habit of using synonymous words in apposition seems to have reach-

ed its peak during the early sixteenth century (25.53 paraphrases per 1000
words in EModEl). After that there is a slight decline in their frequency
(17.83 per 1000 words in subperiod EModE2 and 16.09 per 1000 words
in EModE3), but they remain proportionally more frequent in statutory
writings than in any other type of writing included in our data.

3.1.2. Scientific treatises

The scientific texts in the HC are divided into medical treatises and other
scientific treatises which belong to the field of physical sciences. Medical
treatises are included in the Late Middle English subperiods and the first
two Early Modern English subperiods, whereas other scientific treatises
are only included in the three Early Modern English subperiods. With the
exception of EModE3, both text types are represented by one text in each
subperiod; EModE3 contains two texts labelled under other scientific trea-
tises. Although the use of expository apposition with an explicit marker
150 Päivi Pahta and Saara Nevanlinna

is relatively frequent in all scientific treatises included in our data, it is

particularly so in medical writing.
In technical texts the use of appositional reformulation contributes to
accurate understanding of particular concepts and accordingly helps the
reader to understand the surrounding text. Authors who either argue in
favour of a particular theoretical position or wish to instruct their readers
about a phenomenon or theory are especially interested in presenting their
utterances so that their audiences understand them (Blakemore 1993:
101). In medical treatises, which aim at maintaining or restoring health,
the need to communicate successfully is vital. Although the medical trea-
tises included in our corpus were aimed at medical professionals, i.e., an
audience relatively familiar with the basic concepts of contemporary
medicine, and the degree of shared knowledge between the author and the
addressee is supposedly high, the complexity of the material (both termi-
nology and procedures), seems to have required attention by way of appo-
sitional explanations, as in (47).

(47) a veyne goof) out, icleped vena concaua (i. the holowe veyne)
{The Cyrurgie of Guy de Chauliac 62)
'a vein called vena concava, i.e., the hollow vein, goes out'

Appositions of this type seem to be frequent in other similar texts of the

period. In a study of the names of sicknesses in Middle English (Norri
1992), appositional constructions serving as an explicatory technique in
presenting terminology were frequently found in both academic medical
treatises and surgical manuals (both classified as scientific treatises in the
HC) in contrast to handbooks or remedybooks, which contained far fewer
instances of this technique. The distinction between scientific treatises and
handbooks is also visible in the present material: texts classified as hand-
books generally contain fewer expository appositions with an explicit
marker than those classified as scientific.
The Middle English medical treatises included in the corpus are trans-
lations from Latin. In the vernacularization of scientific Latin texts at
least two factors can be seen as a cause of frequent use of apposition.
Firstly, the model provided by the corresponding constructions in the
original Latin text (cf. Blake 1992b), and secondly, the need to provide a
vernacular translation for technical terms. Both trends, exemplified in (48)
and (49) respectively, are visible in the Middle English translation of a
Latin technical Phlebotomy included in subperiod ME3, in which the
 Expository apposition 151

incidence of apposition is 17.04 per 1000 words, 98 per cent being para-
phrases of technical terms mainly joined by the markers i.e., or, and sc.

(48) Laboryng ... of agriacape gasiliontes, i.e., grene jawnes, my-

nusche of basilica of Jie ryjt arme. (A Latin Technical Phlebo-
tomy 51)
'In treating green jaundice, let blood from the basilica of the
right arm.'
cf. Latin: Laborans agracabe gaselcontis, id est yctericia viridi
(Voigts—McVaugh 1984: 50)

(49) it is to wyt that some minision is made or done by meta-

centesyn, i.e., even drawyng (A Latin Technical Phlebotomy 37)
'It is to be known that sometimes blood is let by metacentesis,
i.e., bleeding from the side on which the sickness is found'
cf. Latin: Notandum igitur quod minutionum alia per antispasim
fit, alia per metacentesim. (Voigts—McVaugh 1984: 36)

Not all paraphrases, however, provide a native variant for a foreign term;
in some instances the second appositive contains a synonymous technical
term of foreign origin. There are cases where an apposition in the Latin
text has been translated, but in many of these cases the English translator
has provided another appositional explanation, which is sometimes attach-
ed to the preceding unit with an explicit marker and sometimes without
it, as in example (50).

(50) Fyrst is cephalica, i.e., capitalis or hedd veyne; 2, mediana or

cardiaca, i.e. veyne of the herte; 3, epatica or bacilica, lyuer
veyne (A Latin Technical Phlebotomy 39)
'The first is the cephalic vein, i.e., the capital or head-vein; the
second is the median or cardiac vein, i.e., the vein of the heart;
the third is the hepatic or basilic vein, that is, the liver vein'
cf. Latin: Prima est cephalica, id est capitalis·, secunda est medi-
ana, id est cardiaca vel vena cordis; tertia est epatica vel basi-
lica (Voigts—McVaugh 1984: 38)

In the translation of Guy de Chauliac's Chirurgia Magna included in

ME4 the overall frequency of appositions is 9.95 per 1000 words, which
is clearly lower than in the Phlebotomy. Being a surgical manual, its aim
is to provide more straightforward, practical advice on surgical procedures
152 Päivi Pahta and Saara Nevanlinna

and perhaps pay less attention to theoretical terminology, although there

are some expository passages. The lower frequency of appositions may
also be connected with the general nature of the particular translation in-
cluded in the corpus. Of the three independent Middle English transla-
tions of this work, that included in the HC has been characterized by its
editor as an attempt to render the text in a more thoroughly English idiom
(Ogden 1971: vi). Providing variant terminology is, however, the main
source of expository apposition in this text, too; 88 per cent of the apposi-
tions were paraphrases of technical terms mainly joined by the markers or
or /., the early variant of i.e. (51).

(51) And j)an f)ou schalt byholde J)e nombre and quantite of f>e spon-
diles (i. pe lynkes ioynt or bones of pe bak).
(The Cyrurgie of Guy de Chauliac 65)
'And then you will see the number and quantity of the verte-
brae, i.e., the links or bones of the back.'

The medical treatises included in the first two Early Modern English
subperiods follow the same model as far as appositions are concerned.
Thomas Vicary's The Anatomie in EModEl dates from the mid-sixteenth
century, although it is mainly based on the work of a thirteenth-century
surgeon (Bennett 1952: 108). The incidence of apposition is 17.31 per
1000 words, and the proportion of paraphrases is 96 per cent. The major-
ity of the paraphrases consist of synonymous or near-synonymous word-
pairs joined by and or or, as in examples (52) and (53).

(52) Why it is called Piamater, is, for because it is so softe and ten-
der ouer the brayne (Vicary, The Anatomie of the Bodie of Man
29)

(53) The fyrst is, that they shoulde keepe and defende the Eye from
duste (Vicary, The Anatomie of the Bodie of Man 35)

Subperiod EModE2 contains William Clowes's surgical text, Treatise

for the Artificiall Cure of Struma from 1602. The incidence of appositions
is 11.33 per 1000 words, paraphrases comprising 87 per cent of all in-
stances. The use of and as a paraphrase marker is higher in Clowes's text
than in any other scientific treatise in our data: 59 per cent of all the ap-
positional constructions are synonymous or near-synonymous words or
phrases joined by and. Although providing a translation for a foreign
 Expository apposition 153

word is also an important function of appositional structures in this text,

stylistic reasons seem to be important, as in (54).

(54) diuers learned mens writings, which heer to repeate were

needles, and to no great vse (Clowes, Treatise for the Artificiall
Cure of Struma 22)

Other pairs occurring frequently in apposition are curing and healing,

diuers and sundry, and end and purpose.
The treatises representing physical sciences are somewhat different from
the medical texts. The incidence of expository apposition containing a
marker is lower and the proportion of paraphrase is generally smaller,
particularly in the sixteenth-century samples. In Robert Record's The
Path-way to Knowledg, Containing the First Principles of Geometrie of
1551 (EModEl), claimed to be the first book treating questions on geo-
metry in English (Bennett 1952: 115), incidence of apposition is 6.12 per
1000 words and the proportion of paraphrase is 56 per cent. The remain-
ing appositions belong to the semantic class of identification, reflecting
the fact that features such as listing or enumerating items are frequent in
this treatise (55).

(55) Nowe adde togyther all the partes of the two lesser squares, that
is to say, sixtene and nyne, and you perceyue that they make
twenty and fiue (Record, The First Principles of Geometrie
E4R)

A similar tendency is seen in Blundevile's Α Briefe Description of the

Tables of the Three Speciall Right Lines Belonging to a Circle written on
a related topic in 1597 (EModE2).

(56) Of which Zones the auncient men were wont to say that three
were unhabitable, that is, the two colde, and the extreame hoat
(Blundevile, The Tables of the Three Speciall Right Lines Be-
longing to a Circle 157R)

The incidence of apposition in Blundevile's text is 6.54 per 1000 words,

paraphrases comprising 67 per cent of all instances.
Robert Hooke's Micrographia of 1665 and Robert Boyle's Electricity
and Magnetism of 1675-1676 (EModE3) differ greatly from each other,
which shows that there can be a lot of variation within individual text
154 Päivi Pahta and Saara Nevanlinna

types. Hooke's text contains 8.25 appositions per 1000 words, whereas in
Boyle's treatise the corresponding figure is only 1.72. The proportion of
paraphrase in both treatises is higher than in the earlier treatises, c. 80 per
cent in each. In the majority of the paraphrases one of the units provides
a more familiar variant for an originally French or Latin word (Cells or
Boxes, a dore or passage, true and genuine), but there are also instances
where the apposition serves merely a stylistic function (shining and
bright).

3.1.3. Religious instruction19

One of the prototypical text categories rich in apposition is religious in-

struction. Religious language is not homogeneous, which is also reflected
in the varying use of apposition in the texts belonging to this category.
The purpose of the writer and the intended audience necessitate con-
straints. In the HC the following distinct text types have been distinguish-
ed: homily, sermon,20 rule, religious treatise.
In the English speech-community the language of religion has had a
special position from the outset. That religious language depends on tradi-
tion is recognized both by believers and non-believers. It is largely based
on foreign models. In the older stages much of the vocabulary was Latin
or anglicized Latin, and partly French. The language of religion is char-
acterized by linguistic archaisms which, owing to the sacredness of the
subject matter, tend to keep it apart from ordinary speech and writing. In
the course of time the linguistic effect of religious terminology and phra-
seology has been evident in all kinds of literature.
During the period of our study the various stages in the general de-
velopment of English are apparent in the language of religion. Theo-
logical or learned religious texts typically contain obscure, difficult or
metaphorical words, phrases, clauses and sentences that need to be clari-
fied by some means. Terms belonging to the common core terminology
such as God, Christ, Virgin Mary, Heaven and the Church may be refer-
red to by lexically common but semantically special terms. Since there
are other words or phrases that are both lexically and semantically
special, religious texts typically contain expositions of difficult passages
such as paraphrases of theological terms or metaphors, or translations of
foreign words and phrases. There is usually some shared knowledge
between the author of religious texts and his customary audience, and a
 Expository apposition 155

writer aims at keeping the balance between the ordinary and the difficult-
to-grasp.
The peak use of expository apposition with an explicit marker falls in
the transition periods ME4 and EModEl, the mean incidence being 4.48
and 6.18 per 1000 words respectively. There is, however, a lot of varia-
tion both between different text types and individual texts, which will be
discussed in more detail in the following two sections.

3.1.3.1. Late Middle English religious instruction

All the four text types pertaining to the category of religious instruction
in the HC are represented in the Late Middle English subsections. The
text type labelled homily is only found in ME3. The only sample of homi-
lies in the HC, an extract from the Northern Homily Cycle, contains very
few appositions. It is written in verse, which may be one of the main rea-
sons for the scarcity of appositions.
Individual sermons included in the Late Middle English subsections
differ greatly from each other in their use of apposition. In ME3 this text
type is represented by Wycliffite sermons. Wycliff and his followers
based their sermons solely on the text of the Bible. They saw English as
the language of instruction and paid less attention to style, not to mention
any technical theological explanations. Their language was a continuation
of Alfredian speech-based prose with hardly any appositions. Early ser-
mons such as those included in John Mirk's Festial (ME3/4) were of the
type sermo ad populum, an edifying discourse without any specific spir-
itual theme. They were addressed to 'good men and women', and con-
tained few expository appositions (only 3.8 per 1000 words).
In the fifteenth-century sermons included in the HC the use of apposi-
tion is more frequent. Gaytryge's Sermon (ME3/4) introduces the articles
of faith to laymen as a response to an episcopal command to educate
them in the faith and Christian virtues (57).

(57) The fyfte comandement byddes vs pat we sla na man, fiat es to

say, bodyly ne gastely noper (Gaytryge, Sermon 6)
'The fifth commandment bids us not to kill any man, that is to
say, either in body or in spirit.'

The incidence of expository apposition with a marker in Gaytryge's Ser-

mon is 12.43 per 1000 words, and and/or comprises 69 per cent of the
156 Päivi Pahta and Saara Nevanlinna

markers. In the Sermons of Fitzjames (ME4), who was a middle-aged

professional preacher of high standard, the incidence of markers is 9.84
per 1000 words, and appositions with and/or comprise 82 per cent of the
markers. Fitzjames' tone of voice in many of the appositions is similar to
that of Gaytryge's (58).

(58) Declyne thou neuer fro them [i.e., the commandments] ne in to

Ϋ ryght honde, ne yet in to the lyfte honde, that is to saye, ne in
prosperyte, ne yet in aduersyte. (Fitzjames, Sermo Die Lüne
A4R)
'Don't ever give up them, neither on the one hand, nor on the
other, that is to say, neither in prosperity, nor in adversity.'

The religious treatises included in the Late Middle English subsections

of the HC show a great deal of variation in their use of apposition. Most
of the religious treatises of the period are translations or adaptations from
Latin or French models. The former are generally regarded as the better
language. The translations from French were influenced by the new, in-
coherent trailing-sentence structure.
The sample from Rolle's translation of the Latin Psalter comprises 157
verses. Rolle translates 'sense for sense', and his text does not seem to
contain any additions in the form of apposition. In his commentary (from
ME2/4) attached to those verses there are, however, altogether 190 in-
stances of expository apposition with a marker, the incidence being 18.30
per 1000 words in the text, which is the highest among the Late Middle
English religious texts. In the majority of cases (162) the marker that is
is used. Only six appositions with that is belong to the class of identifi-
cation. The rest, as well as all the other appositions, represent paraphrase.
The second appositive is usually a lexical synonym or an exposition of a
difficult or metaphorical phrase or clause. Rolle's commentary is closely
modelled on the Latin commentaries of the Church Fathers, such as Peter
Lombard.
In The Book of Vices and Virtues (ME3/4), which was translated from
French, the incidence of expository apposition is lower, 11.00 per 1000
words. The same incidence is seen in Purvey's Prologue to the Bible
(ME3), which is a vernacular text, addressed to the lay reader. Apposi-
tions belonging to the semantic class of identification are fairly frequent
in it (59).
 Expository apposition 157

(59) and eft Dauith seith, "the Lord schal telle in the scripturis of
puplis and of these princis that weren in it, that is, in holi
chirche". (Purvey, The Prologue to the Bible 56)
'And then David says, "the Lord will tell in the scripture about
people and about the rulers that were in it, that is, in holy
church".'

In The Cloud of Unknowing (ME3) composed in the vernacular by an

unidentified mystic writer for an apparently select readership with a great
deal of shared knowledge, the incidence of markers is only 2.74 per 1000
words.
The native coordinative conjunctions and/or are much less frequent as
apposition markers in translations of religious treatises than in original
English texts. In Rolle's Psalter commentary, and/or occurs only in 7 per
cent of the appositions. The corresponding percentage in The Book of
Vices and Virtues is 37, whereas it is 70 in The Cloud, and as high as 74
in Purvey's Prologue, both texts being directly addressed to their audi-
ence.
The two independent translations of Aelred of Rievaulx 's De Institutione
Inclusarum classified as religious rules are also different from each other
in their use of apposition. In ME3 samples, which derive from Latin, the
incidence of apposition is very low (1.30 per 1000 words), while in ME4
samples, which contain a translation of the whole work from French, it is
considerably higher (6.63 per 1000 words).

3.1.3.2. Early Modern English sermons

By the end of the fifteenth century the influence of the Renaissance had
reached England, and Latin received a fresh boost. The vocabulary of
English was greatly enriched from classical Latin sources. The newly
introduced terms necessitated a great deal of explanation and interpreta-
tion.
Throughout the Middle Ages rhetoric had remained central to the
culture. Since rhetoric was specially designed to give emotional or imagi-
native weight to the authoritative truth advocated by Church and State,
rhetorical conventions could be found in all kinds of texts. When classical
texts were rediscovered during the Renaissance, rhetoric was revised as
a complete system. Aristotle's main means of persuasion (ethos — per-
suasion through personality and stance, pathos — persuasion through
158 Päivi Pahta and Saara Nevanlinna

emotion, and logos — persuasion through logic and reasoning) were im-
portant principles followed by public preachers and orators (Cockcroft—
Cockcroft 1994 passim).
The only text type of religious instruction in the Early Modern English
subsections in the HC is sermon. Arousing emotion was especially im-
portant for a sermon to be effective, but it was also important that the
audience should comprehend the message. Educated, largely bilingual hu-
manist preachers addressed their lay audiences in a new type of language
combining native simplicity and humanist latinity. The Early Modern
English sermon-writers were well-versed in rhetoric and highly experi-
enced in professional preaching. In the compilation of the HC, the ser-
mons in each Early Modern English subsection were selected to represent
two different stylistic strata, popular and more formal (see Nevalainen—
Raumolin-Brunberg 1993: 63).
The first Early Modern English subperiod is the peak period in the use
of expository apposition. Although the preachers whose sermons are in-
cluded in EModEl, Latimer and Fisher, are usually presented as being
very different from each other, the incidence of expository apposition in
the samples of their sermons is almost the same (nearly 8 per 1000
words). Among seven variant apposition markers that were found in Bish-
op Latimer's sermons, and/or comprised 78 per cent of the total. Bishop
Fisher's sermons contained four variant markers, out of which and/or
comprised only 40 per cent. Catholic lay audiences accustomed to Latin
readings from the altar probably expected less explanation by way of
paraphrase. Examples (60) and (61) show Latimer's and Fisher's typical
ways of using appositions.

(60) The one is an inclosinge to let or hinder ye bodily ploughynge,

and the other to lette or hynder the holiday ploughyng, the
church ploughinge ... no man wyll herken it but to hinder and
let it (Latimer, Sermons 25)

(61) These fyue great woundes were ingraued wyth sharpe & vyolent
pennes, that is to say, the sharpe nayles, and the speare. (Fisher,
The English Works of John Fisher 1 396)

In EModE2, the incidence of markers was very low in the sermons of

Henry Smith, only 1.73 per 1000 words and there were no and/or apposi-
tions. Richard Hooker had very strong feelings for the Church of England
and he seems to have been a true rhetorician. Paraphrase with and/or
 Expository apposition 159

comprised 86 per cent of the markers, the incidence of markers being

5.48 per 1000 words.
In EModE3, the incidence of markers in the sermons of the Catholic
preacher Jeremy Taylor was only 3.24 per 1000 words, the markers
and/or comprising 68 per cent of the total. In Archbishop Tillotson's ser-
mons the incidence of expositor apposition was 5.00 per 1000 words.
His sermons had their emotional appeal enhanced by appositional para-
phrases with and/or, which constituted as much as 94 per cent of the
total.
The above shows that the great protestant sermon-writers were style-
conscious. It has been argued that their sermons resemble the persuasive
homiletic style of ^ l f r i c and Wulfstan, who often coupled synonyms and
near-synonyms with the native markers and and or.

3.1.4. Philosophical treatises

There is some interesting variation between the philosophical treatises in-

cluded in the HC. In Chaucer's translations of Boethius' De consolatione
philosophiae (ME3) after Latin and French models, and the Tale of Meli-
bee from French, the incidence of markers is 6.78 and 7.90 per 1000
words respectively (cf. Blake 1992a: 504). Being a highly professional
writer and very versatile, Chaucer makes ample use of coordinative ap-
position with and/or as a stylistic device, as in (62).

(62) "A," quod dame Prudence, "ye seyn youre wyl and as yow lyk-
eth, but in no caas of the world a man sholde nat doon outrage
ne excesse for to vengen hym. (Chaucer, The Tale of Melibee
232.C1)
'"Ah," said Dame Prudence, "you speak your will and as you
like it, but in no case should a man take excessive revenge.'

And/or occurs in 81 per cent of the instances in Boethius, and in 78 per

cent in Melibee. Chaucer's main reference points in his Boethius were the
Vulgate, Jean de Meun's French translation and Nicholas Trevet's Com-
mentary. Chaucer followed in the footsteps of Trevet by providing glosses
to his vernacular text (Minnis 1993: viii, 173). He wanted to be faithful
to the sense (sententia), but also followed Jerome's other principle in his
translation of the Bible: word for word. Expository comments and gloss-
es, like in example (63), were considered to be part of the whole.
160 Päivi Pahta and Saara Nevanlinna

(63) telle me, yif thow accordest to leten no torment to the soules
aftir that the body is ended by the deethl (This to seyn, "Undir-
stondestow aught that soules han any torment aftir the deeth of
the bodyV (Chaucer, Boethius 448.C1)
'tell me whether you consent to assigning no torture to the souls
after the body ends in death? (This is to say, "Do you believe at
all that souls are tormented after the death of the body?")'

The HC contains several vernacular versions of Boethius; besides Chau-

cer's Boethius the samples include three Early Modern English trans-
lations. In George Colville's translation of Boethius (EModEl) the influ-
ence of the Renaissance is clearly to be seen, as in example (64). The in-
cidence of expository apposition is as high as 26.49 per 1000 words. The
markers and/or are contained in 88 per cent of all the appositions.

(64) Therfore these worldlye thynges, that is to saye, worldelye suffi-

saunce, power reuere~ce nobles and pleasures, semethe to geue
vnto men the symylitudes or lykenes, of true good, or ells, to
geue certaine vnperfit and fained goodes (Colville, Boethius 71)

Queen Elizabeth's translation from the original Latin of parts of Boeth-

ius in EModE2 provides the direct opposite to the earlier prose trans-
lations as far as apposition is concerned. Though her translation was done
only some thirty years after Colville's, the incidence of expository apposi-
tion is very low, only 0.87 per 1000 words. She no longer had the medi-
eval attitude towards comments and glosses; her text is a slavish trans-
lation from Latin. The only apposition markers occurring in the Queen's
translation are and and or within the class of paraphrase.
The Right Honourable Richard Preston translated Boethius from Latin
towards the end of subperiod EModE3 (1695). The incidence of apposi-
tions, 5.33 per 1000 words, is lower than in Colville, but markedly higher
than in Queen Elizabeth's translation. The markers and and or, constitut-
ing 96 per cent of all the markers, occur almost exclusively in lexical
paraphrase. These results show how idiosyncratic preferences can be re-
flected in the translation of one and the same work.
 Expository apposition 161

3.1.5. Educational treatises

Six educational treatises are included in the Early Modern English section
of the HC, two in each subperiod. Their external aims divide them into
two groups. Each subperiod is represented by one educational treatise
with more theoretical aims and one with purely practical goals (Nevalai-
nen—Raumolin-Brunberg 1993: 64). The first group consists of Thomas
Elyot's The Boke Named the Gouernour (EModEl), Francis Bacon's Ad-
vancement of Learning (EModE2) and John Locke's Directions Concern-
ing Education (EModE3), and the second group contains Roger Ascham's
The Scholemaster (EModEl), John Brinsley's Ludus Literarius, or the
Grammar Schoole (EModE2) and Charles Hoole's A New Discovery of the
Old Art of Teaching Schoole (EModE3). The mean incidence of ap-
positions is highest in the earliest texts and declines steadily towards the
end of the period (7.66 appositions per 1000 words in EModEl, 5.79 in
EModE2, and only 2.30 in EModE3). There is, however, also synchronic
variation within the text type.
In the first two subperiods the use of expository apposition with a
marker is clearly different between the more theoretical discussions and
the more practically oriented treatises. In EModEl, Elyot's text contains
12.28 appositions per 1000 words, whereas in the same amount of text
Ascham only uses 2.49 appositional constructions. Appositional construc-
tions seem mainly to be used for stylistic purposes by both writers, as in
example (65) from Elyot.

(65) Ο what misery was it to be subiecte to suche a minstrell, in

whose musike was no melodye, but anguisshe and dolour?
(Elyot, The Boke Named the Gouernour 27)

Similar trends continue in EModE2: there are 8.27 appositions per 1000
words in Bacon's more theoretical text (66) and the corresponding figure
in Brinsley's dialogue-form treatise is only 3.25 (67).

(66) So Readers in Scyences are indeede the Gardyans of the stores

and prouisions of Scyences, whence men in actiue courses are
furnished (Bacon, Advancement of Learning 4R)
162 Päivi Pahta and Saara Nevanlinna

(67) Prove and confirme what tongue soever your Scholler learnes,
even from the first reading of English (Brinsley, Ludus Lite-
rarius, or the Grammar Schoole 44)

The situation is different in EModE3, where both educational treatises

contain only few appositions: 1.92 per 1000 words in Locke's more theo-
retical treatise and 2.61 in Hoole's text. But like in the earlier texts, the
majority of appositions in Locke's and Hoole's writing are paraphrases
consisting of repetitive word-pairs joined by and or or used either for rhe-
torical or instructional purposes, as in example (68).

(68) Now there are five organs or instruments of speech, in the right
hitting of which, as the breath moveth from within, through the
mouth, a true pronunciation of every letter is made (Hoole, A
New Discovery of the Old Art of Teaching Schoole 2-3)

Further examples of appositions containing and/or in the educational trea-

tises are provided by pairs like accepting or admitting, delites and pleas-
ures, deceit or vntruth, properlie and fitlie and trayne and exercise. Other
markers are rare and they are mainly used in the theoretical treatises, as
in example (69).

(69) If what I have said in the beginning of this discourse be true as

I doe not doubt but it is, viz. That the differences to be found in
the manners and abilitys of men is oweing more to their Educa-
tion then anything else, we have no reason to conclude that
great care is to be had of formeing childrens mindes (Locke, Di-
rections Concerning Education 49)

3.2. The frequency of semantic classes across text types

The different semantic classes of expository apposition were unevenly

distributed in the data. The majority of the appositional constructions with
an explicit marker belonged to the semantic type in which the two units
of apposition are equally specific, i.e., the classes of paraphrase and revi-
sion; c. 93 per cent of the appositions in our study belonged to this type.
The overall distribution of the five semantic classes in our data is shown
in Table 4 on p. 134 above, and Tables 10-14 (see pp. 163-166) show
the distribution of the different classes in each subperiod.
 Expository apposition 163

In the following sections we will briefly comment on the occurrence of

the five different semantic classes in our data.

Table 10. Semantic classes in ME3 (1350-1420; N/1000 words). IDE =

identification, APP = appellation, CHAR = characterization, PAR =
paraphrase, REV = revision.

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

IS Science, medicine 0.04 16.20

STA Document 8.55
XX Philosophy 0.40 0.08 6.97 0.87
IR Religious treatise 0.16 0.03 2.91 0.49
XX Official letter 2.99
XX Document 0.58 2.38
IS Handbook, astronomy 0.70 0.08 1.84 0.31
IR Homily 1.24 0.41
NI Fiction 1.47
NN History 1.27 0.04
IR Rule 0.37 0.74 0.19
IS Handbook, medicine 0.67 0.49
NN Travelogue 1.09
XX Bible 0.19 0.05 0.48 0.10
IR Sermon 0.07 0.37 0.07

Table 11. Semantic classes in ME4 (1420-1500; N/1000 words).

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

IS Science, medicine 1.24 8.71

STA Law 0.09 9.70
IR Religious treatise 0.44 0.02 8.01
IR Sermon 0.40 0.08 0.04 6.04 0.24
IR Rule 1.66 4.97
XX Preface 0.16 5.00 0.48
XX Official letter 0.32 5.10
IS Handbook, other 1.56 2.94 0.09
XX Document 0.19 2.96
164 Päivi Pahta and Saara Nevanlinna

Table 11. Continued.

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

IS Handbook, medicine 0.35 2.47

NI Fiction 0.11 1.94
XX Deposition 2.03
XX Bible 1.58 0.40
NN History 0.16 1.35
NN Saint's life 1.05 0.26
XX Mystery play 0.95
XX Private letter 0.87
NI Romance 0.11 0.52
IS Handbook, astronomy

Table 12. Semantic classes in EModEl (1500-1570; N/1000).

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

XX Philosophy 1.31 25.18

STA Law 25.53 0.17
EX Science, medicine 0.81 16.51
IS Handbook, other 0.90 0.10 10.60
IR Sermon 0.21 7.18 0.53
IS/EX Education 0.19 7.38 0.10
EX Science, other 2.69 3.43
XX Trial 0.44 0.13 2.94 0.19
NN History 0.09 0.09 1.80
NN Biography 1.65
NI Fiction 1.56 0.09
XX Official letter 1.59
NN Travelogue 1.49 0.07
XX Comedy 0.85 0.19
NN Autobiography 0.17 0.70
XX Private letter 0.75 0.09
NN Diary 0.61
XX Bible 0.09 0.52
 Expository apposition 165

Table 13. Semantic classes in EModE2 (1570-1640; N/1000).

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

STA Law 0.34 17.83 0.59

EX Science, medicine 1.21 0.15 9.82 0.15
EX Science, other 1.56 0.31 4.36 0.31
EX Education 0.18 0.09 5.17 0.36
NN Travelogue 0.07 0.34 3.45 0.07
IR Sermon 0.29 3.20 0.19
IS Handbook, other 0.24 3.01 0.33
NN Autobiography 0.24 0.24 2.43 0.24
NN History 0.10 2.48
NN Biography 0.21 0.83 0.42
XX Official letter 1.06 0.18
XX Trial 0.07 1.05 0.07
XX Comedy 0.76 0.34
XX Philosophy 0.87
XX Bible 0.05 0.63
NI Fiction 0.56
XX Private letter 0.09 0.35
NN Diary 0.16

Table 14. Semantic classes in EModE3 (1640-1710; N/1000).

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

STA Law 0.46 16.09

XX Philosophy 0.28 5.10
EX Science, other 0.44 0.09 4.34 0.44
IR Sermon 0.08 3.93 0.16
IS Handbook, other 0.26 0.09 2.55 0.35
NN Travelogue 2.77 0.19
IS Education 0.09 1.86 0.35
NI Fiction 1.50 0.58
XX Official letter 0.17 1.70
NN History 1.20 0.43
166 Päivi Pahta and Saara Nevanlinna

Table 14. Continued.

Text Text IDE APP CHAR PAR REV

categ. type /1000 /1000 /1000 /1000 /1000

NN Biography 1.46
XX Trial 0.58 0.07
NN Diary 0.09 0.45 0.09
XX Private letter 0.15 0.08 0.23 0.08
XX Comedy 0.24 0.16
NN Autobiography 0.18 0.18

3.2.1. Paraphrase

The most frequently occurring semantic class in our data is paraphrase

(see p. 137 above), instances of which are found in all text types and
practically all texts. It is particularly frequent in texts belonging to the
fields of legal, medical or philosophical writing. The linguistic situation
in medieval and early modern England created a need for this type of
usage. It was often necessary to provide an explanation for words of for-
eign origin or dialectal variants (see examples (25) and (28) above). Co-
ordinative apposition consisting of synonymous or near-synonymous
word-pairs linked by and or or was also a common stylistic or rhetorical
device, and is accordingly found in texts which are not otherwise rich in
appositions (example (24) above). It is precisely this construction which
makes paraphrase so much more frequent than the other semantic classes
of apposition.
Although the majority of paraphrases consist of lexical apposition, there
are also a considerable number of longer paraphrases used for interpreting
complex propositions to make the text more accessible to the readers or
listeners (see examples (34) and (35) above). Both forms of paraphrase
contribute to success in communication (cf. Blakemore 1993: 109).

3.2.2. Revision

In revision (see p. 140), two subtypes serving different communicative

functions can be distinguished. The first subtype, which can be labelled
reorientation, is a fairly rare type of apposition in our data. It is mainly
found in religious or philosophical texts in a context where the speaker or
 Expository apposition 167

writer reviews a phenomenon from a different standpoint (examples (38)

and (39) above). The second subtype can be labelled self-correction. It is
more typical of spoken language and therefore mainly found in texts
which are speech-based, such as homilies or sermons and drama. It also
occasionally occurs in legal or scientific texts in order to provide a more
appropriate expression for a lexical item or proposition. In educational
treatises and narrative texts it is used style-consciously either to provide
emphasis or to communicate the author's point of view by way of a meta-
linguistic comment, or to supply the reader with a range of implications
(examples (42) and (43) above; cf. also Blakemore 1993: 111).

3.2.3. Identification

The semantic class of identification (see p. 134) is found in several types

of texts, but there is a tendency for appositions of this kind to occur more
frequently in texts that can be labelled instructional or expository, such as
handbooks, scientific treatises, or religious rules. As in Present-day Eng-
lish, appositions in the class of identification are mostly connected with
enumerating or listing items or procedures (example (17) above; cf.
Meyer 1992: 121).

3.2.4. Appellation

The appositions with an explicit marker belonging to appellation (see p.

135) were rare in our data. In Present-day English this type of apposition
is very common in journalistic writing (Meyer 1992: 115). In our data the
scattered examples are found in a variety of text types including narrative
texts, such as biography, travelogue and history, i.e., in texts in which
naming individual people or places is an essential part of communication
(example (21) above), but also in scientific treatises and handbooks.

3.2.5. Characterization

Providing characteristics of individuals or incidents by way of apposition

with a marker is rare in our data. Characterizing apposition, with the sec-
ond unit providing less specific information, is perhaps a type of con-
struction where the use of a specific marker is more awkward than in the
other semantic classes (cf. p. 136 above). The rare instances of charac-
168 Päivi Pahta and Saara Nevanlinna

terization are found in instructional or expository texts (example (22)

above). In Present-day English characterization is, like appellation, com-
mon in the press genre, where it is important to both name and attribute
characteristics to individuals about whom the audience has little know-
ledge (Meyer 1992: 116).

4. On the use of the markers of expository apposition

This section summarizes briefly some of the general tendencies seen in

the use of explicit markers of expository apposition in our data. A more
detailed discussion of the topic is contained in Pahta—Nevanlinna forth-
coming.
The dominance of expository appositional constructions with the mark-
ers and and or has already been pointed out and discussed in connection
with some text types (pp. 147-162 above). These two markers occurred
in practically all text types and individual texts included in our study, al-
though there is a lot of variation in their frequency across different text
types and individual texts as well as across time. They occur in well over
half of all appositional constructions in the Late Middle English sub-
periods (75 per cent in ME3 and 57 per cent in ME4), making up almost
90 per cent of the total number of the markers in each of the Early Mod-
ern English subsections of the corpus (EModEl 87; EModE2 89; EModE3
89). The texts in which they are most frequently used in our data belong
to the fields of law, science, education, religion and philosophy (for ex-
amples, see section 3.1).
The other markers of expository apposition were found to be far more
restricted in their scope, and they can be divided roughly into two types,
although there is considerable overlap. Markers like that is (to say), (that
is) to wit, namely, and particularly the Latin markers seem to be most
common in fairly formal or technical instructive or expository texts (ex-
amples (70) and (71)), whereas markers like nay and yea, which border
on pragmatic particles, are mainly found in narrative or speech-based
texts (examples (72) and (73)).

(70) they tende to one conclusion or purpose, that is to saye, with a

free and glad wyll to gyue to a nother that thinge which he be-
fore lacked. (Elyot, The Boke Named the Gouernour 148-149)
 Expository apposition 169

(71) if thou list to dissolue & cast out Phlegme, these Pils following
must be taken, viz. Pillulae de Sagap: de Opopan, de Elleboro,
de Euphorbio. (Clowes, Treatise for the Artificiall Cure of
Struma 10)

(72) I say unto thee again, make haste, nay, run, run unto the Breth-
ren. (Penny Merriments 150)

(73) that he sholde after his labour haue full reste. ye & for shorte
labour haue eternal reste (Fitzjames, Sermo Die Lüne A2R)

Our study also showed that most of the markers which are available in
contemporary English for explicitly indicating expository apposition exist-
ed in Late Middle English. The only marker current today which was not
found in our data by 1710 is in other words (see p. 128 and note 9).
Another marker, namely, which is also common in Present-day English as
an expository marker, was undergoing a semantic change during the pe-
riod studied. In the earliest samples of our data it is used as a marker of
exemplifying apposition (74) and the first instances of namely in exposi-
tory apposition date from the fifteenth century (75).

(74) but loke that he examyne truli his Latyn bible, for no doute he
shal fynde ful manye biblis in Latyn ful false, if he loke manie,
namely newe. (Purvey, The Prologue to the Bible 57-58)
'But he should take care to examine closely his Latin Bible, for
undoubtedly he will find many Latin Bibles quite in error, if he
examines several, especially new ones.'

(75) Pis word 'oure' teche}) vs to hate pre pinges, namely: pride,
hate, couetise. (The Book of Vices and Virtues 101)
'This word "our" teaches us to hate three things, namely: pride,
hate, avarice.'

In addition to the markers used in contemporary English, our study also

revealed a number of phrases which have disappeared or become rare,
such as nay, yea or as who say, as in example (76) (for others see Table
3 on p. 129).
170 Päivi Pahta and Saara Nevanlinna

(76) no thing nis leveful to folye in the reaume of the devyne pur-
veaunce (as who seith, no thing nis withouten ordenaunce in the
reame of the devyne purveaunce) (Chaucer, Boethius 454.C1)
'Nothing is allowable to folly in the domain of divine provi-
dence, in other words, nothing is without divine arrangement in
the domain of divine providence.'

The Latin markers scilicet, id est, and viz. and their variants occurred in
both Late Middle English and Early Modern English texts. They were
most frequent in appositions where one of the appositional units was
Latin and the other, vernacular (77). Some of them occurred in fully Latin
passages, but also in fully English passages, as in (78).

(77) in Jje holownesse of J>e lyuere goo{> oute a veyne icleped porta
(i. pe gate) (The Cyrurgie of Guy de Chauliac 62)
'in the hollow cave of the liver a vein goes out which is called
porta, i.e., gate'

(78) His daughter, viz his 2J daughter is upon recovery so as to

come (Henry Oxinden, The Oxinden and Peyton Letters 277)

5. Conclusion

In this study our aim has been to shed some light on the structure and use
of nonrestrictive expository apposition containing an explicit marker dur-
ing the period 1350-1710 and to relate our findings to the situation in
Present-day English. As far as the structure of expository apposition is
concerned, the results of our study show that the basic syntactic types
current in Present-day English, i.e., lexical, phrasal, clausal and sentential
apposition, can also be found in the period of our study. As today, ap-
position is typically a relation between two noun phrases.
Similarly, the various semantic classes distinguishable in contemporary
English (identification, appellation, characterization, paraphrase and revi-
sion) can also be detected in Late Middle and Early Modern English
texts. The dominant semantic class in our data was paraphrase. Several
factors seem to motivate re-phrasing in the form of apposition and par-
ticularly the use of paraphrase in our material. During the period of our
 Expository apposition 171

concern, a large amount of new lexical items were introduced to the lan-
guage from Latin and French, and there was a need for writers to pro-
vide their audience with a native synonym or near-synonym of loanwords
in order to ensure successful communication. The borrowed lexicon
which needed to be explained included words from all spheres of life, but
in the data of this study, providing translations for technical terminology
is one of the functions in which apposition containing an explicit marker
is particularly frequent. In Middle English texts dialectal variation also
provided a source of frequent use of apposition. Although the rise of
standardisation began in late fourteenth-century London, in Late Middle
English there was still a lot of regional and orthographic variation, es-
pecially by scribes who worked far away from the London region and
among writers whose families had migrated to London from other parts
of the country. Like foreign words, dialectal forms could be glossed by
adding an explanation in the form of apposition. The need for an explana-
tion is usually the source of apposition in clausal and sentential para-
phrase, too. Religious and philosophical texts particularly often contain
passages where complicated propositions are made more accessible to the
audience by adding an appositional explanation.
A different source of apposition is provided by stylistic considerations.
The use of coordinative apposition consisting of synonymous or near-
synonymous word-pairs joined by and or or is a well-established stylistic
device going back to the persuasive rhetorical style of Anglo-Saxon writ-
ers like yElfric and Wulfstan. Apposition with and/or dominates through-
out the period of our study. It becomes prominent in writings of the
Renaissance and remains frequent during the whole Early Modern English
period. In some types of writing coordinative apposition has become for-
mulaic, e.g., contemporary legislative writing is characterized by formu-
laic use of archaic appositional constructions, some of which can be dated
back to Old English.
The study also shows that certain types of writing clearly favour the use
of appositional constructions or the use of particular semantic or syntactic
categories of apposition. Expository apposition with an explicit marker
was found to be most frequent in texts belonging to learned or formal
genres, such as scientific or statutory writings. The dominant category in
both types of writing is lexical paraphrase, and the use of apposition is
primarily motivated either by the need to explain technical terminology of
foreign origin or the need to achieve explicitness and precision in com-
munication. Expository apposition is also frequent in texts belonging to
the spheres of religious and secular instruction. Lexical paraphrase is also
172 Päivi Pahta and Saara Nevanlinna

common in these texts. Instructive texts, however, show a greater tenden-

cy to contain longer appositional paraphrases, which help the reader to
process information and profit from instruction.
In a diachronic perspective one the main lines of variation is connected
with the selection of different markers used for explicitly indicating ex-
pository apposition. Various types of unique phrases (see p. 129 above)
are found in this function in the first four subsections of the Helsinki
Corpus included in this survey (ME3-EModE2), being most common in
subperiod EModEl. By mid-seventeenth century (EModE3), the selection
seems to have become stable, and with the exception of in other words
(see p. 128 above), all the markers used for indicating expository apposi-
tion in Present-day English, are found in our data.

Notes

1. Hudson (1990: 314) notes that apposition tends to receive less attention
from theoretical linguists than it deserves.
2. We found the model presented by Charles F. Meyer (1992) very useful for
our study, and we would like to thank him for his valuable comments on
this article. For reviews of his approach see Oostdijk (1994); Pahta (1994);
Acuna (1996: 119-127).
3. In Investigating English style, Crystal—Davy (1974: 59, n. 13) point out
explicitly that they apply the term "apposition" only to identically function-
ing nominal groups and exclude all other structures which may be apposi-
tional in "some notional sense".
4. The London-Lund Corpus of Spoken British English (see Svartvik—Quirk
1980), the Survey of English Usage Corpus of Written British English (see
Greenbaum 1985), and the Brown University Standard Corpus of Present-
day American English (Kucera—Francis 1967). Approximately 120,000
words from each corpus were included in the study (for details see Meyer
1992: 7-8).
5. In Meyer's study of apposition in Present-day English, the majority of ap-
positions (71 per cent) were either nonrestrictive nominal appositions or
non-nominal appositions for which the notions restrictive and nonrestrictive
were irrelevant (Meyer 1992: 82).
6. These are the two main additive functions of appositional constructions
according to Halliday—Hasan (1976: 248-250).
7. Mitchell (1985, 1: 596-612) discusses only paratactically joined apposi-
tions. See also Nevanlinna—Pahta (forthcoming).
 Expository apposition 173

8. The Old English interlinear glosses in several Latin Psalter manuscripts

(e.g. the Vitellius and Canterbury Psalters) contain numerous instances of
phrasal appositive structures in the form of double, triple, or, rarely, quad-
ruple glosses expressing synonymy. The marker of apposition is an I with
a slanting cross-bar short for Latin vel (see Nevanlinna 1984).
9. The earliest example is

for to say, that having in any quantity measured so much, or gone

so far, you are not yet at the end, is only to say, that that quantity
is greter. So that the negation of an end in any quantity is, in other
words, only to say, that it is bigger. (Locke 068/067-P16)

For information on the Century of Prose Corpus see Milic (1990).

10. Of the markers of apposition included in this study, that is and that is to
say can either precede or follow the second appositive in Present-day Eng-
lish, whereas namely, or (rather/better) and the abbreviated forms i.e. and
viz. can only precede the second appositive (Quirk et al. 1985: §17.73).
11. Cf. Huddleston's (1971) view that in Present-day English no apposition
marker is used in characterizing apposition, pp. 123-124 above.
12. In this category our classification differs both from Quirk et al. (1985) and
Meyer (1992). We divide appositions of this type into two semantic classes,
Meyer (1992) distinguishes three classes, and Quirk et al. (1985) have one
class with four subtypes.
13. As regards abstract nouns in apposition it is not always possible to decide
whether they represent one quality or two.
14. In this to sayn as who saith, is is deleted through haplology (see Pahta
—Nevanlinna forthcoming).
15. Quirk et al. (1985) classify revision as a subcategory of reformulation,
whereas we decided to discuss it as a separate class with two subclasses of
its own.
16. In Meyer's study, spoken genres tended to contain fewer appositions than
written genres (Meyer 1992: 126). This reflects Biber's (1988: 46) charac-
terization of spontaneous speech, together with personal letters, as com-
munication between speakers with a high degree of shared knowledge,
whereas most written texts are produced by discourse participants with little
shared knowledge.
17. Different aspects of legal language are discussed e.g. in Charrow (1982),
Crystal—Davy (1974), Danet (1980), Gustafsson (1975), Hiltunen (1990),
Maley (1987), and Mellinkoff (1963) and (1982).
174 Päivi Pahta and Saara Nevanlinna

18. This role of legislation as canonical text has been suggested as originating
in fourteenth-century England, when the function of the judiciary became
separate from that of the King's Council. The judges began to interpret
statutes strictly — not only as "suggestions of policy within whose broad
limits the court can exercise a broad discretion" (Plucknett 1940; cited after
Maley 1987: 28).
19. For background information in this section we are indebted to Blake
(1992a), Crystal—Davy (1974), Drabble (1986), Gordon (1972), Edwards
(1984), Jacob (1993), Leith (1987), Swales (1990), and Webster (1988). See
also Nevanlinna et al. (1993) and Nevalainen—Raumolin-Brunberg (1993).
20. In Middle English texts, the distinction between the text types 'sermon' and
'homily' was observed until the last third of the fourteenth century (Heffer-
nan 1984: 179). Later, the two terms were at times indiscriminately applied
to either of the text types.
To the medieval mind, homilia was "a name applied to an edifying
discourse based on a text of sacred scripture that is itself traditionally tied
to a particular liturgical day or event" (Heffernan 1984: 179). In the Middle
English period, homilies were assigned to the Sundays or other festival
days, even to some weekdays, mostly according to the use of York or Salis-
bury. In the Early Modern English period, special Books of Homilies were
published by the Church of England in 1547 and 1563 for the use of parish
churches (OED).
The purpose of the medieval sermon (sermo ad populum) was to teach
the believers the truths of scripture and the faith. There was a considerable
variety in form (prose - verse) and in theme (religious - popular). Con-
nection to a liturgical event was not obligatory. Rhetorical techniques were
often borrowed from the artes predicandi. References to topical events
were rare. Exempla derived from various sources were attached to sermons
and homilies, which was criticized severely by the Wycliffites (Heffernan
1984).

References

Acufia, Juan Carlos

1996 The puzzle of apposition. On so-called appositive structures in Eng-
lish. (Monografias de Universidade de Santiago de Compostela 190.)
Santiago de Compostela: Universidade de Santiago de Compostela,
Servicio de Publicacions e Intercambio Cientifico.
Aijmer, Karin—Bengt Altenberg (eds.)
1991 English corpus linguistics. Studies in hounour of Jan Svartvik. Lon-
don—New York: Longman.
 Expository apposition 175

Allwood, Jens—Lars-Gunnar Andersson—Osten Dahl

1977 Logic in linguistics. London—New York—Melbourne: Cambridge
University Press.
An Anglo-Saxon dictionary
1955 (Based on the manuscript collection of Joseph Bosworth. Edited and
enlarged by T. Northcote Toller.) London.
AV = The Authorized Version of the English Bible
1953 London.
Bennett, H. S.
1952 English books & readers 1475 to 1557. Being a study in the history
of the book trade from Caxton to the incorporation of the Stationers'
Company. Cambridge: Cambridge University Press.
Biber, Douglas
1988 Variation across speech and writing. Cambridge: Cambridge Uni-
versity Press.
Blake, Norman F.
1992a "The literary language", in: Norman Blake (ed.), 500-541.
1992b "Translation and the history of English", in: Matti Rissanen et al.
(eds.) 1992: 3-24.
Blake, Norman F. (ed.)
1992 The Cambridge history of the English language. Vol. 2. Cambridge:
Cambridge University Press.
Blakemore, Diane
1993 "The relevance of reformulations", Language and Literature 2: 101—
103.
Blockley, Mary
1989 "Old English coordination, apposition, and the syntax of English
poetry", Modern Philology 87/2: 115-131.
Bloomfield, Leonard
1965 Language. (Reprinted.) London: Allen and Unwin.
Bolinger, Dwight
1979 Meaning and form. (English Language Series 11.) London—New
York: Longman.
The Book of Common Prayer.
1945 (Reprinted.) New York: The Church Pension Fund.
Burnley, David
1992 "Lexis and semantics", in: Norman Blake (ed.), 409^493.
Burton-Roberts, N.
1975 "Nominal apposition", Foundations of Language 13: 391-419.
Chambers, R. W.—Marjorie Daunt (eds.)
1931 A book of London English. Oxford: Clarendon Press.
176 Päivi Pahta and Saara Nevanlinna

Charrow, Veda R.
1982 "Linguistic theory and the study of legal and bureaucratic language",
in: Loraine Obler—Lise Menn (eds.), 81-101.
Cockcroft, Robert—Susan M. Cockcroft
1994 Persuading people: An introduction to rhetoric. (Reprinted.) London:
Macmillan.
Collins Cobuild English language dictionary.
1987 Ed. in chief John Sinclair. London: Collins.
Cruse, D.A.
1986 Lexical semantics. Cambridge: Cambridge University Press.
Crystal, David
1985 A dictionary of linguistics and phonetics. (Second edition updated and
enlarged.) Oxford: Blackwell.
Crystal, David—Derek Davy
1974 Investigating English style. (5th edition.) London: Longman.
Curme, George O.
1931 Syntax. (A grammar of the English language, vol. 3.) Boston: D. C.
Heath and Co.
1970 English grammar. (Reprinted.) (College Outline Series 61.) New
York: Barnes and Noble.
Danet, B.
1980 "Language in the legal process", Law and Society Review 14/3: 445-
563.
Denison, David
1993 English historical syntax: Verbal constructions. London: Longman.
Drabble, Margaret (ed.)
1986 The Oxford companion to English literature. (5th edition reprinted
with corrections.) Oxford—New York—Tokyo—Melbourne: Oxford
University Press.
Edwards, Anthony S. G. (ed.)
1984 Middle English prose. A critical guide to major authors and genres.
New Brunswick, New Jersey: Rutgers University Press.
Forshall, John—Sir Frederik Madden (eds.)
1850 The Holy Bible containing the Old and New Testament with the Apo-
cryphal Books in the earlier English versions made from the Latin
Vulgate by John Wycliffe and his followers. Oxford: Oxford Universi-
ty Press.
Fries, Charles C.
1952 The structure of English. New York: Harcourt, Brace and Co.
 Expository apposition 177

Ghadessy, Mohsen (ed.)

1988 Registers of written English: Situational factors and linguistic fea-
tures. (Open Linguistics Series.) London—New York: Pinter Publish-
ers.
Gordon, Ian A.
1972 The movement of English prose. (English Language Series.) (Third
impression.) London-—Beccles—Colchester: Longman.
Greenbaum, Sidney
1985 "25th Anniversary of the Survey of English Usage", World Englishes
4: 261-265.
Gustafsson, Marita
1975 Some syntactic properties of English law language. Turku: University
of Turku.
Halliday, Michael A. K.—Ruqaiya Hasan
1976 Cohesion in English. London: Longman.
HC = Helsinki Corpus
1991 The Helsinki Corpus of English Texts. Helsinki: Department of Eng-
lish, University of Helsinki.
Heffernan, Thomas J.
1984 "Sermon literature", in: Anthony S. G. Edwards (ed.), 177-207.
Hiltunen, Risto
1990 Chapters on legal English: Aspects past and present of the language
of the law. (Annales Academiae Scientiarum Fennicae, Ser. B, Tom.
251.) Helsinki: Academia Scientiarum Fennicae.
Hiltunen, Risto—Marita Gustafsson—Keith Battarbee—Liisa Dahl (eds.)
1993 English far and wide: A festschrift for Inna Koskenniemi. (Annales
Universitatis Turkuensis, Ser. B, Tom. 197.) Turku: University of
Turku.
Hockett, C.
1955 "Attribution and apposition", American Speech 30: 99-102.
Hopper, Paul J.—Elizabeth Closs Traugott
1993 Grammaticalization. (Cambridge textbooks in linguistics.) Cambridge:
Cambridge University Press.
Huddleston, Rodney D.
1971 The sentence in written English. Cambridge: Cambridge University
Press.
Hudson, R.
1990 English word grammar. Oxford: Blackwell.
Jackendoff, Ray
1985 Semantics and cognition. (Current Studies in Linguistics Series.) (Sec-
ond printing.) Cambridge, Massachusetts—London: MIT Press.
178 Päivi Pahta and Saara Nevanlinna

Jacob, E. F.
1993 The fifteenth century 1399-1485. Oxford—New York: Oxford Uni-
versity Press.
Jespersen, Otto
1961 Α Modern English grammar on historical principles. Vol. 4. London:
Allen and Unwin.
Jones, Peter Murray
1989 "Four Middle English translations of John Ardeme", in: A. J. Minnis
(ed.), 61-89.
Kajanto, Iiro—Inna Koskenniemi—Esko Pennanen (eds.)
1983 Studies in classical and modern philology presented to Y. M. Biese on
the occasion of his eightieth birthday 4.1.1983. (Annales Academiae
Scientiarum Fennicae, Ser. B, Tom. 223.) Helsinki: Academia Scien-
tiarum Fennicae.
Kilpiö, Matti
1992 "Beon", in: Dictionary of Old English: Fascicle B. Published for the
Dictionary of Old English Project, Centre of Medieval Studies, Uni-
versity of Toronto. Toronto: Pontifical Institute of Mediaeval Studies.
Koskenniemi, Inna
1968 Repetitive word pairs in Old and Early Middle English prose. Expres-
sions of the type whole and sound and answered and said, and other
parallel constructions. (Annales Universitatis Turkuensis, Ser. B,
Tom. 107.) Turku: University of Turku.
1983 "Semantic assimilation in Middle English binomials", in: Iiro Kajan-
to—Inna Koskenniemi—Esko Pennanen (eds.), 77-84.
Kristensson, Gillis
1974 John Mirk's Instructions for parish priests. (Lund Studies in English
49.) Lund: CWK Gleerup.
Kuöera, Η.—W. Ν. Francis.
1967 Computational analysis of Present-day American English. Providence:
Dept. of Linguistics, Brown University.
LALME = A linguistic atlas of late mediaeval English. 4 vols.
1986 Ed. Angus Mcintosh—M. L. Samuels—Michael Benskin. Aberdeen:
Aberdeen University Press.
Leech, Geoffrey
1981 Semantics. (2nd edition.) New York: Penguin.
Leith, Dick
1987 A social history of English. (Language and Society.) (Reprinted.) Lon-
don—New York: Routledge and Kegan Paul.
 Expository apposition 179

Lewis, Charlton T.—Charles Short (eds.)

1975 A Latin dictionary founded on Andrews' edition of Freund's Latin
dictionary. (Revised, enlarged and in great part rewritten.) Oxford:
Clarendon Press.
Lyons, John
1979 Semantics. Vol. 2. (First published in 1977.) Cambridge—Lon-
don—New York—Melbourne: Cambridge University Press.
Maley, Yon
1987 "The language of legislation", Language in Society 16/1: 25-48.
Matthews, Peter H.
1981 Syntax. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge
University Press.
McMordie, W.
1976 English idioms and how to use them. (Revised by R. C. Goffin.
Fourteenth impression.) Hong Kong: Oxford University Press.
MED = The Middle English dictionary.
1952- Ed. Hans Kurath—Sherman M. Kuhn, et al. Ann Arbor, Michigan:
Michigan University Press.
Mellinkoff, D.
1963 The language of the law. Boston: Little, Brown and Company.
1982 Legal writing: Sense and nonsense. New York: Charles Scribner's
and Sons.
Meyer, Charles F.
1987 "Apposition in English", Journal of English Linguistics 20/1: 101-
121.
1989 "Restrictive apposition: An indeterminate category", English Studies
70: 147-166.
1991 "A corpus-based study of apposition in English", in: Karin Aijmer
—Bengt Altenberg (eds.), 166-181.
1992 Apposition in contemporary English. Cambridge: Cambridge Univer-
sity Press.
Miliö, L. T.
1990 "A new historical corpus", ICAME Journal 14: 26-39.
Minnis, A. J. (ed.)
1989 Latin and vernacular. Studies in late-medieval texts and manuscripts.
(York Manuscript Conferences: Proceedings Series.) University of
York, Center for Medieval Studies: D. S. Brewer.
1993 Chaucer's Boece and the medieval tradition of Boethius. (Chaucer
Studies 18.) Cambridge: D. S. Brewer.
180 Päivi Pahta and Saara Nevanlinna

Mitchell, Bruce
1985 Old English syntax. 2 vols. Oxford: Clarendon.
Mustanoja, Tauno F.
1960 A Middle English syntax. Part 1. (Mimoires de la Societe N6ophilolo-
gique 23.) Helsinki: Societe Neophilologique.
Nevalainen, Terttu—Helena Raumolin-Brunberg
1993 "Early Modern British English", in: Matti Rissanen et al. (eds.) 1993:
53-73.
Nevanlinna, Saara
1974 "Background and history of the parenthetic as who say / saith in Old
and Middle English literature", Neuphilologische Mitteilungen 75/4:
568-601.
1984 "Continuity and variation in Old English psalm glosses", in: Häkan
Ringbom—Matti Rissanen (eds.), 23-36.
1990 "The occurrence of glosses in three Late Middle English texts: Lexi-
cal variation", in: Jacek Fisiak (ed.), Historical linguistics and philo-
logy. (Trends in Linguistics. Studies and Monographs 46.) Berlin:
Mouton de Gruyter, 273-289.
1993 "AS WHO SAY/SAITH revisited: Form and function", Florilegium 11
(Special I.A.U.P.E. Issue. Carleton University Papers on Late Antiquity
and the Middle Ages, ed. D. Wurtele.) Ottawa: Carleton University,
61-81.
Nevanlinna, Saara—Päivi Pahta
forthc. "Middle English nonrestrictive expository apposition with an explicit
marker", in: Jacek Fisiak (ed.), Studies in Middle English. (Trends in
Linguistics. Studies and Monographs.) Berlin—New York: Mouton
de Gruyter.
Nevanlinna, Saara—Päivi Pahta—Kirsti Peitsara—Irma Taavitsainen
1993 "Middle English", in: Matti Rissanen et al. (eds.) 1993: 33-51.
The New English Bible
1961 London.
Norri, Juhani
1992 Names of sicknesses in English, 1400-1550: An exploration of the
lexical field. (Annales Academiae Scientiarum Fennicae. Disserta-
tiones Humanarum Litterarum 63.) Helsinki: Academia Scientiarum
Fennicae.
Oakden, J. P.
1930-35 Alliterative poetry in Middle English, 1: The dialectal survey; 2 (with
E. R. Innes): Survey of the traditions. Manchester: Manchester Uni-
versity Press.
 Expository apposition 181

Obler, Loraine—Lise Menn (eds.)

1982 Exceptional language and linguistics. (Perspectives is Neurolin-
guistics, Neuropsychology, and Psycholinguistics: A Series of Mono-
graphs and Treatises.) New York—London: Academic Press.
ODEE = The Oxford dictionary of English etymology.
1976 Ed. C. T. Onions (with the assistance of G. W. S. Friedrichsen and
R. W. Burchfield). (Reprinted with corrections.) Oxford: Clarendon
Press.
OED = The Oxford English dictionary. Vols. 1-12 with supplement and
bibliography.
1961 Ed. James A. H. Murray—H. Bradley—W. A. Craigie—C. T. Onions.
Oxford: Clarendon Press.
1989 Second edition.
Ogden, Margaret S.
1971 The Cyrurgie of Guy de Chauliac. (EETS, O.S. 207.)
Onions, C. T.
1971 Modem English syntax. A new edition of An advanced English syntax
completely revised by B. D. H. Miller. London: Routledge and Kegan
Paul.
1986 A Shakespeare glossary. Enlarged and revised by Robert D. Eagleson.
Oxford: Clarendon.
Oostdijk, Nelleke
1994 Review of Meyer 1992, ICAME Journal 18: 83-93.
Pahta, Päivi
1994 Review of Meyer 1992, Neuphilologische Mitteilungen 95: 385-387.
Pahta, Päivi—Saara Nevanlinna
forthc. "Markers of expository apposition in a diachronic perspective".
Peltola, Niilo
1960 "Appositional constructions in Old English prose", Neuphilologische
Mitteilungen 61: 159-203.
Plucknett, Τ. F. Τ.
1940 A concise history of the common law. London: Butterworths.
Prins, A. A.
1952 French influence on English phrasing. Leiden: Universitaire Pers Lei-
den.
Quirk, Randolph—Sidney Greenbaum
1973 A university grammar of English. London: Longman.
Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik
1972 A grammar of contemporary English. London: Longman.
1985 A comprehensive grammar of the English language. London: Long-
man.
182 Päivi Pahta and Saara Nevanlinna

Ringbom, Häkan—Matti Rissanen (eds.)

1984 Proceedings from the Second Nordic Conference for English Studies.
(Publications of the Research Institute of the Abo Akademi Founda-
tion.) Abo: Abo Akademi.
Rissanen, Matti
1993 "The loss of wit 'know': Evidence from the Helsinki Corpus", in:
Risto Hiltunen et al. (eds.), 195-206.
Rissanen, Matti—Ossi Ihalainen—Terttu Nevalainen—Irma Taavitsainen (eds.)
1992 History of Englishes. Berlin: Mouton de Gruyter.
Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.)
1993 Early English in the computer age: Explorations through the Hel-
sinki Corpus. (Topics in English Linguistics 11.) Berlin—New York:
Mouton de Gruyter.
Robinson, Fred C.
1985 Beowulf and the appositive style. Knoxville, Tennessee: University of
Tennessee Press.
Ryd6n, Mats
1966 Relative constructions in early sixteenth century English with special
reference to Sir Thomas Elyot. (Universitatis Upsaliensis Studia Ang-
listica Upsaliensia 3.) Uppsala: Almqvist and Wiksell.
Samuels, M. L.
1970 Linguistic evolution. With special reference to English. (Cambridge
studies in linguistics 5.) Cambridge: Cambridge University Press.
Sopher, H.
1971 "Apposition", English Studies 52: 401-412.
Strang, Barbara Μ. H.
1970 A history of English. London—Colchester: Methuen and Co.
Svartvik, Jan—Randolph Quirk (eds.)
1980 A corpus of English conversation. (Lund Studies in English 82.)
Lund: Lund University Press.
Swales, John M.
1990 Genre analysis. English in academic and research settings. (Cam-
bridge applied linguistics series.) Cambridge: Cambridge University
Press.
The Book of Common Prayer. See above.
Thomas Thomas's Dictionarium linguae latinae et anglicanae 1587.
1945 Ed. R. C. Alston. (English Linguistics 1500-1800.) Menston: Scolar
Press.
Visser, F. Th.
1972 An historical syntax of the English language. Part 2. (Second impres-
sion.) Leiden: Brill.
 Expository apposition 183

Voigts, Linda Ε.—Michael R. McVaugh (eds.)

1984 A Latin technical phlebotomy and its Middle English translation.
(Transactions of the American Philosophical Society 74, Part 2.)
Philadelphia.
Vulgata = Biblia sacra iuxta Vulgatam Clementinam
1950 Ed. R. P. Alberto Colunga—Laurentio Turrado.
Wallner, Björn
1987 "On the ./. periphrasis in the N.Y. Chauliac", Neuphilologische
Mitteilungen 88/3: 286-294.
Webster, Jonathan
1988 "The language of religion: A socio-linguistic perspective", in: M.
Ghadessy (ed.), 85-107.
Wrenn, C. L. (ed.)
1973 Beowulf. With the Finnesburg fragment. (Third edition. Fully revised
by W. F. Bolton.) London: Harrap.
Genre conventions: Personal affect in fiction
and non-fiction in Early Modern English

Irma Taavitsainen

1. Aim and plan of the study

This study discusses Early Modern English genres and text types and pro-
poses a new way of looking at them. The focus is on differences between
fiction and the adjoining text types. The starting point of the assessment
is the level of individual texts; genres defined by external factors form the
first level of abstraction and text types the second.
Variation in features marking subjectivity provides a possible key to a
description of text types. In this study a broad range of such features is
used to assess whether there are any distinctive traits in early prose fiction
that tell this mode apart from other narrative but non-literary text types;
and whether and how statistical analysis of linguistic features can cast new
light on our understanding of early fiction and its evolution.
The aim of this study is to chart the difference between adjoining text
types in terms of conventionalised discursive properties, i.e. to pinpoint
text-type markers. Personal affect may provide a tool for a more detailed
definition of these properties because it shows a great deal of variation in
both quality and volume. The co-occurrence patterns of features marking
subjectivity may also make a more detailed description of the qualities of
personal affect possible.
The plan of my study comprises several phases. First, I give the theo-
retical frame of the study. This theory is based on several sources and my
previous empirical studies (Taavitsainen 1993, 1994, 1995a, 1995b, and
forthcoming a and b). The inspiration for this study came from structur-
alist literary criticism, and I developed my research task so that it is suit-
able for corpus-based statistical analysis.
186 Irma Taavitsainen

The hypothesis is that the co-occurrence patterns of features marking

subjectivity provide a means to a more detailed description of genres and
their affinities. I have used factor analysis to identify adjoining text types
and matrices of linguistic features that may be of importance in defining
the quality of personal affect. The genres under scrutiny in this phase are
plays, fiction, biography, autobiography, diary, travelogue, history, hand-
books, science, education, and philosophy in the Early Modern English
section of the Helsinki Corpus (HC hereafter). The purpose is to find out
relevant groupings of adjoining and perhaps overlapping genres, and lin-
guistic features that need to be assessed in more detail, for the ultimate
aim of defining genre conventions and their changes. This is hardly pos-
sible on the basis of factor analysis only, and therefore this method has
been complemented by other quantitative and qualitative methods. Factor
analysis arranges the measured features into hierarchical scales according
to their importance in reflecting the underlying dimensions of variation,
and thus it helps to select features for further assessment.
The third step is to select the genres with closest affinities to fiction for
closer scrutiny and make a microanalysis of linguistic features that are po-
tential text-type markers. This study concentrates on fiction and its nearest
genres. The material of this phase consists of 6 texts of early fiction, and
24 other texts: travelogues, diaries, biographies, autobiographies, and his-
tory writing. This grouping is suggested by factor analysis. In this assess-
ment, the factor analysis is complemented with other statistical methods
for the purpose of pinpointing some conventionalised discursive proper-
ties. T-tests and variance analysis are employed to verify whether the lin-
guistic features, indicated by factor analysis as important in describing the
underlying dimensions of personal affect, are significant in telling the ad-
joining text types apart. The additional methods are necessary for testing
whether the features with highest factor loadings really belong to the class
of conventionalised discursive properties that I set out to establish in this
study.
The final phase of the study consists of qualitative analysis of passages
that show high frequencies of features studied by statistical methods. Four
different aspects prove important: "surge" features of personal affect, es-
pecially interjections; interactive focus; immediate context; and space-
building modality. At the end of the study, conclusions are drawn and
suggestions made for future studies.
 Genre conventions: Personal affect 187

2. Genre theory

2.1. Genres and text types

The HC uses genres as structural elements. When assigning texts to gen-

res, we charted the fields of writing in each period, and classified texts
according to external evidence of provenance and function. Labels like
"travelogue", "religious treatises", "sermons", and "biography" are based
on this external evidence. They provide a structural element in the corpus,
and we chose to call them text types (1986; for the principles of text se-
lection, see Rissanen et al. 1993). In this paper, however, I follow the
usage of Biber (1988) according to which the term genre applies to clas-
sification according to external evidence, and text types are groupings
made according to internal linguistic features of texts. Thus genres and
text types are two kinds of abstraction used in text taxonomies, but based
on different criteria. External evidence, such as textual origin, purpose and
audience, provide one set of criteria, and the term genre is used of such
groupings; internal linguistic features provide another way of classifica-
tion, leading to text types.1 For example, moon prognostications can be
classified according to their external starting point, i.e. the moon by itself
or in combination with other celestial bodies, their purpose and their pro-
venance in handbooks for doctors, almanacs and household literature (Taa-
vitsainen 1988). When the internal evidence is examined, they can be clas-
sified as instructional texts according to their use of imperative forms,
second person pronouns and prescriptive features of language. In contrast,
parts of them, for example more comprehensive biblical allusions, are
classed as narratives as they contain past tense verbs, a chronological se-
quence of events and third person pronouns.
The distinction beween the two types of abstraction is useful and neces-
sary if we wish to chart the linguistic features that serve as a matrix in
distinguishing various text types and pinpoint differences between adjoin-
ing groupings (see below). Both genres and text types are abstractions
made on the basis of individual texts. Thus the level of the text itself is
the only reliable starting point.
Genre criticism of Renaissance texts has focused on literary writings.
The present study takes a somewhat different approach. Its incentive came
from Tzvetan Todorov's claim that every literary genre has non-literary
counterparts that stand near it (1990: 11), and the argument that distinc-
tions between literary and non-literary styles are at best arbitrary (Carter—
188 Irma Taavitsainen

Nash 1990: 3). If this is true, it must be possible to verify it in more de-
tail. The idea of testing it fitted into my current thinking about genres and
text types, and textual conventions within them (Taavitsainen forthcoming
c).
It seems that text types may be very different from one another in their
stylistic markedness, so that some linguistic features may be more salient
than others at any one time. These features may change in the course of
time though the same label may persist, for example, politeness formulae
are more prominent in Renaissance texts than earlier, instructions in hand-
books may be given directly in imperative forms or by descriptive pas-
sages, or in dialogue form. The recognition of such salient features could
be helpful in defining genre styles of past periods. It is evident that no
feature alone is sufficient for describing genre styles, but a matrix is need-
ed. Yet the features within the matrix can be ordered, and a hierarchical
scale may be helpful. Furthermore, the dynamics of textual features along
a time scale may be revealed through the changing order of importance
of matrix features occurring in a diachronic sequence of sychronic descrip-
tions.

2.2. Genre conventions

The social practice of texts and genres has received increasing attention.
In a given society, the recurrence of some linguistic features may be insti-
tutionalised, and individual texts in that genre are produced and perceived
in relation to that norm. Late medieval and renaissance dedications pro-
vide a good example, and letters with their well-defined politeness strat-
egies are another case in point. Instructions for epistolary compositions
were in great demand on the continent and in England, and guidebooks on
how to write letters in the correct way were reproduced in several editions
(Guillen 1986: 72-74). 2 Besides such regulation from "above", the influ-
ence could come more unconsciously from "below", as the conventions
may have been created under the influence of social prestige or in imita-
tion of older masters. Chaucer must have served as a model for later writ-
ers in the art of storytelling. The broadening of the audience, on the one
hand, and of the topics, on the other hand, may have had an influence, so
that features of colloquial speech may have found their way more readily
into the language of comedies and fiction (see Taavitsainen forthcoming
b). Furthermore, the formal difference, whether verse or prose, must be
related to the changing functions of the media (see 2.4).
 Genre conventions: Personal affect 189

Both literary and non-literary text types are choices among the discur-
sive possibilities available in the language system. Both genres and text
types guide text production and regulate the reception of texts; the social
situation determines the choices, within the frame created by already exist-
ing writings pertaining to the same situation. Both authors and readers are
thus influenced by the existing system, and changes are brought about by
bringing in additional linguistic features and leaving out old ones, or by
broadening the target groups, functions, etc. (cf. Fowler 1982: 11). A text
type, whether literary or not, is a codification of discursive properties, i.e.
linguistic features, but its individual members contain these features in
various degrees and in various combinations. Text types may be very
heterogeneous in respect to the occurrence of linguistic features; there is
variability even within narrowly defined groups of texts (Meurman-Solin
1993a, Taavitsainen 1993), though some text types may be very limited
and have strict rules. 3
It is because text types exist as conventions that they function as guide-
lines for both readers and writers. Conventionalised opening phrases like
"once upon a time" serve as signals to the audience and tune their ex-
pectations. At the same time the same conventionalised phrases act as
models for authors, as they provide the norms. Individual linguistic fea-
tures, and their combinations, may work in the same way and provide
models for imitation. This idea is especially relevant when dealing with
periods in which imitation was a virtue. In the Late Middle Ages and the
Renaissance the goal of a poet was to bring the accepted form of versifi-
cation, the sonnet for instance, to perfection; the aim was to recreate the
old, not to renew it. Writing belonged to the public world, and it was far
removed from the private concerns of modern authors. For example, lit-
erature intended for the court circles strove to achieve the accepted norms.
This principle must have worked on a broader basis as well, and influ-
enced other writers of the period. The time span of the HC covers the
transition from medieval conceptions of literary production as imitation to
the modern outlook based on individual creativity. This change should
also show in the selection of linguistic features.
Any verbal property, optional at the level of language, may be obliga-
tory in discourse; the choice a society makes among all the possible codi-
fications determines its way of writing both in relation to genres and text
types. Personal letters are different from official correspondence, scientific
discourse has its own rules, and so on. These conventions may vary great-
ly in a historical perspective.
190 Irma Taavitsainen

The institutionalised discursive properties of various text types may stem

from any level of language, from semantic, syntactic or lexical sources,
and the difference between one text type and another may be situated on
any of these levels (Todorov 1990: 18). Other names for such discursive
properties are distinctive features or text-type markers. The above view
may be criticised for being reductive, but it gives an explicit formulation
to the theory and indicates a case for empirical studies.
The problem can be approached, for example, with a statistical study of
linguistic features in various text types. It is evident that no single feature
is sufficient to describe a text type, or even the style of a single writer, but
a matrix is needed. The co-occurrence of linguistic features working in the
same direction is important. The co-occurrence patterns of a great number
of linguistic features make up several linguistic matrices that form a multi-
dimensional scale of linguistic variation. Yet if the aim is different the
approach can be modified. If the aim is to pinpoint a difference between
adjoining genres, as in this study, the notion of a distinctive feature be-
comes important, as according to the present assumption the difference
between one text type and another may be found in such features, though
it is their co-occurrence that determines the style. An overall assessment
with a matrix of features may suggest stylistic markers of particular period
and genre styles. Linguistic features to be studied must be selected so that
they form a coherent whole but still show enough variation to form differ-
ent co-occurrence patterns.

2.3. Fiction and non-fiction

The difference between fiction and non-fiction is one of the key questions
of aesthetics and literary theory, starting from Aristotle. The difference has
been defined in various ways, e.g. in terms of referentiality, truth values,
modes of knowledge, speech act theory, audience response, and purpose
of the text. For example, the question of mimesis can be broadened to in-
clude non-fictional writing such as prose dialogue in which characters are
created for instructional purposes in works like Walton's The Compleat
Angler (Imbrie 1986: 66). Other prose genres with close affinities to fic-
tion include biography, autobiography and travelogue, in which fictive and
historical elements can blend and pose the mimetic questions of "design"
and "truth" (Hart 1974: 221).
 Genre conventions: Personal affect 191

Early fiction and especially the early novels are well-known for their
emotive qualities, with frequent outbursts of personal feelings. Literary
style became highly conventionalised in this respect; an example is
Euphuistic style, which was "an artistic embodiment of a world view" and
had a courtly appeal with its elaborate patterns and analogies (Margolies
1985: 46). John Lyly wrote for the high social circles and was widely imi-
tated. Yet there were other kinds of fictional writing, such as pastorals, ro-
mances, and literature with a broader social basis, and other patterns must
have influenced fictional writing as well. In particular, short fiction for a
wider and more general audience with different tastes must have had other
models.
It has been said that the purpose of early fiction was to evoke awe and
wonder, the defining characteristic of romance being the immediate, pow-
erful impact on the reader of the separate episodes, each of which calls for
a full response and are both primitive and sophisticated at the same time.
Characters are said to manifest states of uncontrolled passion that burst
into either extravagant lamentation or violent action. It has also been said
that emotion in early fiction may be mentioned but it is never particular-
ised, and therefore remains, like "noble chivalry", a value so generalised
that it means practically nothing and does not provide an experience for
the reader (Margolies 1985: 36).4
The subtle and refined intricacy of early works like Arcadia or Euphues
was created for a culturally learned audience that appreciated literary con-
ventions beyond story-telling. In the course of time the readersip broaden-
ed, and there was a circular line of development. When the literature
originally intended for a courtly audience spread to the bourgeoisie, the
aristocracy rejected what the middle classes imitated; when Euphues gain-
ed popularity with a common audience, it became an object of mockery
among the aristocracy. The same happened to romance (Margolies 1985:
14-16). This circular movement can be found even earlier in literary his-
tory. Chaucer's original audience consisted of the court circles and the
upper middle classes that had dealings with the court, such as rich mer-
chants. By the peak popularity of Euphues and Arcadia, the same stories
circulated among the lowest literate classes in jest books like Me/77 Tales.
192 Irma Taavitsainen

2.4. Prose and verse

The period in focus poses special problems as the relationship between

prose and verse was in a state of flux. In the 15th and 16th centuries, fic-
tion was still largely written in verse, Sir Philip Sidney's Arcadia being
a good example, but prose started to be used for this purpose increasingly.
Prose fiction was firmly associated with an instructive function and moral-
ising purposes were prominent. The conventions of prose fiction were in
the process of formulation, and verse fiction of the previous period should
not be ignored when charting the origin of these conventions. Some fea-
tures of medieval verse fiction may have been adopted into the prose fic-
tion of the 16th century, and its generic markers may have persisted in
early prose fiction.

2.5. From manuscript to print

Another important change that was simultaneously going on had to do

with the transmission of literature through print. It is well known that ear-
ly printers imitated manuscript lay-out and illumination. Reading was still
a social activity rather than a personal exercise in the Early Modern pe-
riod. Manuscripts were produced for a certain, well-defined readership, for
local circulation, and they provided the model even when printing had
become more common. The conventions of reader involvement were carri-
ed over so that early works often had the personal, immediate quality of
direct address. Authors wrote for an assumed readership long after the be-
ginning of printing. This is especially noticeable in translations, which
often had an extensive commentary for the readers' benefit. In fiction the
authors and readers were thought to share a literary experience, and the
change to a more diffused presence of the author or commentator was
slow. The change from manuscript circulation to print involved an objecti-
fication of the text, which happened only gradually. The first works clear-
ly perceived as printed, in contrast to the more private circulation by
manuscripts, are from the very end of the 16th century (Margolies 1985:
22-28).
 Genre conventions: Personal affect 193

3. Personal affect

Personal affect means the expression of subjective emotions, feelings,

moods and attitudes. It has been defined in various ways in the literature,
but no generally accepted set of linguistic features expressing personal
affect in Present-day English has been defined. Cultural environment has
a powerful effect upon emotional expression, as different display rules are
obeyed (Frijda 1986: 62). Cultural conventions are crucial to both the
amount and kind of personal affect; this applies to gestures as well as lin-
guistic behaviour, and both to spoken and written language.
According to my earlier studies, the degree of involvement in the form
of personal affect varies greatly in intensity and means of realisation from
text to text. It provides a fruitful object of study, and a way of describing
genre conventions. For example, the volume of personal affect varies
greatly in Middle English religious treatises so that the differences are
great even among mystical authors (Taavitsainen 1993 and 1994b). It is
an essential component of participant relations and finds outlets in various
forms. It may be present explicitly as direct guidance and comments to the
reader, as for example in religious texts. More indirect ways of involve-
ment include the creation of a mock reality in which the readers are sup-
posed to live with the depicted characters, feel their emotions and imagine
themselves in their situations. For example in fiction, the pretence of re-
corded speech creates a sense of immediate presence and widens the scope
of interaction with the audience; in religious texts readers are supposed to
join in the prayers. Personal attitudes are also expressed by the point of
view in texts, subjectivity being the prevailing mood in texts with high
"volume" of personal affect. 5 In general, emotive focalisation implies a
noticeably idiosyncratic way of representation (cf. Toolan 1988: 73).

3.1. Towards a definition of the quality of personal affect

In earlier studies, affect has been classified into positive and negative fea-
tures. Yet it is evident that other, more subtle and perhaps more impor-
tant shades can be distinguished. Participant relation is important in its
various forms: feelings about oneself often pair off with attitudes to some-
one else. Emotions can be reflected on someone else, and a personal point
of view can open one's mind and way of seeing things to others. Besides
negative and positive expressions of personal affect, two other dimensions
194 Irma Taavitsainen

have been proposed for further study. They are "surge/predisposition" and
"self/other" (Martin 1992: 533). In the present study I understand "surge"
as an outburst of emotion, an expression that conveys intensified personal
charge between the participants of communication, whether text-internal
or exophoric in reference. The dimension "self-other" includes various as-
pects of involvement and interactivity versus egocentric focus.
Negative and positive expressions of personal affect were assessed in
Biber's and Finegan's study (1989b) with adjectives and adverbs, but their
results were not particularly good: in a cluster analysis only 15 out of
more than 400 texts showed relatively high frequencies of personal affect
features. It may well be that the features in their study did not grasp the
essence, and better results may be achieved by different means, along oth-
er dimensions.
My earlier studies showed that different genres have their own codes,
and that these can only be grasped by close reading and detailed textual
analysis (Taavitsainen 1993). The present study concentrates on fiction
and non-fiction. The measurement of features reflecting involvement and
emotional charge could perhaps illuminate the problem of the relationship
between early fiction and other narrative text types of the same period. In
addition, a diachronic approach seems to be helpful in approaching the
question of conventions within a changing system of genres, and a longer
time span may also reveal distinctive features and give evidence for de-
scribing the evolution of text types. According to my theory, personal af-
fect may be realised in different ways in the participant relation, and such
features may well serve as formatting factors in text production and re-
ception.

3.2. Point of view

In my definition the subjective point of view provides a key to the assess-

ment of personal affect. Our conceptualisation of reality begins with the
individual, the mental self, and "I" serves as the reference point (Ben-
veniste 1966 [1971]; Mühlhäusler—Harre, 1990; Heine et al. 1991: 12;
Fludernik 1993: 46, 431^132; Adamson 1994: 195). It has been claimed
that every interpretation of history is based on the present moment, just
as space starts with "here", and the notion of other people with "I". Four
different spaces have been discerned in this egocentric illusion of "I-here-
now-in this case"; they form a scale which orders social, spatial, temporal,
and logical spaces in an inward/outward progression. Social space with "I"
 Genre conventions: Personal affect 195

stands closest to the core, and the first person singular is the basic unit in
this reference system. Deixis may therefore become important in this di-
mension. Five major types of deixis have been identified: social and dis-
coursal, as well as those of person, place and time. Egocentricity is crucial
as the speaker in any language act situates referents both temporally and
spatially to him- or herself, speaking "here" and "now" (Toolan 1988:
127).
Reader involvement is another important issue connected with point of
view. Involvement may be created in various ways, and one challenge is
to find out both the evident and subtle mechanisms used for this purpose
and the various genre conventions. For example, advice on how to inter-
pret the text may be given, or the author may assume that the reader
shares his views and feelings, which is a more indirect way of influencing
the audience.

4. Material

The focus of this study is on early fiction. The HC includes passages of

fictional texts targeted at various layers of society. The following groups
can be discerned:

1) popular fiction; (a) and (c)

2) sociological literature; (d)
3) middle-class fiction; (b) and (e)
4) romantic fiction; (f).

The popular trend is well represented: (a) Merry Tales contains comic
turns and moralising authorial comments. It is the earliest English jest
book, (b) Thomas Deloney represents middle class fiction. The last of the
Elizabethan realists, he started with ballads but turned later to write narra-
tive prose. His work continues and develops the tradition of jest books and
anecdotes; about one third consists of separate jests, but he uses them for
social and moral commentary, thus developing them into a whole. His
book is a forerunner of later novels in this sense (Davis 1969: 239-244).
(c) Penny Merryments belongs to the layer of popular literature written for
the broadest possible audience. It contains stories like "Honest John" and
"Loving Kate" describing the marriage plans of two in-servants for an
196 Irma Taavitsainen

audience of the same level of society (Spufford 1981: 60). (d) Thomas
Harman's book is meant as an "alarum" to forewarn honest citizens, but
in fact it belongs to sociological literature, (e) Robert Armin was a chief
actor of comic parts in the early 17th century. He wrote one play with
genuine dramatic power and some prose works, (f) The final text, very
different from the others, is by the first professional woman writer, Aphra
Behn, a dramatist and novelist. Her most famous novel Oroonoko owes
much to her childhood memories of plantation life in Surinam (Sampson
1970: 929).

4.1. Parameters of the HC: interactive textual form

The parameter value "Interactive" has been assigned to plays and texts in
the dialogue form in the HC, but texts with the value "X" also contain
interactive passages. Thus a large part of the present material consists of
dialogue without time constraints or paralinguistic gestures and intonation
to aid the interpretation, but imitating and perhaps condensing some fea-
tures of spoken discourse. Thus the different levels of spoken language
have to be taken into account, as well as the proportion of direct and in-
direct narration.6 The word counts of direct speech versus narration in fic-
tion show a great deal of variation:

Merry Tales: narration 5093 words; direct speech 1333 words;

total 6426 words
Harman: narration 3107; direct speech 2012; total 5119
Armin: narration 3914; direct speech 1257; total 5171
Deloney: narration 3002; direct speech 4321; total 7323
Penny Merriments: narration 1690; direct speech 4882; total 6572
Behn: narration 5120; direct speech 357; total 5477.

The time constraints of spoken language may cause reduced features of

syntax such as THAT-deletion, and the more involved quality in the use
of first and second person pronouns. Immediate context and indexical use
of language may be a characteristic of spoken language as the context is
usually clear; in written language the context has to be created by differ-
ent means. In general, spoken language is less edited and integrated, but
rapid adjustments to subtle nuances of interpersonal meaning are possible,
and spoken language has its own kind of complexity (Halliday 1989: 78-
84). Thus personal affect features are likely to be found in passages that
 Genre conventions: Personal affect 197

imitate speech, and its linguistic features are likely to be among those of
spoken language. Interaction proved extremely important for the present
assessment, but the parameter itself was not very useful (see note 11).

4.2. Parameters of the HC: informal language

Variation in language may be described in several ways, and there are

subtle shades of difference between various labels. Formal versus informal
is a pair of such labels connected with situation, and employed as a para-
meter in the HC. Non-standard language may be found at the informal end
and such language may be socially or regionally marked. Imitations of
vulgar or provincial speech are well known from earlier periods. Non-
standard speech may be emotionally loaded, swearing being a good ex-
ample.
In fiction, dialogues are important and they are likely to contain delib-
erate informal features. Early short fiction in the present material consists
of jest books and meriy tales for the broadest possible audience. If the
classical rules of adjusting the style of writing according to audience edu-
cation are at all valid in this connection, the distinctive features should be
expected at the more informal, spoken end of the scale. On the other hand,
the moralising and edifying function may also be reflected at the level of
linguistic features, and indeed the edifying purpose may be dominant.
Conventionalised politeness strategies are directly relevant to the present
attempt to trace personal affect features in various text types. For the
influence of this parameter, see note 11.

4.3. Parameters of the HC: imaginative versus non-imaginative

In the Early Modern section of the HC, fiction is assigned the parameter
value "NI" ("Narration Imaginative"), while other texts have the value
"NN" ("Narration Non-imaginative"). As the aim of this study is to search
out possible distictive features of early fiction, this parameter was selected
for further scrutiny (see section 7).
198 Irma Taavitsainen

5. Linguistic features

Linguistic stylistics uses a broad range of linguistic features with various

functional properties (see e.g. Leech—Short 1981). For my aim, Biber's
list of involvement features (1988: 89) is helpful, but not applicable as
such to the present purpose or to a historical corpus because of, for ex-
ample, linguistic and semantic changes. As pointed out before, no set of
linguistic features describing the quality or qualities of personal affect has
been defined in Present-day English, and the problem is even greater with
the past stages of language and in past sociocultural environments.
The choice of linguistic features was based on my previous studies and
my reading of these texts, and background knowledge of early fiction.
Some of the features, like personal pronouns, were an obvious choice, but
I tried to be more inventive and include features that might prove impor-
tant and illuminate the problem of defining personal affect (see above
3.1).
In order to achieve a more detailed description of this aspect of partici-
pant relations in texts, I selected a fairly broad range of linguistic features
that might be thought to be relevant to personal affect. Personal focus and
deixis are potentially important. Yet the range of features should be broad-
er as co-occurrence patterns can be assumed to provide a key to defining
the quality of personal affect in various text types. Therefore I selected a
number of other features which had been established as relevant along the
dimensions spoken-written / informal-formal / spontaneous-more process-
ed in earlier studies. Their affinity to personal affect could be more in-
direct and grasped via matrices (see 2.1).
A brief list of the selected features is given, with the working assump-
tions based on earlier studies. The features and the aspects that become
important in the assessment will be discussed in detail later.

5.1. Personal pronouns (subject forms): I, WE; YE/YOU; HE,

SHE, IT, THEY

All subject forms of personal pronouns were assessed; other forms were
not included in this phase of the study, but they are included in the dis-
cussion later (see section 9). In earlier studies, personal pronouns proved
to be powerful in determining the kind of involvement in texts, and the
interpretation is straightforward: first person pronouns express ego-in-
 Genre conventions: Personal affect 199

volvement, second person pronouns involvement with the addressee, and

third person pronouns involvement with the subject-matter; the alternative
is to use nouns instead (Chafe 1985). In general terms, first and second
person pronouns mark the presence of a narrator and an addressee, while
third person pronouns mark relatively inexact reference to persons outside
the immediate context of interaction. According to Biber's study (1986),
third person pronouns co-occur frequently with past tense and perfect as-
pect forms, as a marker of narrative, or reported (versus immediate) styles.
A frequent use of the pronoun it has been interpreted as a sign of a rela-
tively inexplicit lexical content due to strict time constraints and non-
informational focus, and thus it has been labelled as a feature of spoken
language.

5.2. Direct questions: WH-, HOW

Questions, like second person pronouns, are typically interactive and in-
dicate a concern with interpersonal functions and involvement with the
addressee (Biber 1986). Yes/no questions were excluded from the present
analysis because they could not be accurately identified by automatic
analysis.

5.3. Proximal versus distal context

Demonstrative pronouns THIS/THAT, THESE/THOSE can refer to an entity

outside the text, an exophoric referent, or to a previous referent in the text
itself, often to an abstract concept (e.g "this shows ..."). Their use without
a nominal referent has been interpreted as a feature of spoken language
due to faster production and lack of editing (Chafe 1985). Adverbs of pro-
ximity NOW and HERE are also relevant to the immediate context and
situationally bound discourse, and thus to egocentric focus.

5.4. Reductive features

Reductive features have been considered typical of the spoken language.

Subordinator THAT-deletion is a form of syntactic reduction relevant to
the present concerns; in Biber's study it had the second highest factor
loading in the involvement factor (0.91); contractions were also important
200 Irma Taavitsainen

in his study (factor loading 0.90), but in the present study they are not as
relevant as the only contractions are verbal forms like V, '// of modals
and 'c/ of past tense forms. They become frequent only in the last period
of the HC.

5.5. Exclamations

Exclamations have been described as belonging to the purely emotive

level of language (Quirk et al. 1985), but in addition to expressions of
emotions, other functions can be distinguished. According to my previous
study (1993) they were most important in expressing personal affect in
Middle English religious texts. The following lexical forms were found in
the Early Modern part of the HC: Ο! OH! A! AH! LO! LOO! LOE! LA!
ALAS! ALLAS! FY! FIE! FYE! EH! BENEDICITE! GRAMERCY! AMEN!
WHY! WHAT! HOW! TUT! TUSH! PSHA! They occur in a variety of func-
tions that range from creating an illusion of spoken language to genre-
specific conventions of story telling in the written form (Taavitsainen
1995a, 1995b, and forthcoming d; see also Blake 1992).

5.6. Swearing words

Swearing words overlap with exclamations to some extent, but according

to my pilot study they were even more restricted and occurred in few text
types (Taavitsainen forthcoming b). Therefore it seemed fruitful to assess
them separately. The following swear words and phrases were found:
MARY, BY SAINT ..., FOR SHAME, FOR ... SAKE, BY MY ..., DEVIL,
GOGS..., COCKES BONES!

5.7. Pragmatic particles

Pragmatic particles also overlap with exclamations to some extent, but

mostly their functions are different (see Jucker forthcoming). The lexical
items included in the present assessment are: WELL, (I) PRAY (THEE/
YOU), PRITHEE, I SAY, YEA, NAY.
 Genre conventions: Personal affect 201

5.8. Private verbs

Private verbs had the highest factor loading in Biber's study (0.96). In his
study they are treated as one class, without further subcategories. I de-
cided to apply a more detailed semantic classification that could be helpful
in defining the various kinds of personal affect, or personal point of view
in texts. Certain restricted classes of verbs can be identified as having
specific functions, such as verbs of cognition that refer to mental activities
and sensory verbs such as SEE, HEAR, and FEEL. All occurrences were
counted at first, but in the qualitative assessment the difference between
the first, second and third person are discussed. In my study I applied the
following semantic subcategories, based on my own judgement:

1. Emotion: LOVE, HATE, HOPE, DESIRE, COVET, LONG, ENVY,

DREAD, FEAR
2. Sensory perception: SEE, HEAR, TASTE, SMELL, FEEL
3. Observation: OBSERVE, PERCEIVE, DISCERNE, NOTE, DISCOVER
4. Mental activity: ESTEEM, DEEM, ESTIMATE, JUDGE, FIND, THINK,
CONSIDER, GUESS, DOUBT, REASON, CONCLUDE, SUSPECT, SUP-
POSE, PRESUME, ASSUME, ACCEPT, EXPECT, FORGET, REMEM-
BER, RECALL, BELIEVE
5. Mental state: KNOW (WIT), UNDERSTAND, HOLD (AN OPINION)
6. Imagination: IMAGINE, FANCY, DREAM.

5.9. Modals

Modality in language can be understood broadly as encoding speaker at-

titudes and point of view. It is expressed in a boad range of features,
among which modal auxiliaries are important and easily detectable in a
computerised corpus. No definitive set of linguistic features encoding
modality in language has been defined, but it can be assumed that such
features are closely connected with personal affect, and thus they may oc-
cur in the matrices sought out in this study.
It is possible to distinguish three functional classes of modals: 1) those
marking permission, possibility, or ability; 2) those marking obligation or
necessity; and 3) those marking volition or prediction (Quirk et al. 1985).
Chafe (1985) includes possibility modals among the evidentials that mark
202 Irma Taavitsainen

reliability, and necessity modals among those evidentials that mark some
aspect of the reasoning process. Politeness strategies may also be relevant
in the assessment of the use of modals. The following groupings were
taken as the basis of assessment; the concept of boulomaic modality was
introduced later, as it proved significant.

Possibility: CAN, COULD; MAY, MIGHT

Necessity: SHALL, SHOULD; MUST; OUGHT
Volition: WILL, WOULD ('L(L), 'DE)

6. Factor analysis

6.1. Method

Factor analysis has proved to be a fruitful way of assessing textual affini-

ties, and the method has been applied for various purposes (Biber 1988;
Meurman-Solin 1993a; Taavitsainen 1993; Kytö forthcoming. For an al-
ternative method, cluster analysis, see Biber 1989). As the aims of the
present study are different from those of earlier studies using factor analy-
sis, it may be helpful to specify the difference in more detail. In Biber's
study the purpose is to give a comprehensive overview of variation across
speech and writing, and of the complexity of text types (Biber 1988 and
Biber—Finegan 1986). Meurman-Solin's study tests Biber's method with
a historical corpus of early Scots, and she concentrates on Biber's Factor
1, interpreted as the involvement factor. In another study she elaborated
her results by relating them to the audience parameter of her corpus of
early Scots (1993b). Kytö's study is another application of Biber's in-
volvement features, but this time to the subcorpus of Early American Eng-
lish.
My own pilot study (1993) was designed to test whether various types
of religious texts could be distinguished from one another on the basis of
personal affect features. The linguistic features were selected on the basis
of qualitative studies of Rolle's affective style, and my own reading of
Middle English mystics. The features had to be detectable in an untagged
corpus with a great deal of dialectal variation, and therefore lexical
searches were the most appropriate. Exclamations which had not been as-
 Genre conventions: Personal affect 203

sessed before proved to be most powerful in distinguishing genre/subgenre

styles. The pilot study thus indicated a clear path to follow in a further
study. It is obvious that stylistic features like parallelism, alliteration and
repetition are not readily captured by statistical methods, but statistical ap-
proaches have to be complemented by qualitative readings.
The present study is a further development along the same lines of
thought. I concentrate on the volume and quality of personal affect, and
whether features of personal affect are conventionalised to the degree that
they can be used in distinguishing genre/subgenre styles. The aim is to
proceed further with the assessment of co-occurring linguistic features in
order to test with other methods whether any of them serve as distinctive
discursive properties that tell the adjoining genres apart. The factor analy-
sis is complemented by other statistical methods for this aim. Another aim
is to define the quality of personal affect, and qualitative assessment is
used for this purpose.
Factor analysis seems to be very sensitive to the number of linguistic
features under scrutiny: the broader their scope, the more general the de-
scription of a single factor. In Biber's study the number of features was
67 and factor analysis grouped them into 7 underlying dimensions. 7 Fac-
tor 1 was the most powerful in explaining underlying variation, and it was
labelled "involved versus informational production". In Meurman-Solin's
study the number of features was 19, all from Biber's factor one. Her re-
sults yielded a more detailed description of involvement. She suggested
a division into author-centred texts, addressee-oriented texts, and a third
kind of involvement focused on evidentiality. In my pilot study the num-
ber of features was five and they were deliberately selected so that they
all contributed to the same direction and fell on one factor. This was de-
sirable as the emphasis of the study was on factor scores of texts and the
internal homogeneity versus heterogeneity of genres in respect to personal
affect. It proved that the volume of personal affect within texts grouped
according to external evidence varied a great deal, and confirmed my as-
sumption that the quality of personal affect provides a fruitful area for
more detailed research.
In the present study the number of features has been greatly increased
as the aim is to assess the quality of personal affect as well as the amount
of variation along the various dimensions within the larger frame. Many
of these features have not been included in earlier studies and thus there
was no previous knowledge of how they would correlate with other fea-
tures. Personal pronouns are central to the present approach, and the as-
204 Irma Taavitsainen

sumption is that the three kinds of involvement suggested by Chafe and

confirmed by Meurman-Solin's factor analysis would be found in some
form, and the additional features would help to define their nature in more
detail. According to my pilot study (1993) exclamations were the most
powerful feature in distinguishing Middle English genre styles, and it is
evident that the use of interjections in a historical perspective is worth
studying in more detail. The results confirm this assumption. Yet the qual-
ity of personal affect proved to be extremely complicated, and results
were achieved thanks to the combination of the various methods; factor
analysis alone was not sufficient for this purpose.

6.2. Factorial structure

The linguistic features were gathered by WordCruncher and an explora-

tory factor analysis was performed.
To start with, all features are treated equally. Factors group the lin-
guistic features that co-occur with high frequency. The features loading on
the same factor are interpreted to share a communicative purpose. The
method also arranges the linguistic features into a hierarchical order ac-
cording to their weights on the factor, i.e. their correlation with the factor
denotes their importance in representing an underlying dimension.
Consequently, factors represent co-occurrence patterns of linguistic fea-
tures which are arranged in a hierarchic order according to their weight
on the factor. Co-occurring features may be interpreted to be functionally
related, constituting an underlying dimension of variation.
The best number of factors has to be established by experimentation in
order to find the solution with the most reasonable linguistic explanation.
If too few factors are selected, the result is a loss of information, while
too many factors lead astray in the interpretation of underlying dimen-
sions. The features may be combined (e.g. singular and plural, / and we),
some may be excluded, and others added. 8 It has also been emphasised
that interpretations made on the basis of factor analysis should be verified
by additional considerations (see section 9).
 Genre conventions: Personal affect 205

Eigenvalues

1.5 -

1-

0.5 -

0 1 1 1
Fl F2 F3

Figure 1. Eigenvalues: scree plot.

206 Irma Taavitsainen

In the present study, a three-factor solution was considered best in de-

scribing the qualities of personal affect and in indicating relevant group-
ings of texts. It gives enough information, and the communalities of fea-
tures reveal that no potential factors remain hidden.
The correlation patterns reveal that exclamations, with the highest factor
loading, correlate very strongly with swearing, pragmatic particles, verbs
of sensory perception, verbs of emotion, expressions of proximity and first
person pronouns, modals of necessity and volition; but also with second
person pronouns and direct questions. Of the features of Factor 2, the pro-
noun it correlates strongly with verbs of observation and modals of pos-
sibility and necessity; and likewise the most salient features of Factor 3,
the third person pronouns s/he, THAT-deletion and verbs of emotion and
imagination, correlate with one another very strongly.

Table 1. A three-factor solution (salient loadings having weights larger than 0.5).

Factor 1 Factor 2 Factor 3 Sumsqr

EXCLAMATIONS 0.888 0.045 0.201 0.831

SWEARING 0.582 -0.046 0.188 0.376
IT -0.145 0.682 -0.189 0.522
THAT-deletion 0.275 0.026 0.580 0.413
SENSORY PERCEPTION 0.517 -0.061 0.378 0.413
OBSERVATION -0.157 0.444 -0.179 0.254
ACTIVITY (MENTAL) 0.127 0.600 0.294 0.462
STATE (MENTAL) 0.430 0.335 0.487 0.534
NOW 0.652 0.102 0.104 0.446
HERE 0.653 -0.091 0.004 0.435
DIRECT QUESTIONS 0.838 0.178 0.215 0.780
PRAGMATIC PARTICLES 0.843 0.108 0.185 0.756
PERSON1 0.623 -0.064 -0.099 0.402
PERSON2 0.828 0.327 0.110 0.804
PERSON3 -0.176 -0.129 0.973 0.995
POSSIBILITY 0.155 0.720 0.220 0.591
NECESSITY 0.416 0.696 0.154 0.682
VOLITION 0.693 0.312 0.351 0.700
EMOTION/IMAGINATION 0.272 0.369 0.545 0.507
Sumsqr 5.836 2.570 2.499 10.904
 Genre conventions: Personal affect 207

Figure 2a: 3 factor solution combined variables

Fl

F2
208 Irma Taavitsainen

Figure 2b: 3 factor solution combined variables

Fl

F3
 Genre conventions: Personal affect 209

Figure 2c: 3 factor solution combined variables

F2

F3
210 Irma Taavitsainen

Factor 1 (eigenvalue 5.836):

The strongest features contributing to this factor are exclamations (factor

loading 0.888), pragmatic particles (0.843), direct questions (0.838), sec-
ond person pronouns (0. 828), modals expressing volition (0.693), first
person pronouns (0.623), expressions of proximity (here 0.653 and now
0.652), swearing (0.582) and sensory perception verbs (0.517). This factor
can be interpreted as "Interaction" (see below) .

Factor 2 (eigenvalue 2.570):

The most important features contributing to this factor are modal auxilia-
ries of possibility (0.720) and necessity (0.696), the pronoun it (0.682),
private verbs expressing mental activity (0.600) and observation (0.444).
This factor is interpreted as "Reasoning" (see below).

Factor 3 (eigenvalue 2.499):

The third person pronouns he and she contribute to this factor very strong-
ly (0.973); THAT-deletion is next in importance (0.580), then come private
verbs expressing emotions and imagination (0.545) and mental state
(0.487). This factor is named "Narration" as it seems to reflect narrative
concerns.

For illustration, see Figures 2a, b, and c (pp. 207-209).

6.3. Interpretation of factors

Factor scores of individual texts tell how important the co-occurrence of
the features constituting the factor are for each text. Thus the basic unit
of measurement is an individual text, and texts of a given genre may be
very different from one another in respect to various factors. Each text has
its own place in this three-dimensional scale (see Table 2, pp. 211-212).
Examples of texts with high factor scores on at least one factor are given
below as "typical texts".
 Genre conventions: Personal affect 211

Table 2. Factor scores of texts (for abbreviations, see pp. ix-x and 345-363 in
this volume).

CAT PERIOD TEXT WORDC PARI INF INT Fl F2 F3

FICT 3 PENNY 6560 NI INF INT 3.092 0.516 0.504

PLAY 1 UDALL 4620 XX INF INT 2.446 0.364 0.052
PLAY 1 STEVENSO 5950 XX INF INT 2.278 -0.068 0246
PLAY 3 VANBR 7190 XX INF INT 2.180 0.460 -0.054
PLAY 2 SHAKE 6170 XX INF INT 2.137 -0.082 0.442
PLAY 2 MIDDLET 5640 XX INF INT 1.977 -0.143 -0245
PLAY 3 FARQUHAR 5550 XX INF INT 1.976 -0.118 -0232
FICT 2 DELONEY 7320 NI INF INT 1.165 -0.454 1.166
AUTO 1 MOWNT 5740 NN X X 0.861 -0.383 0303
HAND 3 WALTON 5370 IS X INT 0.697 0.822 -0.391
FICT 1 HARMAN 5120 NI X X 0.523 -0.656 1.771
FICT 2 ARMIN 5170 NI INF INT 0.446 -0.731 0.869
HAND 2 GIFFORD 6190 IS X INT 0.085 0.352 1.715
DIAR 2 HOBY 6050 NN INF X 0.010 -1.550 -1.642
SCIE 2 BLUNDEV 6420 EX X X -0.054 0.626 -1.113
SCIE 1 RECORD 6700 EX X X -0.079 0.138 -1284
AUTO 3 FOX 5560 NN X X -0.145 -0.806 0.534
HAND 2 MARKHAM 6100 IS X X -0.153 1.083 -0.470
EDUC 1 ASCHAM 4900 IS/EX X X -0.215 -0.235 -0.756
TRAV 1 LELAND 6860 NN X X -0.216 -1.136 -1.487
TRAV 2 JOTAYLOR 8850 NN X X -0.231 -0.367 -0.784
PHILO 2 ELIZ 6880 XX X INT -0.233 2.281 0.614
DIAR 3 EVELYN 6070 NN INF X -0.236 -0.937 -0.980
SCIE 2 CLOWES 6620 EX X X -0.236 0.339 -1.056
FICT 1 MERRY 6430 NI INF X -0.244 -0.385 0.803
HAND 3 LANGF 6000 IS X X -0.249 1.005 -0.946
DIAR 2 MADOX 6470 NN INF X -0.268 -0.412 -0.326
TRAV 3 FIENNES 5140 NN X X -0.272 -1.126 -0.893
DIAR 3 PEPYS 5140 NN INF X -0.274 -0.474 -0.157
PHILO 1 COLVILLE 9890 XX X INT -0.280 1.947 -0.062
TRAV 1 TORKINGT 7240 NN X X -0.348 -1.177 -0.968
TRAV 3 FRYER 5330 NN X X -0.359 -0.563 -0.980
EDUC 3 LOCKE 5200 IS X X -0.370 1.054 0377
PHILO 3 PRESTON 8820 XX X INT -0.375 2.367 0.102
212 Irma Taavitsainen

Table 2. Continued.

CAT PERIOD TEXT WORDC PARI INF INT Fl F2 F3

BIO 1 ROPER 5440 NN X X -0.433 0.026 0248

HAND 1 FITZH 5150 IS X X -0.442 1.282 0.039
DIAR 1 MACHYN 6280 NN INF X -0.443 -1.453 -1.127
DIAR 1 EDWARD 6780 NN INF X -0.471 -1.047 -0.549
EDUC 2 BRINSLEY 5540 EX X X -0.480 1.561 0221
HIST 3 MILTON 5820 NN X X -0.493 -0.717 -0.045
HIST 1 FABYAN 5420 NN X X -0.525 -1.117 -0.174
HAND 1 TURNER 4850 IS X X -0.535 0.047 -0.709
TRAV 2 COVERTE 5930 NN X X -0.567 -0.898 0.144
BIO 2 PERROT 4810 NN X X -0.576 -0.444 1249
HIST 2 STOW 4810 NN X X -0.591 -0.943 -0.331
EDUC 1 ELYOT 5540 IS/EX X X -0.609 0.344 0.489
EDUC 2 BACON 5680 EX X X -0.612 0.432 -0.949
SCIE 3 BOYLE 5220 EX X X -0.616 1.303 -1.043
SCIE 3 HOOKE 6060 EX X X -0.652 0.858 -1241
HIST 2 HAYWARD 5280 NN X X -0.659 -0.497 -0.105
EDUC 3 HOOLE 6120 IS X X -0.669 1.194 -0.239
HIST 1 MORERIC 5670 NN X X -0.697 0.117 1.123
AUTO 2 FORMAN 4120 NN X X -0.853 -0.740 1.791
FICT 3 BEHN 5480 NI X X -0.876 -0.219 2.163
SCIE 1 VICARY 6180 EX X X -0.883 0.617 -0.6*3
BIO 3 BURNETR 6170 NN X X -1.121 -0.439 2.187
HIST 3 BURNETC 5820 NN X X -1.231 -0.819 2.828

A personal point of view, egocentric focus, or interactive communication

are central to this study; these are all aspects of personal affect that might
be relevant to defining its quality particularly in fiction. Thus their co-
occurrence patterns should reflect more specialised qualities of personal
affect. Unlike earlier studies of involvement, first and second person pro-
nouns contribute to the same factor. Thus the well-established division in-
to ego-involvement, involvement with the second person, and involvement
with the subject-matter or evidentiality is not dominant. The interpretation
of Factor 1 is not straightforward as both parts of the dichotomy between
"self' and "other", i.e. the first and the second person pronouns, are im-
portant in this factor.
 Genre conventions: Personal affect 213

6.3.1. Factor 1 (eigenvalue 5.836), "Interaction"

This factor can be called "Interaction", and it is most powerful in explain-

ing underlying variation in this study. It is evident that the interactive
quality has to be assessed in more depth. Texts that score highest belong
to the genre of fiction. Penny Merriments from the last period (EModE3)
scored highest on Factor 1, and had positive values on both others; exam-
ples of texts help to define the qualities of personal affect in them. Com-
ments of stylistic features not captured by the statistical analysis are also
included in the discussion (see 6.1):

FICTION: PENNY, factor scores: Factor 1: 3.092 (Factor 2: 0.516, Factor

3: 0.504)
(Ka) Who shall be my Father to give me.
(Jo) Thou mayest ask Jack Wheeler, but I know he had rather had
thee himself.
(Ka) Oh fie no, I will not ask him, he will take it for an affront, I
will rather ask old father Bandol for he us'd to call me Daughter,
and he will take it kindly.
(Jo) Do then.
(Ka) Does it not make you ashamed to talk of these things.
(Jo) No I promise thee, I am proud of it, and so art thou I believe,
but that thou wilt not confess it.
(Ka) I would it were once over.
(Jo) So would I, i'd as live as a groat. {Penny Merriments 119)

This text is written in dialogue form with first and second person pro-
nouns, direct questions, interjections, contracted forms, and some popular
phrases. It is obviously written in imitation of natural language, with a
high emotional loading between the participants.

FICTION: DELONEY, Factor 1: 1.165 ( Factor 2: -0.454, Factor 3: 1.166)

Who was it that checkt thee, I pray thee tell mee? was it not your old
gossip, dame dayntie, mistresse trip and go? I beleeue it was.
Why man if it were she, you know shee hath beene an old house-
keeper, and one that hath known the world; and that shee told mee
was for good will.
Wife (quoth hee), I would not haue thee to meddle with such light
braind huswiues, and so I haue told thee a good many times, and yet
I cannot get you to leaue her company. Leaue her company? why
214 Irma Taavitsainen

husband so long as she is an honest woman, why should I leaue her

company? Shee neuer gaue mee hurtfull counsell in all her life, but
hath alwayes been ready to tell me things for my profit, though you
take it not so. Leaue her company? I am no gyrle I would you should
well know, to bee taught what company I should keepe: I keepe none
but honest company I warrant you. Leaue her company ketha? Alas
poore soule, this reward she hath for her good will. I wis I wis, she
is more your friend, then you are your owne. Well let her be what
she will sayd her husband ... {The Pleasaunt History of... lack of
Newberie 73)

In the above text, Factors 1 and 3 are equally strong. The text is mostly
in dialogue form, though a narrator is present explicating what is going on
in the story. Direct questions, pronouns of the third, first and second per-
son, interjections, assertations and repetitions are frequent in the above
passage and elsewhere in the work.

FICTION: HARMAN, Factor 1: 0.523 ( Factor 2: -0.656, Factor 3: 1.771)

"What! your neuewes?" quoth she. "My neuewes?" quoth this parson;
"I trowe thou art mad." "Nay, by god!" quoth this good wife, "as
sober as you; for they tolde me faithfully that you were their vncle:
but, in fayth, are you not so in dede? for, by my trouth, they are
strau[n]gers to me. I neuer saw them before." "O, out vpon them!"
quoth the parson; "thye be false theues, and this night thei compelled
me to geue them al the money in my house." "Benedicite!" quoth this
good wife, "and haue they so in dede? as I shall aunswere before
god, one of them told me besides that you were godfather to him, and
that he trusted to haue blessinge before he departed." "What! did
he?" quoth this parson; "a halter blesse him for me!" "Me thinketh,
by the masse, by your countenance you loked so wildly when you
came in," quoth this good wife, "that somthing was amis." "I vse not
to gest,"... (A Caveat or Warening for Commen Cursetors 40)

Factor 3 is strongest in this text, but there are passages in which features
of Factor 1 dominate. The above scene is in dialogue form and it provides
a highlight of a story, used as an exemplum, with a strong emotional load-
ing with exclamations, interactive questions and mild swearing.
 Genre conventions: Personal affect 215

FICTION: ARMIN Factor 1: 0.446 (Factor 2: -0.731, Factor 3: 0.869)

But they come; and up they came, and to the king they goe, who,
being with the lord treasurer alone, merry, seeing them two, how Will
had got another foole, knew there was sport at hand. How now! sayes
the king, What news with you? O, Harry! sayes he, this is my owne
uncle; bid him welcome. Wei, said the king, he is welcome. Harry,
sayes hee, heare me tell thee a tale, and I will make thee rich, and
my uncle shall be made rich by thee. Will tels the king how Terrils
Frith was inclosed. Tirrels Frith! sayes the king; what is that? Why,
the heath where I was borne, called by the name of Tirrels Frith: now
a gentleman of that name takes it all in, and makes people beleeue
it is all his, for it took the name from him; so that, Harry, the poore
pine, and their cattle are all undone without thy help. And what
should I doe? sayes the king. Marry, sayes Will, send to the Bishop
of Hereford; hee is a great man with Terril: commaund him to set the
Frith at liberty againe, who is now imprisoned by his means. And
how shall I be rich by that? sayes the king. The poor will pray for
thee, sayes Will, and thou shalt bee rich in heauen, for on earth thou
art rich already. (A Nest of Ninnies 44)

The basis of Armin's work is third person narration reflected by the high
factor score of Factor 3, but direct quotations with questions, exclama-
tions, and pragmatic particles are incorporated into the text, and first and
second person pronouns alternate with the third person. These are features
that fall on Factor 1. The result is a combination of different view points
(see below, discussion of Factor 3).
Some of the non-literary texts score high on this factor, Mowntayne's
autobiography being the extreme example:

AUTOBIOGRAPHY: MOWNTAYNE, Factor 1: 0.861 (Factor 2: -0.383,

Factor 3: 0.303)
"No, (sayed I,) as I ham borne to dye, contentyd I ham so to doo
whan God wyll; but to be made awaye after sowche slyghte, I wolde
be verye lothe; and therfor, yfe that yow have nothynge to showe for
your dyscharge, acordynge as I have requyryd of yow, I tel yow trwe
that I wyll not dye. Take yow good heed therfor to your selve, and
loke that I myscary not, for yfe that awghte come unto me but good,
yow and yours are lyke to knowe the pryse of yt, be yow well assur-
216 Irma Taavitsainert

yd therof. Whan dyd yow ever see anye man put to deathe, before he
was condemnyd to dye?" "That ys trwe, (sayd he;) and are yow not
condemnyd?" "No, (sayd I,) that I ham not, nether was yet ever ar-
aynyd at anye sesyones." "Than, (sayed he,) I have been greatly mys-
eynformyd. I crye yow marsy; for I hade thowghte that yow had been
bothe araynyd, and also condemnyd to dye, beynge sent hether for to
suffer yn thys plase, bycawse that yow were here agaynyste the
quene with the ducke of Northethomeberland." "Well, (sayed I,)
thoos materes hathe bene alredye suffysyently answeryd before your
betteres; but I praye yow, syr, and a man myghte aske yow, whoos
man are yow, or to whome doo yow belonge?" "Maiye! (sayd he,)
I ham not ashamyd of my maister, I wolde thow showldest knowe yt,
as thow arte". (Narratives of the Days of the Reformation 200-201)

The story is half in dialogue form, half in indirect narration. Some of the
speech quotations have a very strong emotional loading, with first and sec-
ond person pronouns, interjections, pragmatic particles and swearing.

HANDBOOK: WALTON, Factor 1: 0.697; (Factor 2: 0.822, Factor 3:

-0.391)
(Ven.) Gentleman Huntsman, where found you this (Otter)?
(Hunt.) Marry (Sir) we found her a mile from this place a fishing;
she has this morning eaten the greatest part of this (Trout); she has
only left thus much of it as you see, and was fishing for more; when
we came we found her just at it: but we were here very early, we
were here an hour before Sun-rise, and have given her no rest since
we came; sure she will hardly escape all these dogs and men. I am
to have the skin if we kill her.
(Ven.) Why, Sir, what's the skin worth?
(Hunt.) 'Tis worth ten shillings to make gloves; the gloves of an
(Otter) are best fortification for your hands that can be thought on
against wet weather.
(Pise.) I pray, honest Huntsman, let me ask you a pleasant question,
do you hunt a beast or a fish?
(Hunt.) Sir, It is not in my power to resolve you, I leave it to be
resolved by the Colledge of Carthusians... (The Compleat Angler
210-211)
 Genre conventions: Personal affect 217

The other non-literary text that has a fairly high score on Factor 1 is Wal-
ton's handbook, written in dialogue form according to the conventions of
instructive texts. It is worth noting that Factor 2 has an even higher score,
in accordance with the instructive purpose of this text.

DIARY: HOBY Factor 1: 0.010, (Factor 2: -1.550, Factor 3: -1.642)

The positive score on Factor 1 obviously depends on the high frequency

of the first person singular pronoun. See the example on p. 248.

6.3.2. Factor 2 (eigenvalue 2.570) "Reasoning"

Factor 2 is perhaps the easiest to interpret as its salient features, e.g.

modal auxiliaries of possibility and necessity, and private verbs of mental
activity and observation, mark the subjective reasoning process. This is in
accordance with my earlier studies of the personal viewpoint in scientific
texts in the 17th centuiy. The highest factor scores are found in philo-
sophical and educational writing. 9

PHILOSOPHY: PRESTON Factor 2: 2.367, (Factor 1: -0.375, Factor 3:

0.102)
(Ph) ... Let me ask thee, can that, dost thou think, which needeth
nothing want Power?
(Bo) No, I am not of that Opinion.
(Ph) Thou thinkest right indeed; for if there be any thing which, upon
any occasion of Performance, doth shew a Weakness or want of
Power, it must, as to that, necessarily need foreign Aid.
(Bo) So it is.
(Ph) And therefore Sufficiency and Power are of one Nature.
(Bo) So it truly seems.
(Ph) And thinkest thou that things of this kind are to be undervalued
and contemn'd, or rather to be reverenced of all? (Preston, Boethius
124)

Boethius' De Consolatione philosophie is written in dialogue form accord-

ing to the scholastic tradition. The language is different from Walton's es-
pecially in respect to Factor 1. The contrast to imitation of natural speech
is clear.
218 Irma Taavitsainen

EDUCATION: BRINSLEY, Factor 2: 1.561; (Factor 1: -0.480, Factor 3:

0.221)
Very reason must needs perswade every one of this. For, if they bee
apt much before five yeeres of age, to learne shrewdnesse, and those
things which are hurtfull, which they must bee taught to unlearne
againe; why are they not as well fit to learne those things which are
good and profitable for them, if they be entred and drawne on in such
manner, as they may take a delight and finde a kinde of sport and
play in the same? This delight may and ought to be in all their pro-
gresse, and most of all in their first entrance, to make them the better
to love the schoole, and learning, as we shall see after. (Ludus Liter-
arius, or the Grammar Schoole 9)

This educational text scores high on Factor 2 only. It has high frequencies
of modal auxiliaries expressing possibility and necessity, the pronoun it,
and private verbs of mental activity and reasoning. Scientific writings and
handbooks also score high on this factor, although the quality of involve-
ment is different in them (see note 17).

SCIENCE: BOYLE Factor 2: 1.303 (Factor 1: -0.616; Factor 3: -1.043)

And this brings to my mind, that it has been observed, that Diamonds
draw better whilst rough, than they do after they are cut and polish'd;
which seeming to contradict what has been observed by others and
by us also, that Amber, for instance, attracts more vigorously if the
surface be made very smooth than otherwise, it induces me to conjec-
ture, that, if this Observation about Diamonds be true, as some of my
trials have now and then inclined me to think it, and if it do not in
some cases considerably depend upon the loss of the (Electrical) Sub-
stance of the Stone, by its being cut and ground, the Reason may
possibly be, that the great rapidness with which the Wheels that serve
to cut and polish Diamonds must be mov'd, does excite a great de-
gree of heat, (which the senses may easily discover) in the Stone, and
by that and the strong concussion it makes of its parts, may force it
to spend its effluviable matter, if I may so call it, so plentifully, that
the Stone may be impoverish'd and perhaps also, on the account of
some little change in its Texture, be rendred lesse disposed to emit
those effluvia that are Instruments of Electrical Attraction. (Electricity
& Magnetism 37-38)
 Genre conventions: Personal affect 219

HANDBOOK: FITZHERBERT Factor 2: 1.282 (Factor 1: -0.442, Factor 3:

0.039)
I coulde peraduenture shewe the housbandes dyuerse poyntes that the
wyues deceyue them in: and in lyke maner, howe husbandes deceyue
theyr wyues: but if I shulde do so, I shulde shewe mo subtyll poyntes
of deceypt, than eyther of them knewe of before. And therfore me
semeth beste to holde my peace, least I shoulde do as the knyght of
the Tour dyd, the whiche had many fayre doughters, and of fatherly
loue that he oughte to them, he made a boke, to a good entente, that
they myghte eschewe and flee from vyces, and folowe vertues. In the
whiche boke he shewed, that if they were wowed, moued, or styred
by any man, after suche a maner as he there shewed, that they shulde
withstande it. (The Book of Husbandry 98)

6.3.3. Factor 3 (eigenvalue 2.499), "Narration"

Factor 3 has the third person pronouns he/she as its most salient feature.
THAT-deletion and emotive and imaginative verbs are also important. The
third person is connected with narration, but in contrast to earlier studies,
past and present tense were not important (see note 8). This, again, may
be connected with the frequent use of first person narration with past tense
verbs in scientific texts that have high factor scores on Factor 2. The high-
est scores on this factor are found in non-literary texts, though several fic-
tional texts have high scores as well.

FICTION: BEHN, Factor 3: 2.163; (Factor 1: -0.876, Factor 2: -0.219)

'Twill be imagin'd Oroonoko stay'd not long before he made his sec-
ond visit; nor, considering his quality, not much longer before he told
her, he ador'd her. I have often heard him say, that he admir'd by
what strange inspiration he came to talk things so soft, and so pas-
sionate, who never knew love, nor was us'd to the conversation of
women; but (to use his own words) he said, most happily, some new,
and, till then, unknown power instructed his heart and tongue in the
language of love, and at the same time, in favour of him, inspir'd
Imoinda with a sense of his passion. She was touch'd with what he
said, and return'd it all in such answers as went to his very heart,
with a pleasure unknown before. Nor did he use those obligations ill,
that love had done him, but turn'd all his happy moments to the best
advantage; and as he knew no vice, his flame aim'd at nothing but
220 Irma Taavitsainen

honour, if such a distinction may be made in love; and especially in

that country, where men take to themselves as many as they can
maintain. (Oroonoko 156)

Aphra Behn's text has a very high score on Factor 3. In contrast to other
writings of the same genre, it represents long fiction, and a much later pe-
riod. The story is told in the third person, with long monologues inserted
into the narration. A narrator is present, intruding with first person com-
ments as in the passage above, but such remarks are not very frequent as
the negative score on Factor 1 indicates.

FICTION: HARMAN, Factor 3: 1.791, (Factor 1: 0.523, Factor 2: -0.656)

Α Roge is neither so stoute or hardy as the vpright man. Many of
them will go fayntly and looke piteously when they see, either meete
any person, hauing a kercher, as white as my shooes, tyed about their
head, with a short staffe in their hand, haltinge, although they nede
not, requiring almes of such as they meete, or to what house they
shal com. But you may easely perceiue by their colour that thei cary
both health and hipocrisie about them, wherby they get gaine, when
others want that cannot fayne and dissemble. Others therebee that
walke sturdely about the countrey, and faineth to seke a brother or
kinsman of his, dwelling within som part of the shire; either that he
hath a letter to deliuer to some honest housholder, dwelling out of an
other Shyre, and will shewe you the same fayre sealed, with the
superscription to the partye he speaketh of, because you shall not
thinke him to runne idelly about the countrey; (A Caveat or Waren-
ingfor Commen Cursetors 36-37)

The basis of Harman's work is third person narration as in the above pas-
sage; see also p. 214 for another type of narration inserted in the work.

FICTION: DELONEY Factor 3: 1.166; see the example on pp. 213-214.

FICTION: ARMIN, Factor 3: 0.869, (Factor 1: 0.446, Factor 2: -0.731)

Will Somers, in no little credit in the king's court, walking in the
parke at Greenwich, fell asleepe on the stile that leads into the walk,
and many that would haue gone that way so much loued him, that
they were loth to disease him, but went another way; I, the better
sort, for now adaies beggars are gallants, while gentiles of right blood
seeme tame ruffians: but note the loue Will Sommers got. A poore
 Genre conventions: Personal affect 221

woman, seeing him sleepe so dangerously, eyther to fal backward, or

to hurt his head leaning so against a post, fetcht him a cushion and
a rope; the one for his head, and the other to bind him to the post,
from falling backward: and thus hee slept, and the woman stood by,
attending as the groom of his chamber. (A Nest of Ninnies 42)

Armin's text has positive scores on Factors 3 and 1. Third person narra-
tion forms the basis, but in places it alternates with direct speech and in-
teractive patterns, with first and second person pronouns, direct questions,
and mild swearing. The above example has an imperative form directed
to the reader; controversal statements reflect the theme of carnivalism, i.e.
reversal of ideas.

FICTION: MERRY TALES, Factor 3: 0.803, (Factor 1: -0.244, Factor 2:

-0.385)
A woman ther was whych had had .iiii. husba~des. It fortunyd also
that this fourth husband died & was brought to chirch vppon ye bere/
who- this woma- folowyd & made gret mone & wext very sory. In
so mych that her neybours thought she wold sowne & dy for sorow/
wherfor one of her gossyps cam to her & spake to her in her ere &
bad her for goddes sake to comfort her self & refrayne that lamen-
tac~on or ellys it wold hurt her gretly & p~auenture put her in ieop-
erdy of her lyfe. To who- this woman answeryd & sayd I wys good
gosyp I haue gret cause to morne if ye knew all/ for I haue byryed
.iii. husbandys besyde thys man/ but I was neuer i~ the case y' I am
now/ for there was for there was not one of the- but whe~ that I fol-
owid the corse to chyrch yet I was sure alway of an other husba~d
before that ye corse cam out of my house/ & now I am sure of no
nother husband & therfore ye may be sure I haue gret cause to be sad
and heuy. (A Hundred Mery Talys 19-20)

Merry Tales contains third person narration with THAT-deletion and im-
aginative and emotive verbs. Direct quotations are given at the turning
points of the plot and other highlights of the story; yet the score on Factor
1 is negative, which marks the absence or rarity of features like exclama-
tions, swearing, pragmatic particles, and first and second person pronouns.

HISTORY: BURNET, Factor 3: 2.828; (Factor 1: -1.231, Factor 2: -8.19)

He thought that nobody served him out of love: and so he was quits
with all the world, and loved others as little as he thought they loved
222 Irma Taavitsainen

him. He hated business, and could not be easily brought to mind any:
but when it was necessary, and he was set to it, he would stay as
long as his ministers had work for him. The ruin of his reign, and of
all his affairs, was occasioned chiefly by his delivering himself up at
his first coming over to a mad range of pleasure. (Burnet's History
of My Own Time 168)

BIOGRAPHY: BURNET, Factor 3: 2.187; (Factor 1: -1.121, Factor 2:

-0.439)
He loved to talk and write of Speculative Matters, and did it with so
fine a thread, that even those who hated the Subjects that his Fancy
ran upon, yet could not but be charmed with his way of treating
them. Boileau among the French, and Cowley among the English
Wits, were those he admired most. Sometimes other mens thoughts
mixed with his Composures, but that flowed rather from the Impres-
sions they made on him when he read them, by which they came to
return upon him as his own thoughts; than that he servilely copied
from any. For few men ever had a bolder flight of fancy, more sted-
dily governed by Judgment than he had. No wonder a young man so
made, and so improved was very acceptable in a Court. (The Life and
Death of John Earl of Rochester 8)

Two non-literary texts by the same author have the highest scores on Fac-
tor 3. They represent the extreme ends on the positive side on this factor
and on the negative side on Factor 1 reflecting the absence of interactive
features. Private verbs of emotion and imagination are prominent, as are
third person pronouns.

AUTOBIOGRAPHY: FORMAN, Factor 3: 1.791 (Factor 1 -0.853; Factor

2 -0.740)
Ther was a man of good reputation and wealth, that dwelte not far
from Simon's master, that had a proper fine maiden to his only
daughter, the which being but yonge of yeares and younger then
Simon, that loved Simon wonderfull welle, and wold suerly see him
once a daie, or ells she wold be sicke. And often she wold com to
Simon's master, and entreat him very kindly on holly daies that she
might see him or speake with him, and somtymes to goe to pastymes
with her; and she loved him soe well that yf forty youth were at play
before the dore, in a spacious place which ther was, yf Simon were
not amonge them, she would not be ther; but yf he were there, none
 Genre conventions: Personal affect 223

could kepe her from thence. Yf Simon stode by his master or mis-
tress at the dore she would com and stand by him, and wold not goe
from him tille necessity did comple. And Simon's master well per-
ceyvinge the grete affection of the gentlewoman towardes Simon,
wold often say unto her, mistress An Yene love my boy welle me-
thinkes; and she would aunswer, yea forsothe, yf yt will please youe
to give him leave to go rone with us, wee shall give youe thankes,
sir. Wherupon oftentymes he wold give him leave; and as for Simon,
he loved her not but in kindnes, but because she was soe kind to
Simon, he wold doe anythinge he could doe for her. And his love on
her syde lasted longe, as herafter shalbe showed. (The Autobiography
and Personal Diary of Dr. Simon Forman 9)

Forman's autobiography and Perrott's biography score high on this factor;

both have negative scores on the two other factors in spite of occasional
exclamations, first person passages and direct questions. Negative scores,
however, tell that they are very rare (see 6.2).

BIOGRAPHY: PERROTT Factor 3: 1.249 (Factor 1: -0.576, Factor 2:

-0.444)
Shortly, after Sir John Perrott returned from France, and came to the
Court of England, where he lived at great Charge, and at soe high a
Rate, as he grew into great Dept, and ranne so farre into Arrearages,
that he began to mortgage some of his Lands, and yet did owe some
seven or eight Thowsand Pounds, being like to Allexander the Great
in this, who agaynst his Expedition into Persia did put most Part of
his Possessions belonging to the Crowne of Macedon in Pawne. And
being asked by Perdica, his chiefe Commander, what he left behynd
hym, answered Hope. So this Knight spending his Patrimony (as
many of his Yeres and Calling do now-a-dayes, wastfully, and above
their Habiliitie) had lefte but a bare hope to recover his Estate.Yet he
at Length did begin to bethink hymself, and to look back into his de-
caying Fortune, and soe grew much agrieved at his owne Prodigality.
Insomuch that on a Time he walked out of the Court, into a Place
where commonly the Kinge did use to come about that Howre; and
there he began (either as knowing that the Kinge would come that
Way, or els by Chaunce,) to complayne as it were agaynst hymself,
unto hymself: How unfortunate and unwise he was, soe to consume
his Livinge, having wasted a great Part of that in few Yeares, which
his Auncestors had gotten and enjoyed for many Yeres? And must I
224 Irma Taavitsainen

(quoth he) be the man that shall overthrow my Howse, which hath
continued soe longe? It had byn better I had never byn born. (The
History of That Most Eminent Statesman, Sir John Perrot 31-32)

The above passages of texts with analyses of their linguistic features aim
at highlighting the complicated texture of these early works. No simple
answer to the question posed at the beginning, whether there are any sali-
ent features that tell various text types apart and reveal relevant genre con-
ventions, emerges. The quality of personal affect in these texts also needs
further investigation.

6.4. Textual affinities

Table 3. Factor scores of genres, with means (M), standard deviations (D) and
frequencies (F), i.e. numbers of texts.

Category Factor 1 Factor 2 Factor 3

Autobiography Μ 0.87573 -0.04595 -0.64288

D 0.80068 0.86119 0.22790
3 3 3

Biography Μ 1.22783 -0.70997 -0.28559

D 0.96999 0.36263 0.26958
F 3 3 3

Philosophy Μ 0.21783 -0.29573 2.19818

D 0.35229 0.07251 0.22169
F 3 3 3

Diary Μ -0.79672 -0.28022 -0.97870

D 0.55666 0.17279 0.47631
F 6 6 6

Education Μ -0.14267 -0.49266 0.72499

D 0.60611 0.17382 0.66015
F 6 6 6
 Genre conventions: Personal affect 225

Table 3. Continued.

Category Factor 1 Factor 2 Factor 3

Fiction Μ 1.21268 0.68433 -0.32160

D 0.63329 1.37172 0.45020
F 6 6 6

Handbook Μ -0.12693 -0.09960 0.76536

D 0.96102 0.44740 0.47233
F 6 6 6

History Μ 0.54932 -0.69949 -0.66273

D 1.23272 0.27179 0.43532
F 6 6 6

Play Μ 0.03489 2.16561 0.06882

D 0.27114 0.18096 0.26879
F 6 6 6

Science Μ -1.06343 -0.41997 0.64697

D 0.22794 0.34377 0.40757
F 6 6 6

Travel Μ -0.82784 -0.33217 -0.87796

D 0.53415 0.12920 0.34024
F 6 6 6

Total Μ 0.00000 -0.00000 -0.00000

D 0.99647 0.97327 0.91928
F 57 57 57

The next step is to assess the means and standard deviations of the genres.
In the structure of the corpus, the genres are represented by two texts in
each subperiod, which means a great deal of generalisation at this level.
The samples were often deliberately selected to stand far apart from one
another, and thus the standard deviations are important as indicators of the
possible heterogeneity of texts grouped together on external evidence. On
the other hand, the representatives of some genres seem to be very close
to one another in respect to their personal affect features. 10
226 Irma Taavitsainen

For the purpose of this study, textual affinities revealed by the three fac-
tors are of great importance (see Table 3 above). If the ultimate aim of
genre studies is to probe into the essential differences between adjoining,
overlapping genres, the above description of features that correlate and
texts in which they occur serves to indicate the groupings within which
the differences should be found. The overlappings are clear. Factors 1 and
3 are closely interwoven in some texts: while Mowntayne's autobiography
scores high on Factor 1, some fictional works are best described by Factor
3. High scores on both factors are not uncommon, as the above examples
show. Factor 2 is different as it describes philosophical, educational and
scientific texts, as well as handbooks; only Walton's handbook reached a
high factor score on Factor 1 because of the more colloquial features of
the dialogue. In the following assessment I shall concentrate on fiction and
the adjoining genres indicated by Factors 1 and 3.
It is obvious that the parameter values ascribed to plays and some other
texts (INTERACTIVE, INFORMAL, see 4.1 and 4.2) show up well in the
above assessment." The generic conventions of plays are very different
from those of the other genres, and therefore they are excluded from the
second phase of my search for distinctive discursive properties in fiction
and adjoining genres. They are completely in dialogue form, in imitation
of spoken language, and the conventions must necessarily be different.
The narration of events takes place through direct speech quotations, re-
created on stage each time the play is performed. The situations that form
the context of linguistic behavior are acted, and thus the reference system
must be different from that of purely written texts. (See Figures 3 a, b, c.
They indicate how far apart plays stand from the other genres in focus.)

7. Combining quantitative and qualitative approaches

In this phase of the study the material was limited to comprise fiction and
the genres that stand nearest to it: history, biography, autobiography, diary
and travelogue. Factor analysis proved the close affinity of these genres,
defined by external evidence, in respect to the linguistic features that re-
flect personal affect. The pattern revealed by factor analysis is, however,
intricate and requires further consideration (see 6.3).
 Genre conventions: Personal affect 227

Figure 3a: 3 factor solution combined variables

F2

BOETH

HA ND
]
«
I

BIO FICT

HIST A JTO
TRAV
DIAR

- 1 0 1 2
228 Irma Taavitsainen

Figure 3b: 3 factor solution combined variables

FL

PLAY

HCT

HA *JD AU' ι υ

BOETH
SC :IE

HIST BIO

-1.5 -0.5 0.5 1.5

F3
 Genre conventions: Personal affect 229

Figure 3c: 3 factor solution combined variables

F2

BOETK

SC :IE
m ίΡ

PLAY

mr
HIST AU' ro

TRAV
DIAR

-1.5 -1 -0.5 0 0.5 1 1.5

230 Irma Taavitsainen

The genres selected for scrutiny are all relevant to the evolution of early
fiction. The difference between them and fiction may be difficult to pin-
point, but it may be attempted via the features indicated by Factors 1 and
3, though the distinction does not seem to be striking on the basis of fac-
tor scores only. Fiction may have its own devices in story telling. Ques-
tions of deixis and interactive focus may grow in importance in adjoining
text types that by definition employ first person pronouns, and there may
be subtle differences not captured by factor analysis. The categories may
be fuzzy, and it may prove difficult or impossible to tell these non-fiction
genres apart from fictional accounts with similar contents. The questions
to be asked are numerous. Where are the differences, then? Are there any
that separate the adjoining genres? How do non-literary texts of adjoining
genres stand in relation to fiction? Are there any differences in respect to
the interactive focus or narrative concerns? Can more subtle shades of per-
sonal affect be discerned? Is there any difference in the intensity, empha-
sis, or quality of affect? Does the moralising function of early fiction
show in this analysis? And finally, can this analysis reveal mechanisms of
reader involvement in these texts?
According to earlier considerations, the ultimate aim of genre theory is
to define the institutionalised discursive properties of various genres,
which can be interpreted as distinctive features that form a matrix and
contribute to genre conventions in a longer time span. If a significant dif-
ference can be verified in the occurrence of a feature in fiction and text
types that stand near it, this feature must be worth studying in more detail.
It may illuminate the special character and evolution of these text types.

7.1. Imaginative versus non-imaginative narration

For a sharper focus, the material was divided into two groups along the
corpus parameter of imaginative versus non-imaginative narration (NI ver-
sus NN) which divides fiction and the adjoining non-fictional genres into
groups of their own. 12 This parameter is built into the corpus, and this
is the first time to my knowledge that this parameter in the HC has been
systematically applied in research. The linguistic features of personal af-
fect showed a great deal of variation and possible matrices were indicated
by factor analysis, and in the next phase of the study the significance of
the individual features included in these matrices was tested along this
parameter. The aim was to find possible text-type markers. T-tests were
performed in order to find out whether the difference between fiction and
 Genre conventions: Personal affect 231

non-fiction was significant in respect to the same variables that were ap-
plied in the factor analysis. The results of the t-tests of features discussed
below are as follows:

Table 4. Results of t-tests.

Ν of observations: Probability Means St. d.

NN 24, NI 6 NN NI NN NI

Exclamations 0.046 0.156 3.98 0.30 3.54

Swearing 0.028 0.046 1.24 0.15 0.95
THAT-deletion 0.019 1.24 2.84 1.55 1.16
Second person 0.041 1.73 8.88 3.37 6.15
Modals, volition 0.040 3.71 9.91 3.14 5.27
possibility 0.030 3.52 5.33 2.13 1.40
Deictic features
NOW 0.031 1.43 2.69 0.98 1.03

The above table gives the results of t-tests of features with the greatest
factor loadings on Factor 1 and Factor 3; only features that proved signifi-
cant are given. 13 The texts of early fiction with the parameter value NI
("Narration Imaginative") formed one group, and texts in the adjoining
groups with the parameter value NN ("Narration Non-imaginative") the
other.
In addition, an analysis of variance was done with the influence of the
period and the parameter NI versus NN as dependent variables. In general
the results confirmed those of the t-tests; in some cases the analysis of
variance added to the earlier information, as the difference between fiction
and non-fiction, given as insignificant by t-test, proved significant accord-
ing to the analysis of variance. For example, of the tested features prag-
matic particles (probability < 0.0028), direct questions (p < 0.0005) and
this (p < 0.0043), and the second person singular (p < 0.0017) proved
significant as indicators of the difference between imaginative and non-
imaginative narration when the influence of period was eliminated. Ac-
cording to the analysis of variance, exclamations and swearing were the
most outstanding text-type marker of fiction (p < 0.0001).
These statistical tests were used to complement the findings of factor
analysis. They indicate that there are significant differences between the
occurrences of some linguistic features in imaginative and non-imaginative
narration, fiction and the adjoining genres. Some of the features belong to
232 Irma Taavitsainen

the spoken versus written dimension of language, some reflect interaction,

and some are connected with the reference system. As one of the aims of
this paper is to define the quality of personal affect in these texts, I have
grouped these features according to my interpretation, in view of their
relevance to making a difference between imaginative narration and non-
imaginative narration, fiction and related non-literary text types. The aim
is to chart what this difference really consists of.

8. "Surge" features of personal affect

According to my pilot studies personal affect is an umbrella term for the

features among which the markers of the fictional mode can be sought.
One of the main dimensions is the surge of personal feelings and attitudes.
Reflections of the speaker's or writer's mental state have linguistic outlets,
encoded in the texts, and they seem to be of special relevance to the pres-
ent purpose.

8.1. Exclamations, swearing, and pragmatic particles

Table 5. Results of t-tests and f-tests.

Ν of observations: Probability Means St. d.

NN 24, NI 6 NN NI NN NI

t-tests
Exclamations 0.046 0.156 3.98 0.30 3.54
Swearing 0.028 0.046 1.24 0.150.9
f-tests
Exclamations 0.0001
Swearing 0.0001
Pragmatic particles 0.028

In modern linguistics interjections are described as linguistic gestures

which express a speaker's mental state, action or stance, or reaction to a
situation. Thus they encode speaker attitudes and communicative inten-
tions and are context-bound (see Ameka 1992). In spoken language their
 Genre conventions: Personal affect 233

interpretation often depends on paralinguistic features like intonation and

accompanying gestures. In written genres, the linguistic context is the only
clue to their interpretation. Interjections range from onomatopoetic ad hoc
formations to single words, utterances and set collocations; such secondary
interjections merge with swear words and phrases, oaths, and also with
greeting formulas, expressions of politeness, and other routines. Some-
times interjections have been included in the class of particles, because
they are uninflected. The borderlines between various categories are fuzzy
and difficult to draw. Because of the close affinity of secondary interjec-
tions and pragmatic particles, these groups are discussed together with in-
terjections and exclamations.

8.1.1. Occurrences in Early Modern English

In my statistical analysis, the difference between the occurrence of inter-

jections in fiction and in the adjoining genres proved highly significant.
The use of exclamations and swearing seems to be the most powerful text-
type marker of fiction. In quantitative calculations the frequencies are
given per one thousand words of text, but at this point it may be useful to
return to the absolute occurrences:

Table 6. (Absolute) occurrences of exclamations.

Exclamations Total EModEl EModE2 EModE3

Fiction 94 19 27 48
Autobiography 7 6 0 1
History 2 2 0 0
Biography 1 0 0 1
Diary 1 0 1 0
Travelogue 0 0 0 0

The distribution is very uneven across the genres; of the excluded genres
plays had even higher concentrations of exclamations (see Taavitsainen
1995a).
234 Irma Taavitsainen

8.1.2. Conventions in a longer diachronic perspective: Late Middle

English

Inspired by the results of this study, I extended my assessment of inter-

jections and swear words to the preceding period (1350-1500) with the
help of the Late Middle English part of the HC, and the Chaucer Con-
cordance (for details, see Taavitsainen forthcoming a and b). The aim was
to see how far back in time the convention could be traced, and whether
the distribution of these linguistic features showed a similar pattern earlier.
The difference between fiction and non-fiction proved highly significant
in the Late Middle English material as well.14
Interjections were particularly frequent in Chaucer's works, and the ex-
ample he set must have inspired authors after him. The stylistic quality of
interjections ranges from solemn invocations that only occur in some
genres to burlesque exaggerations of brisk action with collocations of
interjections, equally restricted in scope. In some cases Chaucer modifies
and colours the meanings of these words with shades of irony, which use
is not found in any other writer's work in this period.
An emotive loading is often present in speech quotations and passages
that employ interjections. The audience is supposed to live with the char-
acters and their emotions. Interjections may also have a conative function,
directing text internally to another character, or the reader may be request-
ed to pay special attention; this is specially frequent with the originally
emphatic particle lo. In addition to the element of personal affect, inter-
jections have textual functions. In several genres they are used to organise
discourse and mark turn-taking in the perfomance of the text, aiding the
hearer to distinguish various speakers' turns when read aloud. This is es-
pecially clear in drama and romances, in which the use of interjections at
the beginning of each turn is almost regular. In Chaucer's works inter-
jections have acquired other textual functions as well. Collocations of ex-
clamations are used to signal important turning-points in the plot; the pat-
terns are genre-specific (see Taavitsainen 1995b). An extreme develop-
ment can be seen in Chaucer's use of interjections with a special future
meaning, as a foregrounding device. In these passages interjections are
embedded in the narration to create apprehension in the audience by omi-
nous exclamations like alas and lo.ls
 Genre conventions: Personal affect 235

8.1.3. The art of story-telling

Nearly all interjections in the Late Middle English and the Early Modern
English period are found in direct speech quotations, frequently in sen-
tence-initial position. When compared with the Late Middle English func-
tions of interjections, the marking of turn-taking is not evident in the Early
Modern material. This issue is connected with several important points
concerning the consumption of these texts. With the spread of literacy
reading habits changed, and with cheaper production methods and ma-
terials, books and booklets became more readily available for a wider
audience. The change was gradual, and reflected the transition from medi-
eval to modern (cf. 2.5).
The instances of interjections in the present material are especially nu-
merous in Merry Tales (EModEl) and Penny Merriments (EModE3) which
are both collections of short funny stories. They continue the tradition of
medieval fabliaux, which can also be seen in the use of interjections and
swear words. Medieval narration in verse has been transformed to fit the
prose diction in these stories (see below). According to the fabliau tradi-
tion, the characters represent the lower or middle layers of society and the
speech quotations have the pretence of natural talk with colloquial idioms,
oaths and swear words.
Some of the phrases obviously belonged to the stock-in-trade of the
genre. Alas occurs frequently in emotive vocatives with adjectives of sym-
pathy like poor and good, and in passages of regret and lamentation as the
very stereotype of such feelings.
Chaucer's art of story-telling is reflected in the use of interjections in
marking turning-points in the plot. This tradition is carried over to Early
Modern short fiction. Alas is used in this way to highlight important pas-
sages and lend them vividness and emotional colouring:

sayd to ye maltman alas I haue let my boget fal in to ye water &

there is .xl. li. of money therin yf thou wylt wade in to ye water &
go seke it & get it me agayne I shall gyue ye .xii. pence for thy la-
bour ... This maltman within a whyle after with grete payne & depe
wadynge founde ye boget & came out of the water & sawe not his
felowe there & sawe that his clothys & money were not there as he
left them suspectyd ye mater and openyd the boget and than founde
nothynge therin but stonys cryed out lyke a mad man and ran all nak-
236 Irma Taavitsainen

yd to london agayne and sayde alas alas helpe or I shall be stolen.

For my capons be stolen. My hors is stolen. My money and clothys
be stolen and I shall be stolen myself. And so ran aboute the stretys
in london nakyd and mad cryenge alway I shall be stole. I shall be
stolen ... (A Hundred Mery Talys 149)

The collocation fye, for shame, fye provides the climax of one of the short
comic anecdotes in Harman. It is a watch word and when uttered at the
wrong moment it provides the jest of the story around which the narration
is built. The choice of a phrase like this to mark the culmination of a
comic tale is in accordance with the conventions derived from medieval
fabliaux and Chaucer's use of cumulative lists of interjections at the
turning-points of such stories (see Taavitsainen 1995b).

geue me a watche worde a loud when hee goeth aboute to haue his
pleasure of the, and that shall be "fye, for shame, fye" and I wyll bee
harde by you wyth helpe ...
"And are you not ashamed? neuer a whyte," sayth he, "lye downe
quickely." "Now, fye, for shame, fye," sayth shee a loude, whyche
was the watche word. At the which word, these fyue furious, sturdy,
muffeled gossypes flynges oute, and takes sure holde of this betrayed
parson, sone pluckinge his hosen downe lower, ... (Harman, A Caveat
or Warening for Commen Cursetors 72)

In Armin's work the use of interjections is somewhat more elaborate than

in the above mentioned texts. The passage with the most marked use takes
an outsider's angle with third person pronouns, and yet it uses Ol several
times as direct speech quotations to shift the angle; ah is also used in this
way. This conflict of viewpoints is effective and it does not occur in other
texts of the present material.
Exclamations, together with short questions and quick changes of turns
in the fictional conversation create a vivid narration with an intensive
emotional tone, perhaps in imitation of pangs of pain in the first passage,
and anxiety in the second. Thus there is a link to the use of interjections
in passages that provide the climax of the story:

The pyper and the minstrel, being in bed together, one cryed O! his
backe and face; the other, Ol his face and eye: the one cryed Ol his
pype! the other, Ol his fiddle! Good mussicke or broken consorts,
they agree well together; (Armin, A Nest of Ninnies 11-12)
 Genre conventions: Personal affect 237

Jack fed, and feeding greedily, (more to anger the cooke, than dis-
appoint Sir William) ever as he burnt his mouth with hast, dipt the
pie in the water to coole it. Ol says the cooke, it is Sir William's
owne pie, sirra. Ol sayes Jack hang thee and Sir Willy too. I care not,
it is mine now. Save Sir William some, sayes one; save my lady
some, sayes another. By James, not a bit, sayes Jack. (Armin, A Nest
of Ninnies 14)

The function of foregrounding, found in Chaucer's works,16 is not obvi-

ous in the present material. The nearest usage can be found in Aphra
Behn's text, which is different from the other fictional works here. It is a
long narration that can be considered a forerunner of the romantic mode
of story-telling. The use of interjections seems to be consciously targeted
at manipulating readers' emotions, as these exclamations are embedded in
long monologues or in the middle of narration. The interjection provides
a side remark to provoke the readers' emotions and contributes to the
overall affective tone of the text (see Taavitsainen forthcoming d).17

At this character, his old heart, like an extinguish'd brand, most apt
to take fire, felt new sparks of love, and began to kindle; and now
grown to his second childhood, long'd with impatience to behold this
gay thing, with whom, alas! he could but innocently play. But how
he shou'd be confirm'd she was this wonder, before he us'd his
power to call her to court, (where maidens never came, unless the
king's private use) he was next to consider ... (Behn, Oroonoko 157)

He cou'd not be convinc'd he had no cause to sigh and mourn for the
loss of a mistress, he cou'd not with all his strength and courage
retrieve. And he would often cry, Oh, my friends! were she in wall'd
cities, or confin'd from me in fortifications of the greatest strength;
did inchantments or monsters detain her from me; I wou'd venture
through any hazard to free her: But here, in the arms of a feeble old
man, my youth, my violent love, my trade in arms, and all my vast
desire of glory, avail me of nothing. Imoinda is as irrecoverably lost
to me, as if she were snatch'd by the cold arms of death: Oh! she is
never to be retrieved. If I wou'd wait tedious years, till fate shou'd
bow the old king to is grave, even that wou'd not leave me Imoinda
free (Behn, Oroonoko 160)
238 Irma Taavitsainen

8.1.4. Non-literary text types

Although there is a significant difference between the frequencies of in-

terjections, swear words and pragmatic particles in imaginative and non-
imaginative narration, examples are found in other genres as well. Mown-
tayne's autobiography has several occurrences. They are all found in di-
rect speech quotations.

"Ye", sayed he. "And ys thys the beeste servys that yow can doo my
lorde your master? Fye, for shame, fyel wyl you folowe now the
bludye stepes of that wyckyd man your master! who ys unworthye,
before God I speake yt, bothe of the name and place that he hathe
and ys calyd unto. What sholde moufe yow for to handyll me after
thys sharpe sorte as yow have done, so spytefiillye, beynge here not
yet iij dayes under your kepyng? Wyl yow become a tormentor of
Godys people and prophetes? wyl yow now seas from kyllynge of
bolokes, calvys, and shepe, which ys your ockapasyon (being a buch-
er), and to gyve over your seife moste crwellye to sarve your mast-
ares tourne in sheddynge of Ynnosente blode? Ο man, with what an
avaye (heavy) harte maye yow laye your selve down to slepe at
nyghte, yf that God of hys great marsy doo suffer yow to lyve so
long yn thys your so wycked atempte and enterpryse! (Mowntayne,
Narratives of the Days of the Reformation 201)

After execucyon of which dedely dede ye sayd syr Piers toke great
repentaunce Inso moche that lamentably he sayd alas what haue we
done we haue now put to deth hym that hath ben our Soueraygne and
drad lorde by the space of ,xxii. yeres by reason whereof I shall be
reprochyd of all honoure (Fabyan, The New Chronicles of England
and France 170R.C2)

Now truly, quoth an old gentilman to a young feloe, ye ar far to

blame to mislyke your aunt for she may do you pleasure and I wold
God I had such an aunt. Fy, quoth he, wold I had your land on condi-
cion you had xxty such aunts. (Madox, The Diary of Richard Madox
88)

Even the texts with the lowest negative scores may contain occasional ex-
amples of interjections:
 Genre conventions: Personal affect 239

Soon after I told him, I was glad to find his Style so reformed, and
that he had so entirely overcome that ill habit of Swearing; Only that
word of calling any "damned", which had returned upon him, was
not decent. His Answer was: (Oh that language of Fiends, which was
so familiar to me, hangs yet about me: Sure none has deserved more
to be damned than I have done.) (Burnet, The Life and Death of John
Earl of Rochester 154)

In these examples the interjections are reflexions of spoken language,

spontaneous reactions to the immediate situation.

8.2. THOU and YE; sociolinguistic considerations

According to the statistical assessment, the difference between the use of

the second person singular in fiction and the adjoining genres is signifi-
cant. The absolute occurrences of the second person singular are given be-
low:

Table 7. (Absolute) occurrences of second person singular.

Total EModEl EModE2 EModE3

Fiction 102 20 29 53
Autobiography 23 9 4 10
History 2 0 0 2
Biography 1 0 1 0
Diary 3 0 0 3
Travelogue 0 0 0 0

The historically plural pronoun (YE; oblique form YOU) for polite address
to a single person had been in use from the Early Middle English period.
By c. 1600, the plural form (with YE/YOU as nominative form) had be-
come the unmarked form of address among the upper classes, and the sec-
ond person singular was restricted to expressions of "power and solidar-
ity" as well as heightened emotionality (Mühlhäusler—Harre' 1991: 152;
for a summary of uses, see Calvo 1992). A classical example of the use
240 Irma Taavitsainen

of this pronoun in trials is the well-known adjuration to Raleigh: "I 'thou'

thee thou traitor". Vacillation of the pronouns is common in Shakespeare's
plays, and various functions have been attributed to these shifts. In the
seventeenth century the use of thou gradually became restricted to reli-
gious language (Mühlhäusler—Harre' 1991: 153).
A detailed study of the legend of St Katherine in Southwell Minster MS
7, written about 1500, gives new evidence of the state of the process at
the very beginning of the Early Modern period (Nevanlinna—Taavitsainen
1993). It shows that the use of the second person pronouns was in a state
of flux, as expected, but the emotional use was so well established that a
change of pronoun could be employed as a deliberate device in marking
an emotionally heightened passage and a change in attitude, from respect
to contempt in a saint's legend; similar examples are well known from
Shakespeare's works. In some other passages the use of the polite plural
pronoun implies respect and emotional distance, but with growing famili-
arity it changes to thou. The use seems to be well under control in most
examples, at least where the main characters are concerned. The second
person singular was one of the surge features of personal affect available
in the language in this period, and thus worth a detailed assessment in the
present connection.18

8.2.1. Unmarked and special uses in fiction

The second person singular thou occurs frequently in fiction. The explana-
tion of the high frequency can be found in the social classes that these
texts depict, as the second person singular was the normal pronoun of ad-
dress among country folk and lower classes (Wales 1983: 116). It is evi-
dent that most instances are unmarked as they occur in quotations of
speech among social equals of these classes. Thus it would be exaggerated
to attribute affective meanings to the pronouns in most cases, for example
in the first example below, but the second passage certainly contains an
emotional colouring which is made explicit by other features contributing
in the same direction:

This poller then sayd to hym go thy way streyght to thend of yt long
entre & there thou shalt se whether it be he or no & I wyl holde thy
horse here tyll thou come agayn. This maltman thynkyng to fynde ...
(A Hundred Mery Talys 147)
 Genre conventions: Personal affect 241

(Tweedle) I drinke to thee with all my hart: why thou whoreson

when wilt thou be maried? Ο that I were a young wench for thy sake:
but tis no mater though I be but a poore woman... (Deloney, The
Pleasaunt History of... lack of Newberie 79)

Alongside the normal, unmarked use of thou, special meanings can be

attributed to the use of this pronoun in fiction. In Armin's work, the fool
addresses Sir William with thou·, the king addresses the fool with ye in
return. This is well in accordance with the reversal of values in his work.

Sir William demaunded why hee eate the pye? Because I had a
stomacke, sayes Jack. Would nought else serve, sayes the knight, but
my pye? No, Willy, sayes he, thou would not be angry then, and the
cooke had not been turned away: but all is well — thou art rich
enough to buy more. (Armin, A Nest of Ninnies 15)

This lusty jester, forgetting himself, in fury draws his dagger, and be-
gings to protest, nay; saies the king, are ye so hote? (Armin, A Nest
of Ninnies 48)

In EModE3 all examples in fiction are found in Penny Merriments (1685);

Aphra Behn's text contains third person narration; you occurs only six
times in direct quotations of fictional conversation. The high frequency of
the singular form in EModE3 in Penny Merryments occurs in a dialogue
between a Quaker and his maid (34 out of 53 instances of thou)·, its use
in marking Quaker affiliations is well-known (Mühlhäusler—Harre 1991:
153). The text in Penny Merriments is a parody with exaggerated emotion-
ality and comical imitation of religious language. The use of thou is in
accordance with the special usage among Quakers, but there are other
surge features of personal affect, and the mocking tone is obvious:

(Quaker) I What a War is there even now, betwixt the Inward and the
Outward Man! Satan, Satan, I say unto thee, avoid, by Yea and by
Nay, I charge thee tempt me not: Oh! how the Outward Man pre-
vails! and I can hold no longer; nay, the Light within does say unto
me, That Mary is a Sister, and that Gods Lambs may play, so that
they can but keep it secret from the Wicked; therefore Satan, though
I defie thee and all thy Works, yet will I go unto Mary as I have said:
Mary, why Mary, I say unto thee Mary.
242 Irma Taavitsainen

(Qua) Then Mary, I plainly say unto thee, sit thee down, by yea and
nay I must Touze thee, ingeniously I must.
(Ma) I fie, Master, fie; what is't ye do? the Saints ought not to defile
each other, we shall lose our Credit among the Prophaned; nay,
Master, why Master, Ο fie! wherefore is it you Kiss me so? Ο if my
Dame should know on't!
(Qua) I say unto thee, fear not, fear not I say, thou art a Sanctified
Sister, and one of the Infallible Congregation; and as for thy Dame,
I say she is departed; therefore Mary, again I say unto thee, that the
Spirit within does move me to refresh thee; I burn, I fry, and can
forbear no longer.
(Mary) Oh! Master, Master, I adjure thee, that thou forbear, nay,
Master, Master, Ο Master!
Ο fie! take away thy hand, what is't thee dost? I say unto thee, nay,
nay, I say unto thee nay; Ο let me alone, why dost thee tempt me to
go astray like one of the Wicked? (Penny Merriments 147-148)

The other passage in Penny Merriments, "Loving Kate", is equally affec-

tive, with a dialogue of high emotional charge between the participants
(see the example above). There are several features that contribute to the
overall tone, exclamations for instance. Thou is regularly used by John
when addressing Kate, but the pronoun vacillates in Kate's answers. Inti-
macy and affection are strong qualities in this text, but it is difficult to
know whether the vacillation reflects the change of the unmarked use
among the social class in question, or heightened emotionality.

(Ka.) You need not doubt that, for he cannot keep it from me, and
five pound I have gathered since I came to service, besides my Mis-
tress owes me above half a years wages.
(Jo.) Ο what a happy man shall I be, what a good housewife thou
hast been, thou hast good cloathes too ... (Penny Merriments 116)

(Ka.) Must I dance too.

(Jo.) Ay pretty one, every body will strive to dance with the Bride.
(Ka.) Ide rather dance with thee John than with them all.
(Jo.) So thou shalt my dear. (Penny Merriments 118)
 Genre conventions: Personal affect 243

8.2.2. Surge and social hierarchy in non-fiction

In non-fictional genres the characters belong to higher social classes.

Autobiography employs thou frequently. Mowntayne's situation in his
autobiography was much like Raleigh's, and the use of thou is in accord-
ance with it. The first part of the passage below contains polite, civilised
discussion, but personal affect is certainly present in Mr. Gryffyn's
speech:

"Syr, what make yow here? are you not a Londynar?" "Yes, and yt
lyke your lordshyp." "Howe longe have yow be here prysonar?"
"Halve a yeare, my lorde." "Who sent yow hether?" "Forsothe, my
lorde, that dyd the counsel." Than sayd the hye shyryffe, "My lorde.
thys ys the man that I tolde your lordshyp of; I beseeche yow be
good lord unto hyme, for he hathe bene as quyete a prysonar as ever
came within thys gayell, and hathe usyd hymselve as honestly toward
hys keapar." "Yow speake wel for hym," sayd my lorde; "stand asyed
a whyell tyl yow be called." Yn the meane tyme mr. Gryffyn had a
caste at me, sayenge thus: "Thou arte bothe a traytor and a herytyke"
"No, and yt lyke your worshup, I ham nother of bothe." "Ys not thy
name Mowntayne?" "Yes, forsothe, I wyll never deny yt." "And art
not thow he that my lorde chansler sent hether with a wiyte?" "I am
the same man." "Wel! (sayed he,) and thow be not hangyd I have
marvell. Thow wylte scape narrowly, I beleve." (Mowntayne, Narra-
tives of the Days of the Reformation 206-207)

In other autobiographies it also occurs in emotionally coloured quotations.

It is particularly frequent in the journal of George Fox from the latter half
of the 17th century. This deals with the same topic, imprisonment and jus-
tice, which are highly affective themes:

Thats a snare said I: which all people may take notice of: for I ought
to bee sett free from ye goaler: & this Courte: if I am a free man: as
thou says I ought to bee: & yett thou tendrest ye oath before I am at
liberty but thou ought to lett mee bee at liberty & then thou mightst
have donne thy will. But hee cryed give him ye booke (Fox, The
Journal of George Fox 82)
244 Irma Taavitsainen

In Forman's autobiography and Perrott's biography from EModE2 the pro-

noun of address is in accordance with the social hierarchy and power rela-
tions in addressing social inferiors. Both examples contain pious wishes
and reference to suffering, and are thus connected with the very personal,
even emotive level:

And his master wold say to him, Simon, thou moste suffer as well as
I myselfe; thou seeste we cannote remadie yt as yet, but God will
send a remadie on daie. No we because Simon's master had beaten
him for his mistris sake, herself being in faulte, ... (Forman, The
Autobiography and Personal Diary of Dr. Simon Forman 10)

To whom he sayd, Well Boy, God blesse thee, and I give thee my
Blessing. I wold to God thou wert a shore, and the Queenes Ship
saffe; then I should care the lesse for my seife (Perrott (?), The
History of That Most Eminent Statesman, Sir John Perrot 114)

The other examples from EModE3 are found in Evelyn's diary and Mil-
ton's history writing. All examples in Evelyn's text are from a religious
context, a prayer, which use conforms to the regular pattern at the time.
Milton's history writing provides a good example of the pathetic use of
this pronoun in an emotionally toned passage:

He caus'd his Royal Seat to be set on the shoar, while the Tide was
coming in; and with all the state that Royalty could put into his
countnance said thus to the sea: Thou Sea belongst to me, and the
Land wheron I sit is mine; nor hath any one unpunish't resised my
commands: I charge thee come no furder upon my Land, neither pre-
sume to wet the Feet of thy Sovran Lord. (Milton, The History of
Britain 281)

9. "Self'-"other" dimension

Another possible dimension for the definition of the quality of personal

affect is "self' versus "other". The dichotomy implies an interpersonal fo-
cus, and a key to it is provided by personal pronouns.
 Genre conventions: Personal affect 245

9.1. THOU and YOU — "the other"

The primaiy function of second person pronouns is to express orientation

towards other participants, either the reader or other characters within the
textual world. Second person pronouns had a high factor loading on Factor
1 (0.821) and further statistical tests indicated the significance of this
feature (t-test: ρ < 0.041). It is important to find out who is addressed by
whom. Keeping the moralising function of early fiction in mind, we have
to ask who this second person is in these texts. Are the readers addressed
directly? And if not, how are they taken into account in these genres? I
went through the examples to see if they included direct authorial com-
ments to the readers using the second person pronouns. It was rather sur-
prising that only few could be found:

By thys tale ye may lerne a good generali rule of phesyk.

By thys tale ye may se that the olde p~uerbe ys trew that yt is as gret
pyte to se a woman wepe as a gose to go barefote.
By thys tale ye may perceyue that women ofte tymes be wyfe and
lothe to lose any tyme. (A Hundred Mery Talys 20)

Α Roge is neither so stoute or hardy as the vpright man ... requiring

almes of such as they meete, or to what house they shal com. But
you may easely perceiue by their colour that thei cary both health and
hipocrisie about them, wherby they get gain ... you shall not thinke
... (Harman, A Caveat or Warening for Commen Cursetors 36-37)

here you haue heard the difference twixt a flat foole naturall, and a
flat foole artificiall; one that did his kinde and the other who fool-
ishly followed his owne minde: on which two is written this Rime ...
(Armin, A Nest of Ninnies 12)

The scarcity of examples of this use in HC is even more surprising when

we consider the transition from manuscript transmission to print (see
above), which emphasises a more personal attitude. It is necessary to go
beyond the extracts included in the corpus and look at the books as a
whole for this purpose. A good example is the conclusion in Penny Merri-
ments. It is a poem, directed to the reader, which makes the parody of
Quakers explicit:
246 Irma Taavitsainen

Now Reader, see what here is plainly shown

The Character of such who weep and groan
Who hate all sins but such as are their own.

Divide us from the Wicked, still they cry,

And whilst they seem the Flesh to mortifie,
They hatch vile frauds, and close Adultery.
... {Penny Merriments 152)

But if the moral commentary and personal involvement are not usually
direct, by addressing the reader overtly, how are they incorporated in these
texts? Sometimes this is done more covertly: e.g. Deloney repeats the
same moral theme in his stories. Sometimes a more direct commentary is
found in prefaces or titles. The name of Harman's book, A Caveat or
Warening Signals for Commen Cursetors Vulgarely Called Vagabondes
(1567), signals the moral content, and the pseudoeloquent term "curse-
tors", invented by the author, is revealing. Harman's books had a socio-
logical bias and claimed to be doing a public service by explaining to the
honest citizenry the ways of the criminal classes (Hughes 1991: 156). This
is also found in the direct comment to the reader, quoted above.
Although fiction could have didactic aims, there is a difference between
instructional texts and fiction. Guiding the reader directly is not very
prominent when compared with handbooks and 17th-century scientific
writings. We cannot speak of second-person oriented texts in the same
way, though the frequency of the pronouns is statistically significant. The
overlap with instruction is not conspicuous in this respect, though single
examples like the above can be found.19
Instead of direct reader orientation, the fictional characters of conversa-
tion provide the poles of interaction, and a two-part system with / and
you, the self and the other, must be introduced. Even if the you referred
to in fiction turns out to be a fictional addressee, the use of the pronoun
always involves the actual reader, either by means of direct address or by
the invocation of a more generalised you, with a pretense of being appli-
cable to the current reader by virtue of its gnomic truth value (Fludernik
1995).
 Genre conventions: Personal affect 247

9.2. First person — "the self'

This brings us to "the self' — the first person singular, which had a fairly
high factor loading on Factor 1 "Interaction" (0.623) but not significant
in the present assessment (t-test: ρ < 0.22). By definition autobiography,
diary, and travelogue report on personal experience in the first person sin-
gular. So it is not surprising that no great differences are found. Yet the
matter can be taken further by defining the role and quality of the / in
these texts.
The interactive role of the second person presupposes an interactive first
person, so I assessed the first person singular pronouns to find out whether
the / of these texts was interactive, defined in relation to the other, or an
I with an egocentric, solipsistic focus.
This assessment reveals a difference. It can be stated that in fiction the
self exists only in relation to other; / or we is always paired off with
someone else in the second person, mostly a fictional character, but some-
times the author is paired with the reader (see the examples above).

FICTION:
"that," sayd she, "is your vncles house." "Nay," saith one of them,
"he is not onely my vncle, but also my godfather." "It may well be,"
quoth she, "nature wyll bind him to be the better vnto you." "Well,"
quoth they, "we be weary, and meane not to trouble our vncle to-
night; but to-morowe, god willinge, we wyll see him and do our duty:
but, I pray you, doth our vncle occupy husbandry? what company
hath he in his house." "Alas!" saith she, "but one old woman and a
boy, he hath no occupying at al: tushe," quoth this good wife, "you
be mad men; go to him this night, ...
"Now, by my troth," quoth one of them....
"Where dwellest thou!" quoth this parson. "Alas! sir," saithe this
roge, " / haue smal dwelling, and haue com out of my way; and /
should now," saith he, "go to any towne nowe at this time of night,
they woulde set me in the stockes and punishe me." "Well," quoth
this pitifull parson, "away from my house, either lye in some of my
out houses vntyll the morning, and holde, here is a couple of pence
for thee." (Harman, A Caveat or Warening for Commen Cursetors
38)
248 Irma Taavitsainen

In contrast, diaries record the first person only; the extreme example is
found in Hoby's text in which the viewpoint is purely egocentric and sol-
ipsistic. In the same way travelogues record personal experience, some-
times expressed with the more collective we. The texts also include third
person narration and more neutral description of external reality; occasion-
al instances of other pronouns may occur in such contexts.

DIARY:
Friday 7
After priuat praiers I wrett my notes in my testement, which I
geathered out of the Lector the night before: then / did eate my
breakfast, then / walked abroad and talked of good thinges, so that
I found much Comfort: after / Cam hom I wrett my sermon that was
preached the saboth day before, then / went to priuat praier, and so
to dinner: after which / taked a litle with som of my frendes, and
exercised my body at bowles a whill, of which / found good: then /
cam home and wrought tell 4, then / praied with Mr Rhodes, and
after walked abroad: and when / Came hom / praied priuatly, and
sonne after went to supper: after which I went to the Lector, and then
to bed: (Hoby, Diary of Lady Margaret Hoby 70)

TRAVELOGUE:
Wedynsday, the xv Day of Julii, the ffather Warden of Bedelem cam
to vs with lordis of Jherusalem — And Rama thane beyng turkys —
The great Türke havyng in Dominyon All the holl londe, And in
shorte tyme they concludyd what sume ower patrone should pay for
or tribute. And thanne we war suffered to com on londe.
The same Day at iij of Cloke at aftir noon, we com on londe, And as
we came owt of the boott we war receyvyd by the Turkys and Sarra-
syns, and put in to an old Cave by name and tale, ther Screvener ever
wrytyng ower namys man by man As we entyred in the presens of
the seyd lordis, And ther we lay in the same Grotte or cave all nyght
upon the stynking Stable grounde, as well nyght as Day, ryght evyll
intretyd by the seyd Turkes Mames. (Torkington, V Oldest Diarie of
Englysshe Travel! 23)
 Genre conventions: Personal affect 249

TRAVELOGUE:
From London to Amwell bery /Amwell Bury/ which is in Hartford-
shire 19 mile, where I staid a day or two, thence to Bishopstafford
/Bishops Stortford/ in Essex 13 mile, thence to Dunmew /Dunmow/
8 long miles thro' severall little villages, its very deep way especially
after raines; this is a little Market town, they are altogether taken up
about the spinning and prepareing for the Bayes /baize/: all along
between that and Colchester you pass but hälfe a mile ere one comes
to two or 3 houses all along the road, its from Dunmow to Coll-
chester 22 miles, and mostly clay deep way. (Fiennes, The Journeys
of Celia Fiennes 142)

Some passages of dialogue are inserted in these texts. Mostly such inser-
tions are occasional, but half of Mowntayne's autobiography is written in
dialogue form. An assessment of the interactive passages embedded in
these texts revealed an interesting feature. As the focus is on the I, the
narrator of the travelogue, for instance, the real referent of the pronoun
you may be the / of the text. Thus the movement is not reciprocal but
self-centered:

TRAVELOGUE:
Whereto an English Gentleman replyed, that hee was the next good
voyage after at the Hands: / answered him that / was there also. He
demanded in what ship I was? I tolde him in the Rainebowe of the
Queenes: why (quoth he) doe you not know me? / was in the same
ship, and my name is (Witherington). Sir, said I, I do remember the
name well, but by reason that it is neere two and twenty yeers since
/ saw you, I may well forget the knowledge of you. Well said he, if
you were in that ship, / pray you tell me some remarkable token that
happened in the voyage, whereupon / told him two or three tokens;
which he did know to be true. Nay then, said I, I will tell ... (John
Taylor, Pennyles Pilgrimage 131.C2)

/ and you mix in the same way in Mowntayne so that the prevailing direc-
tion of address is towards the self, though some reciprocity is also found
(see also the examples above):
250 Irma Taavitsainen

AUTOBIOGRAPHY:
Than sayed my lord cheffe justyes unto me, "Syr, whate make yow
here? are you not a Londynar?" "Yes, and yt lyke your lordshyp."
"Howe longe have yow be here prysonar?" "Halve a yeare, my
lorde." "Who sent yow hether?" "Forsothe, my lorde, that dyd the
counsel." (Mowntayne, Narratives of the Days of the Reformation
206)

"you"

Figure 5. Egocentric "you".

 Genre conventions: Personal affect 251

In contrast to fiction, the other text types use the first person singular fre-
quently, independently of the dimension "other", with a solipsistic illu-
sion. In several cases the second person you is mostly a projection to I,
someone addressing the self in an egocentric focus. The contrast to the in-
teractive self in fiction is striking.
The first person singular in Aphra Behn's text is different. There is a
narrative frame and the narrator gives personal comments and guidance
to the reader. The dual role of the narrator and character are dedicated to
authenticating the story, claiming historicity for the events. Behn's work
has therefore been characterised as an imaginary "true history" of travel
narrative (McKeon 1987: 111-113). Direct quotations are not very com-
mon. This work seems to belong to a totally different kind of fiction,
foreshadowing romantic prose. The present analysis does not catch the es-
sence of its style.

9.3. Direct questions and THAT-deletion

Direct questions had a very high factor loading on Factor 1 (0.838), and
they are also significant for the difference between imaginative and non-
imaginative text types. They represent the prototypical form of interaction
and are closely linked with the occurrence of the second person. The quo-
tations in fiction imitate natural speech, and contain features typical of the
spoken level of language such as exclamations, swearing and pragmatic
particles, but reductive features of the surface form are also present as
THAT-deletion proved significant. Besides direct quotations, this feature
occurs in indirect narration in Armin's text. In addition, contracted forms
are found in fiction.

The foole, being wakened, lookes about him; when he had thanked
the woman, asked what newes? sayes the man, Sir, here is your uncle
come out of the country to see you. God a mercy cousin! sayes Will
Sommers; I thank thee for thy labour, you cannot uncle me so. Yes,
truly, sir, I am your own deare uncle, M. William, and with that
wept. Are you my uncle? sayes Will. I, sir, sayes hee. Are you my
uncle? sayes hee againe. I, sure, and verely too. But are you my
uncle, indeed? By my vusse I am, sayes the old man. Then, uncle, by
252 Irma Taavitsainen

my vusse, welcome to court, sayes Will Sommers. But what make

you heere, uncle? He ups and tels his comming to him. Will takes
him by the hand: Come, saies hee, thou shalt see Harry, onckle —
the onely Harry in England; (Armin, A Nest of Ninnies 43—44)

10. Indexical features of proximity

Besides surge and interaction, deictic features proved important. In the

factor analysis both here and now loaded on Factor 1 (factor loadings
0.653 and 0.652 respectively), but only now proved significant in the
second phase (t-test: ρ < 0.031). Demonstrative pronouns did not load on
any factor, but the difference between the use of this in fiction and non-
fiction proved significant in the analysis of variance.
Orientational features of language which locate utterances in relation to
space and time form the deictic patterning in texts (Simpson 1993: 13).
The main types of indexical features are the pairs of adverbs and demon-
strative pronouns here/there, now/then, this/that; personal pronouns are
also indexical, I and you being the basic dichotomy. Other indexical fea-
tures are deictic verbs that express the direction of motion in respect to
the speaker like come/go, bring/take, tense present/past, and even inter-
jections can be included as they derive their meanings purely from the im-
mediate context, at least in spoken language.
In the present assessment, the indexical features of proximity proved
significant. The deictic adverb of present time now and the deictic adverb
or adjective this, indicating physical proximity to the speaker, were impor-
tant. Deictics strongly invoke the point of view in texts (Ehrlich 1990: 96-
97; Toolan 1990: 127); they create orientation, and indicate the way in
which egocentricity and the dimension "self-other" works.
In addition, it has been pointed out that words like this and that include,
besides definiteness, an indication of why the speaker expects the ad-
dressee to be able to identify the referent (Chafe 1975: 39). For example,
the passage below contains a condensed network of this and now occur-
rences which are not restricted to direct speech quotations only.

Thys horsman heryng ye frere thys intrete for hym sayd to oconer
thys Oconer thou seeyst well by thys mannys reporte γ yf I dye now
I am out of charyte & not redy to go to heuen & so it ys y' I am now
 Genre conventions: Personal affect 253

out of charyte in dede but hou seest well y' this frere ys a good man
he is now well dysposyd & in charyte and he is redy to go to heuen
& so am not I therfore I pray the hang vp thys freere whyle that he
hys redy to go to heuyn. This Oconer heryng this mad answere of
hym sparyd the man ... (A Hundred Mery Talys 24)

This device helps to bring the events maximally close to the time of nar-
ration, and create a "proximal" style (Simpson 1993: 18). Together these
features have the function of calling for a more immediate response in the
audience. In the same way as the readers are expected to feel the emotions
of excitement, sorrow, hesitation, surprise or horror expressed by surge
features like interjections, swearing and pragmatic particles, they are guid-
ed by indexical features of proximity. Narration is made more vivid, and
the immediacy of experience is emphasised.
Armin's text uses deictic elements in the same way as interjections, to
shift the viewpoint (see above); come and go in the present and past
tenses alternate in succession. The quick change from proximal to distal
contributes to the special effect of reversal:

But they come; and up they came, and to the king they goe, who,
being with the lord treasurer alone, merry, seeing them two, how ...
(Armin, A Nest of Ninnies 44)

In non-literary texts deictic expressions are fewer and serve to make the
directions explicit in a matter-of-fact way. For example, the text strategies
found in the travelogues of the HC are explicitly iconic (Virtanen 1995).
In the description of the scenery below with neutral expressions of the
points of compass and existential sentences:

TRAVELOGUE:
A mile a this side Akeland Castelle I cam over a bridg of one great
arch on Gaundelesse a praty ryver rising a vj. miles of by west: and
renning by the south side of Akeland Castelle goith a litle beneth it
to the great streame of Were.
Gaundeles rising by west cummith to Westakeland, by S. Helenes
Akeland, by S. Andreas Akeland, and by Bisshop Akelande.
The towne self of Akeland is of no estimation, yet is ther a praty
market of corne.
It standith on a praty hille bytween 2. ryvers, wherof Were lyith on
the north side, and Gaundelesse on the south, and an arow shot or
254 Irma Taavitsainen

more benethe they meete and make one streame, and ren to the este.
And ech of these rivers hath an hille by it, so that Bisshop Castelle
Akeland standith on a litle hille bytwixt 2. great [hills]. (Leland, The
Itinerary of John Leland 70)

11. Modal auxiliaries

Modal auxiliaries and adverbs, evaluative adverbs, generic sentences,

verbs of knowledge, prediction and evaluation, together with deictic el-
ements, contribute to the expression of point of view and degrees of com-
mitment in texts (Simpson 1993: 47). Modal auxiliaries were included in
the present assessment; for adverbs and adjectives, see Meurman-Solin in
this volume. Verbs expressing volition and possibility proved significant
in t-tests (t-tests: modals of volition ρ < 0.040; modals of possibility ρ <
0.030). In factor analysis, modals expressing volition correlated with Fac-
tor 1 (factor loading 0.693), whereas modals of possibility and necessity
correlated with Factor 2 (factor loadings 0.720 and 0.692). These auxilia-
ries mark the reasoning process together with private verbs of mental ac-
tivity, and this factor describes philosophical and educational writings.
Boulomaic modality, expressing "desire" and futurity, seems to be a
component of the fictional world in the present selection of texts. In Pen-
ny Merriments the use of modals of volition is especially prominent, and
they act as fictional space-builders of the make-believe world. The con-
trast to the factual, matter-of-fact narration is striking as the examples
below demonstrate. The examples given below represent the far ends of
the scale. More studies are needed to define the quality and shades of mo-
dality in these texts.

FICTION:
(Jo) If my master will let me have the house we will brue good Ale,
and we will have mault of our own, for we'l keep a stock going in
my Master's Mault-house, with his leave, and there is pasture enough
to keep two beasts, and conveniences for hogs and poultery, so thou
mayest have all things about thee, and keep a maid and live like a
Lady.
 Genre conventions: Personal affect 255

(Ka) This will be brave indeed John, but what shall we do with our
Ale.
(Jo) Sell it my sweet one; Let me see, there are eight rooms in the
house besides the Cellar, and with a little painting and a few benches
it will be very fine, & a handsome sign to draw in company.
(Ka) What shall that be.
(Jo) The three fair maids, I think.
(Ka) Not for a hundred pound I would not have such a sign.
(Jo) Why prithee.
(Ka) Why man they'd think surely we kept a bawdy house ...
(Penny Merriments 117)

DIARY:
The xxvij day of July the duke of Suffoke, maister [Cheke] the
kynges scolmaster, maister Coke, [and] ser John Yorke, to the Towre.
The xxxj day of July was delevered owt of the Towre the duke of
Suffoke; and the sam day rod thrugh London my lade Elssabeth to
Algatt, and so to the qwens grace her sester, with a Ml hors with a
C. velvett cotes.
The sam tyme cam to the Flett the yerle of Ruttland and my lord
Russell, in hold. The qwen['s] grace mad [sir Thomas] Jarnyngham
vyce-chamburlayn and captayne of the garde, and ser Edward Hast-
yngs her grace mad ym the maister of horsse the same tym (Machyn,
The Diary of Henry Machyn 38)

12. Conclusions

The aim of this study was to test whether any differences could be found
between literary and non-literary text types. Another aim was to achieve
a more detailed description of the qualities of personal affect. For these
purposes various methods were combined. Factor analysis was used first
to establish the affinities of texts and to detect matrices of subjective uses
of language, reflecting the underlying dimensions of variation. Further
statistical methods were then employed to verify possible text-type mark-
ers among the linguistic features that could prove important for various
qualities of personal affect in fiction.
256 Irma Taavitsainen

Four different qualities of personal affect were revealed by this study.

They are all characteristic of early fiction: surge features in the form of
emotional outbursts, an interactive dichotomy between self and other, in-
dexical features of proximity, and boulomaic modality. As the ultimate
aim of the study was to chart genre styles and genre conventions, some
of these features were assessed in a longer diachronic perspective.
The quality of affect in the fiction of the Early Modern section of the
HC is veiy different from the kinds of affect I defined in my earlier
studies on religious and scientific prose. The repertoires of affective/emo-
tive features in text types vary greatly and there are differences from text
to text. Some general trends of evolution can be discerned, but further
studies are needed for more specific results.

12.1. "Surge" of personal affect

Of the surge features, the most important are exclamations. Their use
seems to be a deliberate device of story-telling. Emotional involvement is
often prominent especially in fiction, and echoes of earlier writings in the
Middle English fabliau tradition can be discerned; genre conventions are
an important factor in the frequent occurrence of interjections in written
texts. Chaucer uses them in marking turning-points of the plot and fore-
grounding to create awe and apprehension in the audience. The Late Mid-
dle English conventions are continued in a somewhat modified form in
Early Modern short fiction, and even later. Swearing, oaths, and prag-
matic particles are also important as surge features of affect, and the line
of continuation from Middle English is evident. The last sample of fiction
from 1706 belongs to another era and another tradition; interjections are
found in it as well, but they are used in a different way. Thus the conven-
tion continues, although it finds new forms and applications.
A possible surge feature available in the language at the time in focus
was the use of thou. Statistically it proved significant, but a closer con-
sideration revealed that the frequent use of thou in fiction is connected
with the social classes depicted in these texts: country folk and lower and
middle class people among which it was the unmarked pronoun of ad-
dress. Although several texts clearly contain an emotional loading, it can-
not be attributed to this linguistic feature alone; together with others, the
feature may contribute to the affective tone.
In Armin's text and in Penny Merriments the use of thou is marked and
serves special purposes: reversal of values in carnivalism, and parody of
 Genre conventions: Personal affect 257

religious language and Quaker usage.These conclusions were made on the

basis of qualitative assessment and backgound knowledge of the socio-
historical context of these texts.
In non-literary texts such as autobiography the use of thou to express
heightened emotionality is evident. Some instances are found in outbursts
of accusation and contempt, and in one example it is used in a solemn and
affective speech.

12.2. Interactive focus

Another salient feature contributing to the difference between fiction and

the adjoining text types is the reference system of the pronouns. The di-
chotomy between self and other proved important in fiction, whereas the
non-imaginative narrative texts have a more self-centered focus. An inter-
active, reciprocal system functions in these texts with questions and an-
swers, expressions of emotions and responses to the other partner's words
(cf. Factor 1). In contrast, the other text types show a system in which you
is often coreferential with I, and the egocentric, solipsistic focus is domi-
nant.

12.3. Immediate context and space-building modality

Indexical features of proximity are characteristic of fiction. They indicate

that the action is simultaneous with the narration and that the things re-
ferred to are close at hand. Special elements like modal auxiliaries of a
certain kind help to build up the fictional world in these texts.
The above-mentioned linguistic features are mostly found in direct
speech quotations that imitate natural speech, and perhaps condense some
of its features. It has also been suggested that interjections should be re-
garded as deictic elements because their interpretation is totally context-
bound. This is not entirely true of written texts, as textual functions have
also developed.
Linguistic features reflecting personal affect may act as text-type mark-
ers in some genre and period styles; exclamations are especially distinctive
in this respect. These features are essential in reader involvement. They
function in various ways. Readers may be asked to pay attention to certain
points by conative interjections or collocations. Proximal deictic expres-
sions help to bring the events and characters to the immediate experience
258 Irma Taavitsainen

of the reader. Indirectly they too contribute to reader involvement. Per-

sonal affect in the form of charge between participants may be found both
text-internally between characters in the text or involving the readership.

13. Evaluation of the methods

There is a great deal of variation both in volume and form of personal af-
fect. Factor analysis revealed textual affinities and helped to focus the as-
sessment to texts that stand closest to fiction. Factor analysis also in-
dicated potential text-type markers, and its matrix of interactive features
a feasible way of interpreting the final results.
Additional statistical methods were used to verify the implications of the
hierarchical ordering of linguistic features marking subjectivity. T-tests
and f-tests proved the significance of these features in telling fictional
works apart from the adjoining non-imaginative texts. Thus it was possible
to pinpoint text-type markers and achieve empirical evidence of genre
conventions.
This study showed that it is necessary to complement quantitative meth-
ods by qualitative analysis in order to achieve reliable results. Immediate
context is not sufficient, but the books have to be considered in all and
their sociohistorical connections must be taken into account; irony, mock-
ing, and parody, or reversal of values in carnivalism, can only be detected
by careful reading of the texts; only qualitative stylistic analysis can reveal
the nature of involvement and the more subtle shades of meaning in these
texts.

Notes

I am grateful to Professor Edward Finegan for comments and suggestions for im-
provement on an earlier version of this study.

1. Cf. Werlich's classification (1983: 38-41) into the basic text types of de-
scription, narration, exposition, argumentation, and instruction. They are de-
fined according to internal linguistic features.
 Genre conventions: Personal affect 259

2. Guillen also points out the importance of the Renaissance letter to the
development of the early novel in terms of shared functions (1986: 99-100).
3. Görlach emphasises the role of convention in the recognition of certain
well-defined and standardised uses of language as text types in a specific
culture (1991: 200-201). Some recent studies concentrate on very narrow
text types like wedding reports in newspapers, which are then discussed in
great detail, e.g. Suter (1993).
4. "Noble chivalry" in the same way that kitchen furniture may be marketed
as "sophisticated elegance" (Margolies 1985: 36). For a linguistic assess-
ment of emotional features in romances, see Taavitsainen (forthcoming d).
5. Biber—Finegan (1989b: 94) define personal affect as a wide range of per-
sonal attitudes, emotions, feelings, moods and general dispositions, and
place it under the title of "stance" together with evidentiality. According to
Halliday (1978: 33) the relationship between participants includes "the de-
gree of emotional charge". With status and contact it forms the third com-
ponent of the participant relation, but unlike the other two, affect is not
always present: it can be turned on or off, and the "volume" may be ad-
justed (Martin 1992: 533). (For a discussion, see Taavitsainen 1994b). Re-
cent discussions have largely concentrated on the anthropological study of
emotions and discourse (see Lutz—Abu-Lughod 1990) and gender studies
(Poynton 1989). Lexical studies of emotions have also received attention
(e.g. the first part of Busse 1992 is called "Linguistic aspects of emotions",
and it contains several articles on lexis).
6. Of the non-fictional texts, Mowntayne has the highest proportion of direct
speech: 2899 words versus 2842 words of narration, total 5740. No other
text comes near it.
7. The underlying dimensions were interpreted to reflect (1) involvement
versus informational focus, (2) narrative discourse, (3) endophoric versus
exophoric reference, (4) persuasion, (5) abstract and technical discourse, (6)
informational elaboration under strict real-time constraints, and (7) academic
qualification or hedging (Biber 1988: 101-120).
8. For example, present and past tense were also included in the primary
analysis but excluded from the final one. Their communalities in the factor
analysis were low, which showed that they did not really contribute to any
factor; see note 13. It is possible to rotate the factors to simplify the pattern
and make the distinctions stand out more clearly; thus in a rotated solution
each factor is characterised by a few features that are most representative.
The extent to which the clusters of features distinguish among different sets
of texts varies, and the factors are given eigenvalues that show their impor-
tance in explaining the amount of underlying variation.
260 Irma Taavitsainen

I acknowledge the expertise of Ms Leena Sadeniemi at the Computing

Centre of the University of Helsinki in the statistical analyses carried out for
this study.
9. I shall turn to this grouping and the special quality of personal affect in
philosophy and education in a future study.
10. It is, of course, evident that the small number of texts both in fiction and
adjoining genres is not adequate for definitive conclusions, but this material
may give indications of whether the question is appropriate and how to find
a possible answer to it.
11. The corpus parameters "interactive versus x" and "informal versus x" are
important for the present assessment. They are coded in the corpus, but they
have not been systematically applied in research. I wanted to test their
influence on the above factors. T-tests were used for this end. Factor scores
were counted for each text, and the group means of scores were compared
with one another. The results show that the parameter "interactive versus x"
is very significant in Factor 1 (p < 0. 0005), significant in Factor 2 (p <
0.026), but insignificant in Factor 3 (p < 0.058). Likewise, the parameter
"informal versus x" was significant in Factor 1 (p < 0.001), in Factor 2 (p
< 0.013), but insignificant in Factor 3 (p < 0.677).
12. The parameter NI versus NN coincides with the present focus on the differ-
ence between fiction and the adjoining genres because plays are excluded
from this phase of my study. In Late Middle English, Mandeville 's Travels
belongs to imaginative narration although it is a travelogue.
13. Not all features proved significant with respect to the difference between
imaginative versus non-imaginative narration. Non-significant features were
e.g. so, then, that, here, and past tense versus present tense.
14. My hypothesis was that exclamations act as a marker of the fictional mode.
To test my hypothesis I divided the material into two categories: the fic-
tional and the non-fictional, the former group comprising the material from
The Canterbury Tales, Gower, Caxton, romances and drama. A t-test was
performed and the results proved that the difference in the use of exclama-
tions in the two groups is highly significant. To shift the angle, another t-
test was performed with the texts by well-known literary authors, Chaucer,
Gower, Caxton and Malory, on the one side, and the anonymous texts and
non-literary authors, including the mystical authors of religious treatises, on
the other side; drama was excluded from this t-test. The difference was
highly significant in this case as well (Taavitsainen forthcoming a).
15. In Chaucer's fabliaux the collocations of interjections are a regular feature
in highlighting the stories and bringing turning-points of the plot to the
audience's attention: The Miller's Tale provides a case in point. For com-
parison, the same story in a prose version included in "Jokes and Jests" in
the Penny Merryments contains only one interjection, in the only direct quo-
tation.
 Genre conventions: Personal affect 261

16. An example of foregrounding can be found in the Squire's Tale:

So ferde this tercelet, alias the day!

Though he were gentil born, and fressh and gay,
And goodlich for to seen, and humble and free,
He saugh upon a tyme a kyte flee,
And sodeynly he loved this kyte so
That al his love is clene fro me ago,
.... withouten remedie! (621-629)
17. This is the mechanism of involvement in Middle English mystical writings.
It is most frequent in Julian of Norwich's treatise and Richard Rolle uses
it in his affective works (see Taavitsainen 1993).
18. A change in the pronouns is used as a dramatic effect in one speech act
which reflects a change of attitude when a scholar announces the end of his
loyalty to the tyrant (lines 806-810). The passage has a strong emotional
loading.
19. A typical second-person oriented text guides the reader in phrases like you
shall see/iudge/marke/ you must vnderstand/marke/ you may see/perceave,
etc. The interpersonal aspect is often present in phrases like I will show you
(Taavitsainen 1994a).

References

Adamson, Sylvia
1994 "Subjectivity in narration: Empathy and echo", Subjecthood and sub-
jectivity. Paris—London: Ophrys, 193-208.
Ameka, Felix
1992 "Interjections: The universal yet neglected part of speech", Journal
of Pragmatics 18(2/3): 101-118.
Benveniste, Emile
1966 Problems in general linguistics. Trans, by Mary Elizabeth Meek.
[1971] Coral Cables, Flo.; University of Miami Press.
Biber, Douglas
1986 "Spoken and written textual dimensions in English: Resolving the
contradictory findings", Language 62: 384—414.
1988 Variation across speech and writing. Cambridge: Cambridge Uni-
versity Press.
1989 "A typology of English texts", Linguistics 27:3-43.
Biber, Douglas—Edward Finegan
1986 "An initial typology of English text types", in: Jan Aarts—Willem
Meijs (eds.), Corpus linguistics II. Amsterdam: Rodopi, 19-46.
262 Irma Taavitsainen

1989a "Drift and the evolution of English style: A history of three genres",
Language 65/3: 487-517.
1989b "Styles of stance in English: Lexical and grammatical marking of
evidentiality and affect", Text 9-1: 93-124.
Blake, Norman F.
1992 " Why and what in Shakespeare", in: Toshiyaki Takamiya—Richard
Beadle (eds.), Chaucer to Shakespeare: Essays in Honour of Shin-
suke Ando. Cambridge: D. S. Brewer, 179-193.
Busse, Wilhelm G. (ed.).
1992 Anglistentag 1991: Proceedings, Düsseldorf. Tübingen: Max Nie-
meyer.
Calvo, Clara
1992 "Pronouns of address and social negotiation in As you like it", Lan-
guage and Literature 1/1: 5—27.
Carter, Ronald—Walter Nash
1990 Seeing through language: A guide to styles of English writing. Ox-
ford—Cambridge, Mass.: Blackwell.
Chafe, Wallace L.
1975 "Givenness, contrastiveness, deflniteness, subjects, topics and point
of view", in: Charles N. Li (ed.), Subject and topic. New York—San
Fransisco—London: Academic Press, 25-55.
1985 "Linguistic differences produced by differences between speaking
and writing", in: D. R. Olson—N. Torrance—A. Hildyard (eds.), Lit-
erature, language and learning: The nature and consequences of
reading and writing. Cambridge: Cambridge University Press, 105-
123.
Davis, Walter R.
1969 Idea and act in Elizabethan fiction. Princeton, New Jersey: Princeton
University Press.
Ehrlich, Susan
1990 Point of view: A linguistic analysis of literary style. London—New
York: Routledge.
Fludernik, Monika
1993 The fictions of language and the languages offiction. London—New
York: Routledge.
1995 "Pronouns of address and 'odd' third person forms: The mechanics
of involvement in fiction", in: Keith Green (ed.), New essays on
deixis?, 99-129.
Fowler, Alastair
1982 Kinds of literature: An introduction to the theory of genres and
modes. Oxford: Clarendon Press.
Frijda, Nico Η.
1986 The emotions. Cambridge: Cambridge University Press.
 Genre conventions: Personal affect 263

Görlach, Manfred
1991 "Text types and the linguistic history of Modern English", in: Claus
Uhlig—Rüdiger Zimmermann (eds.), Proceedings of the Conference
of the German Association of University Professors of English XII.
Tübingen: Max Niemeyer, 195-215.
Guillen, Claudio
1986 "Notes toward the study of the Renaissance letter", in: Lewalski
(ed.), 70-101.
Halliday, Μ. A. K.
1978 Language as social semiotic: The social interpretation of language
and meaning. London—New York—Melbourne, Auckland: Edward
Arnold.
1989 Spoken and written language. Oxford: Oxford University Press.
Hart, Francis R.
1974 "Notes for an anatomy of modern autobiography", in: Ralph Cohen
(ed.), New directions of literary history. London: Routledge and
Kegan, 221-247.
HC = Helsinki Corpus
1991 The Helsinki Corpus of English Texts. Helsinki: Department of Eng-
lish, University of Helsinki.
Heine, Berd—Ulrike Claudi—Frederike Hünnemeyer
1991 Grammaticalization: A conceptual framework. Chicago—London:
The University of Chicago Press.
Hughes, Geoffrey
1991 Swearing: A social history of foul language, oaths and profanity in
English. Oxford—Cambridge, Mass.: Blackwell.
Imbrie, Ann E.
1986 "Defining nonfiction genres", in: Lewalski, 45-69.
Jucker, Andreas H. (ed.)
1995 Historical pragmatics. Amsterdam—Philadelphia: John Benjamins.
Jucker, Andreas H.
forthc. "The origin of the discourse marker well: A case for historical prag-
matics".
Kytö, Merja (comp.)
1996 Manual to the diachronic part of the Helsinki Corpus of English
Texts: Coding conventions and lists of source texts. (3rd edition.)
Helsinki: Department of English, University of Helsinki.
Kytö, Merja
forthc. "Robert Keayne's notebooks: A verbatim record of spoken English in
early Boston?", in: Susan Herring—Pieter van Reenen—Lene Sches-
ler (eds.), Textual parameters in older languages.
264 Irma Taavitsainen

Leech, Geoffrey Ν.—Michael Η. Short

1981 Style in fiction: A linguistic introduction to English fictional prose.
London—New York: Longman.
Lewalski, Barbara Kiefer (ed.).
1986 Renaissance genres: Essays on theory, history and interpretation.
Cambridge, Mass.—London, England: Harvard University Press.
Lutz, Catherine A.—Lila Abu-Lughod
1990 Language and the politics of emotion. Cambridge: Cambridge Uni-
versity Press.
Margolies, David
1985 Novel and society in Elizabethan England. London— Sydney: Croom
Helm.
Martin, J. R.
1992 English text: System and structure. Philadelphia—Amsterdam: John
Benjamins Philadelphia.
McKeon, Michael
1987 The origins of the English novel 1600-1740. Baltimore—London:
The Johns Hopkins University Press.
Meurman-Solin, Anneli
1993a "Involvement markers in the characterization of early prose genres",
in: Anneli Meurman-Solin, Variation and change in early Scottish
prose: Studies based on the Helsinki Corpus of Older Scots. (Huma-
narum Litterarum 65 / Annales Academiae Scientiarum Fennicae.
Diss.) Helsinki: Suomalainen Tiedeakatemia, 276-312.
1993b "The author-addressee relationship and the marking of stance in the
characterization of sixteenth- and seventeenth-century genre styles",
Revue Beige de philologie et d'histoire, Fasc. 3: Langues et litte-
ratures modernes, 733-745.
Mühlhäusler, Peter—Rom Harre
1990 Pronouns and people : The linguistic construction of social and per-
sonal identity. Oxford: Blackwell.
Nevanlinna, Saara—Irma Taavitsainen
1993 St Katherine of Alexandria: The Late Middle English prose legend in
Southwell Minster MS 7. Cambridge—Helsinki: Boydell—Brewer—
The Finnish Academy of Science and Letters.
Poynton, C.
1989 Language and gender: Making the difference. Geelong, Vic.: Deakin
University Press.
[1985] Oxford: Oxford University Press.
Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik
1985 A comprehensive grammar of the English language. London: Long-
man.
 Genre conventions: Personal affect 265

Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.)

1993 Early English in the computer age: Explorations through the Helsinki
Corpus. Berlin—New York: Mouton de Gruyter.
Sampson, George
1970 The concise Cambridge history of English literature. (3rd revised
edition.) Cambridge: Cambridge University Press.
Simpson, Paul
1993 Language, ideology and point of view. London—New York: Rout-
ledge.
Spufford, Margaret
1981 Small books and pleasant histories: Popular fiction and its reader-
ship in seventeenth-century England. Cambridge: Cambridge Uni-
versity Press.
Suter, Hans-Jürgen
1993 The wedding report: A prototypical approach to the study of tradi-
tional text types. Amsterdam—Philadelphia: John Benjamins.
Taavitsainen, Irma
1988 Middle English lunaries: A study of the genre. (Mdmoires de la
Societe Ndophilologique de Helsinki XLVII. Diss.) Helsinki: Societ6
N6ophilologique de Helsinki.
1993 "Genre/subgenre styles in Late Middle English?", in: Rissanen et al.,
171-200.
1994a "On the evolution of scientific writings between 1375 and 1675:
Repertoire of emotive features", in: F. Fernändez et al (eds.), Papers
from the 7th International Conference of English Historical Linguis-
tics, Valencia, Sept. 1992. Philadelphia—Amsterdam: John Benja-
mins, 329-342.
1994b "Subjectivity as a text-type marker in historical stylistics", Language
and Literature 3/3: 197-212.
1995a "Interjections in Early Modern English: From imitations of spoken to
conventions of written language", in: Andreas H. Jucker (ed.), 419-
445.
1995b "Narrative patterns of affect in four genres of the Canterbury Tales",
The Chaucer Review 30/2: 82-101.
forthc. a "Exclamations in Late Middle English", in: Jacek Fisiak (ed.),
Studies in Middle English. Berlin—New York: Mouton de Gruyter.
forthc. b "By Saint Tanne: Pious oaths or swearing in Middle English? An
assessment of genres", in: Raymond Hickey—Stanislav Puppel (eds.),
Festschrift in honor of Jacek Fisiak's 60th birthday. Berlin—New
York: Mouton de Gruyter.
forthc. c "Genres and text types in Medieval and Renaissance English", to
appear in Poetica: An International Journal of Linguistic-Literary
Studies 47: 49-62 (Tokyo, Japan).
266 Irma Taavitsainen

forthc. d "Emphatic language and Romantic Prose: Changing functions of in-

terjections in a sociocultural perspective", European Journal of Eng-
lish Studies. Special Issue. Language theory and practice in current
literary scholarship, ed. by Monika Fludernik.
Todorov, Tzvetan
1990 Genres in discourse. Cambridge: Cambridge University Press.
Toolan, Michael J.
1988 Narrative: A critical linguistic introduction. London—New York:
Routledge.
1990 The stylistics of fiction: A literary-linguistic approach. London—New
York: Routledge.
Virtanen, Tuija
1995 '"Then I saw to antique heddes': Discourse strategies in Early Mod-
ern English travelogues", in: Andreas H. Jucker (ed.). Amsterdam—
Philadelphia: John Benjamins, 499-513.
Wales, Katie
1983 "'Thou' and 'you' in Early Modern English: Brown and Gilman re-
appraised", Studia Linguistica 37/2: 107-125.
Werlich, Egon
1983 A text grammar of English. Heidelberg: Quelle and Meyer.
Towards reconstructing a grammar of point of
view: Textual roles of adjectives and open-class
adverbs in Early Modern English

Anneli Meurman-Solin

1. Introduction

In this study the general approach to the concept of point of view in texts
is based on recent developments in corpus-based sociohistorical linguis-
tics. A text as a linguistic construct is affected by extralinguistic features
in various complex ways, and this is reflected in variation and change
over time. For example, besides their text-organizing function, linguistic
expressions of point of view can be meaningfully related to the author-
addressee relationship (Meurman-Solin 1993b). Studies of features such
as the social functions of texts, the social roles and social networks of
their authors and addressees, the general narrative framework of texts and
the varying narrator roles in them, provide information that may explain
linguistic choices on all levels of language use. In the present study the
hypothesis is that the above-mentioned extralinguistic features condition
textual participant roles. A closer examination of how these roles are ex-
plicitly or implicitly expressed in texts may lead not only to the identi-
fication of a text's ideology (Simpson 1993: 5) but also to a better under-
standing of processes towards the establishment of extra-linguistically
conditioned linguistic conventions and norms. Additionally, it may shed
light on processes towards diversification in the development of genres.
It is the latter of these aims, the presumably multidirectional diachronic
development of genres, that the present study sets out to explore further.
In more specific terms, the more immediate aim of the study is to provide
material for the reconstruction of a grammar of point of view, and to
identify genre-specific features and genre styles in such a grammar in a
268 Anneli Meurman-Solin

wide range of Early Modern English texts. 1 A further aim is to increase

our knowledge of how the social and communicative functions of texts
are reflected in linguistic choices and in patterns of information process-
ing resulting in various degrees of syntactic complexity.
To achieve the aims defined above, the analysis of genres is viewed as
a multi-disciplinary activity (cf. Bhatia 1993: 1). Genre analysis discusses
the more or less 'conventionalized communicative purposes' (Bhatia
1993: 1) of genres and the ways in which such purposes link with social
and cultural factors, and, as Bhatia (1993: 1-2) succinctly puts it, it shares
approaches applied in 'discourse analyses of institutionalized use of lan-
guage in socio-cultural settings with a heavy emphasis on communication
as social action'. To use Hallidayan terminology (Halliday—Hasan 1985
[1990]: 12), the analysis of point of view is here mainly concerned with
the tenor of discourse, but the use of particular linguistic features is
linked with the field and the mode of discourse by stressing 'the nature of
the social action that is taking place' and the way in which explicit lin-
guistic devices are used to organize texts in terms of text categories (such
as non-imaginative narration, argumentative and expository texts).
I became interested in the social and textual roles of authors and their
readers when trying to identify extralinguistic factors conditioning lin-
guistic variation and change in the history of Scottish English, a minority
variety where these roles could be assumed to be affected by how that
variety related to a majority variety under various social and political
pressures. The general assumption was that by describing what was
known about the individual authors, their social status and education, their
social networks, and, on the other hand, about the audiences and ad-
dressees of their texts, it would be possible to identify factors condition-
ing linguistic variation and change, particularly as regards language stand-
ardization, in Scotland the so-called anglicization of Scottish English.
Meurman-Solin (1993a) shows that the most important conditioning
factors in the standardization of Scots are extralinguistic variables such as
dialect, i.e. dialects within Scots, whether the text was printed or not, 2
genre, text category, the author's sex, and, in letters, the participant rela-
tionship. Perhaps predictably, the conditioning of the variable 'genre'
proved to be more difficult to trace, partly because of problems inherent
in compiling a diachronic corpus, where scarcity of texts in earlier time
periods may sometimes fatally decrease its representativeness, and partly
because of the innate heterogeneity of certain genres or overlappings
between them. 3 To suggest solutions to some of these problems, the pres-
ent study provides more information about texts as representatives of
 Reconstructing a grammar of point of view 269

genres and text categories, and about similarities and differences between
them. The study isolates a number of linguistic features reflecting point
of view in texts, and it attempts to interpret their frequencies and dis-
tributions in terms of how they relate to other linguistic and extra-linguis-
tic features of the texts and how they compare with findings in other re-
cently published genre studies (see especially Taavitsainen 1993, 1994
and her chapter in the present volume). Given the special challenges of a
diachronic perspective and the innate problems related to a semantic
analysis of large amounts of data, further study will be required to deal in
more detail with some of the findings presented here.
At present we do not know exactly what the normal, basic, least mark-
ed or canonical (see Enkvist 1988: 138) features of prototypical repre-
sentatives of genres are in a given context at a given time (cf. Enkvist
1988: 127), and therefore both the description of one particular genre and
comparisons between genres necessarily remain tentative. Moreover, the
traditional classification of texts, a framework used as a practical tool also
in the Helsinki Corpus, tends to rigidify the interpretation of findings and
their correlation patterns with extralinguistic factors. Therefore, instead of
restricting the focus to similarities and differences between genres, prior-
ity is always given to properties of individual texts.
First and foremost, my study tries to identify features that could be
claimed to be text-specific stylistic properties from a synchronic point of
view. An attempt is then made to suggest what features can be identified
as period-bound or petrified generic conventions and what diachronic de-
velopments features and their co-occurrence patterns undergo. It is plau-
sible that the conservative or innovative pressures attested can be related
to changes in the social functions of texts and the rise of new ideals and
ways of thinking. 4 Ultimately, it may be possible to identify not only dia-
gnostic features in individual texts but also features clustering in particu-
lar text types, and to build a network of overlappings and cross-cuttings
between them resulting in a scalar or multi-dimensional typology.
A text is viewed as a network of points of view expressed by markers
of attitude such as private verbs, modal auxiliaries, adverbs and adjectives
in the roles of modality and degree, expressions conveying evaluations,
and so on. In my terminology, point of view is directly related to partici-
pant voices in a text, whereas involvement refers to a functional notion in
a set of textual dimensions: 'integration' versus 'fragmentation', 'involve-
ment' versus 'detachment' (Chafe 1982, 1985). As these four notions are
also directly related to degrees of informational density and syntactic
complexity, various aspects of genre styles can be illuminated just by
270 Anneli Meurman-Solin

considering the frequencies and distributions of adverbs and adjectives

and their position and function in sentences.
As comparative material, studies based on large corpora, those using a
multi-feature approach and those relating linguistic evidence to socio-
historically relevant extralinguistic evidence, are particularly significant
here. Degree of involvement as opposed to informational density has been
shown to be relevant in describing genre styles (Biber 1988, Biber—
Finegan 1992), and especially frequencies and distributions of markers of
stance have given interesting evidence for suggesting a typology of texts
(Biber—Finegan 1989). In the Helsinki Scots Corpus extralinguistic fac-
tors such as parameters describing the author or the audience of a text
shed light on the general function of genres in their sociocultural context
(Meurman-Solin 1993a and 1993b), and also on the more pragmatic needs
reflected in linguistic choices in language contact situations. In addition
to these and other studies based on computer-readable corpora, the pres-
ent study has been motivated by ideas supporting the importance of point
of view as a text-organizing agent (for example Simpson 1993) and, on
the other hand, by evidence of grammaticalization processes related to ex-
pressions of attitude (for example Hopper—Traugott 1993).

2. The text corpus and methodological considerations

A corpus of approximately 550,000 words of running text was used,

available in the Early Modern English section of the Helsinki Corpus of
English Texts. The texts, altogether 79, represent sixteen genres, each
type usually illustrated by two texts per each one of the three time pe-
riods: 1500-1570, 1570-1640, and 1640-1710. The genres are: law, hand-
books, science, educational treatises, philosophy, sermons, trial proceed-
ings, history, travelogue, diaries, biography, fiction, comedies, private
letters, official letters and the Bible. Scientific treatises are subdivided
into 'science, medicine' and 'science, other', biographies into 'auto-bio-
graphy' and 'biography, other', and the Bible contains extracts from both
the Old Testament and the New Testament. Private and official letters
each contain letters by about ten different individual writers per period.
As was stated above (p. 268), because of the serious limitations of the
material the findings per genre can only be very preliminary, but the cor-
pus can successfully identify similarities and differences between texts
 Reconstructing a grammar of point of view 271

that serve to regroup them in various ways. It is shown later (see for ex-
ample Tables 1 and 8) that, in fact, several texts share practices that
confirm the conventional typological labels coded into the corpus, where-
as others, as a result of their idiosyncratic practices, do not allow any
straight-forward comparisons within the corpus. In addition to date and
genre, the following extralinguistic parameters coded into the Helsinki
Corpus are taken into consideration in the statistics: relationship to spoken
language, sex and social rank of author, audience description, interaction,
setting and prototypical text category.

3. Adverbs

An analysis of all open-class adverbs ending in -ly and variants was done
for each text (for information on some high-frequency closed-class ad-
verbs, see note 14). It may seem that, apart from practical considerations,
there is no valid justification for restricting the material to open-class ad-
verbs. Yet the exclusion of structural alternatives makes it possible to
trace changes from adverbs integrated in sentence structure to those func-
tioning as sentence adverbs (disjuncts), so that formal features do not
influence the diversification or restriction of their function and meaning.
Moreover, as the position of an adverbial — specified in terms of initial,
medial and final — in a sentence may be relevant in the assessment of its
semantic role and syntactic function (see pp. 277-284), only structural re-
alizations allowing all these positions provide useful evidence for study-
ing diachronic variation and change. In the present study, both the posi-
tion within the sentence structure and the position in the text structure
were considered. A difference was made between sentence-initial dis-
juncts and subject-oriented item subjuncts (see Quirk et al. 1985: 572-
578). Despite the relative infrequency of relevant examples, an attempt
was made to trace developments from medially or finally positioned ad-
verbials integrated in the sentence structure to syntactically peripheral dis-
juncts, typically positioned at the beginning or separated by means of
punctuation (cf. Swan 1988). Significantly, such patterns of variation and
change can be related to grammatical ization processes from propositional
via textual to expressive meaning (Traugott 1982: 257, 1990: 497). In my
further studies, I will look at whether the development of textual and ex-
pressive meanings and sentence adverbial functions can be traced in terms
272 Anneli Meurman-Solin

of how the changes are located in specific genre styles or text types, the
ultimate aim being the identification of subjectifying genre styles.
Altogether 3212 open-class adverbs were analysed. Despite the relative-
ly high total number of instances, the breakdown into subcategories often
results in small numbers per subcategory in individual texts, which some-
times weakens the statistical significance of the results. This problem can-
not be avoided in studies of lexical items in corpora consisting of rela-
tively short text extracts (ranging from 5,000 to 15,000 words). The se-
mantic developments of individual adverbs were taken into account in the
categorization of occurrences (see section 6). This categorization was
based on an analysis of each occurrence in a context long enough to al-
low the interpretation of function and meaning. 5 The problem of homo-
morphy (see for example Quirk et al. 1985: 70-75) was thus solved by
means of a manual classification of occurrences. There were approxima-
tely 1,400 evaluative, 450 modal, 450 time, 400 descriptive, 300 intensi-
fying and 200 focusing adverbs.
The analysis of adverbial scope and orientation was based not only on
positional criteria but also on textual and semantic factors. As it was as-
sumed that the use of evaluative expressions realised by means of adverbs
significantly correlated with verb category, the semantic category of the
verb was defined in each sentence where an adverb expressing evaluation
had been attested. The verb categories used as labels in the classification
were as follows: verbs of action, speech act verbs, verbs of inert percep-
tion and cognition, verbs acting as metalinguistic comments, verbs of vo-
lition, copular and relational verbs. It turned out, however, that only the
first two were sufficiently frequent to allow a more detailed study; a sta-
tistical analysis was performed only of verbs of action (see section 9). In-
formation about whether the verb was used transitively or intransitively,
in the active or in the passive voice was also recorded.

4. Adjectives

Because of the high frequency of adjectives, a random sample of 15,000

adjectives in the WordCruncher word list was analysed; in addition,
adjectives derived from numerals and own and whole were ignored in the
 Reconstructing a grammar of point of view 273

statistics. Information about the syntactic functions (prenominal modifier,

postnominal modifier, subject complement, object complement) was re-
corded, ignoring, however, most participles.6 No distinction was made
between adjectives on the basis of their morphological characteristics.
To allow the compound effect of the two features to be assessed in a
grammar of point of view, the classification of adjectives reflects that of
adverbs (see section 6). Besides their syntactic properties, adjectives were
grouped into the following subcategories: temporal (point of time: the
present study; habitual: the usual reaction; generic: a common practice),
locative (the left ventricle), descriptive (a slow movement), evaluative (an
easy task), state of mind (I am sorry that I am late.), modal (epistemic: a
possible solution; perception: a manifest lie; deontic: a necessary change),
intensifying (a complete victory) and focusing (a particular text, an addi-
tional example). A great majority of the adjectives studied were either de-
scriptive (7,000) or evaluative (4,000); in addition, there were approxi-
mately 1,000 modal, 1,000 adjectives referring to state of mind, 750 to
time, 500 marking focus, 400 indicating place, and 300 with an intensi-
fying function.
The decision to focus chiefly on modal and evaluative adverbs and ad-
jectives of course drastically reduces the range of material which is un-
doubtedly relevant in the reconstruction of point of view. Not only modal
auxiliaries and verba sentiendi but also lexical choices of various kinds
(see for example Simpson 1993: 108-109) and features related to types of
transitivity are of great significance. The findings will, however, be com-
pared with the degree of involvement attested by means of features in-
cluded in Biber's Factor 1 (see Meurman-Solin 1994 and forthcoming; cf.
also Taavitsainen in the present volume). Moreover, the significant corre-
lation patterns between various types of attitudinal expressions in the ma-
terial included in this study stress their converging roles in a grammar of
point of view (see pp. 316-317, 324-328, Appendix 2 and Figure 1, p.
274).
274 Anneli Meurman-Solin

C/5 JI Ρ
ω
α Cu ώ
ω .s
*Ή -2§
6 Ii
11
X
Q S
o .
S £
ω
'Σλ
I § 'S
.Ξ ω
δII Ηπ IIf
< s <
ω
1? b-
S ^ -S
ο Ό B·
Λ οΕ ΈΟ
ω c
t η -α
ο
ο Ο.
S
5? οο
Λ έί W
c χΟ » Q
Λ
υ "Ξ ιι g
'Ξ •SPiiÄ
«1 ® ·§-
οο
•ο £ ·Η
c C
λ .ä· ω c
rC C β ·- ω
SP βο Λ ΛΜ
UΌ Ό
<Η Ο Cs
Έ Ε
.S«••3
P ο 3
s i 1
I ω II

ίί
a
 Reconstructing a grammar of point of view 275

5. General frequencies and distributions in genres

Because of lack o f space, the general frequencies o f open-class adverbs

and adjectives are given per genre in the f o l l o w i n g table; characteristics
o f the individual texts are c o m m e n t e d upon in the discussion that f o l l o w s
(see pp. 2 8 4 - 3 1 0 and A p p e n d i x 2). All mean frequencies in this study are
based on figures normalized per 1,000 w o r d s of r u n n i n g text. 7

Table 1. The general frequency of open-class adverbs and of adjectives in the

Early Modern English part of the Helsinki Corpus of English Texts.
Absolute figures and mean frequencies normalized per 1,000 words of
running text. Standard deviation in brackets. N=number of texts, fol-
lowed by abridged names of genres, attached to names of texts in the
tables throughout the study (for abbreviations used in this and sub-
sequent tables, see pp. ix-x and 345-363 in the present volume; see,
also, Kytö 1996).

Genre Adverbs/Mean/St.d. Adjectives/Mean/St.d.

Ν
Bible 100 2.3 (0.3) 326 7.5 (0.7)
2 (BIBLE)
Law 107 2.8 (1.9) 938 25.5 (1)
3 (LAW)
Autobiography 53 3.4 (2.3) 243 15.7 (1)
3 (BIA)
Comedies 125 3.4 (1.6) 771 21.8 (5)
6 (COME)
Diaries 132 3.6 (2.1) 812 22.4 (7.6)
6 (DIARY)
Trial proceedings 219 4.1 (2.4) 771 16 (5.6)
5 (TRI)
Travelogue 198 5.1 (1.6) 1218 31.9 (10.9)
6 (TRAV)
Sermons 197 6 (2.6) 961 29.1 (8.8)
6 (SERM)
Fiction 218 6.1 (2) 838 23.5 (2.5)
6 (FICT)
Handbooks 207 6.3 (2.4) 1287 38.8 (17.5)
6 (HANDO)
276 Anneli Meurman-Solin

Table 1. Continued.

Genre Adverbs/Mean/St.d. Adjectives/Mean/St.d.

Private letters 222 6.4 (1.6) 927 26.1 (4.4)

3 (LET PRIV)
History 229 6.9 (2.4) 890 26.9 (6.7)
6 (HIST)
Official letters 135 7.5 (1.9) 488 27.3 (3.2)
3 (LET NON-PRIV)
Philosophy 199 7.6 (2) 881 34.2 (5)
3 (PHILO)
Science/medicine 107 8.2 (5.2) 613 47.7 (6.2)
2 (SCIM)
Science/other 215 9.3 (6.5) 1109 46.1 (10)
4 (SCIO)
Biography 166 9.9 (2.7) 465 28 (5)
3 (BIO)
Education 383 11.8 (4.4) 1228 37.2 (3.4)
6 (EDUC)

Total 3212 6.2 (3.5) 14766 28.3 (11.3)

79

As regards adverb realizations, the Bible, law, autobiographies and com-

edies are at the low end of the scale, travelogues, trial proceedings and
diaries taking the position next to them, leaving handbooks, sermons, fic-
tion, private letters and history to represent genres close to the mean
(6.2). Philosophy, official letters and science, particularly science/other,
can be positioned next to the two genres, biography and education, which
represent the high end on the scale. Besides the Bible, philosophy and let-
ters are the most homogeneous genres in the material, but a relatively
high degree of uniformity has been attested in handbooks, education, his-
tory, biography and fiction. This finding is especially significant, as the
conditioning of date has been ignored in the table above. When develop-
ments in the three time periods are taken into account, for example the
heterogeneity of science can be explained by change over time, which is
reflected in the significantly lower mean frequencies in the pre-1570
scientific treatises (science/medicine, i.e. Vicary, 4.5; science/other, Re-
cord, 3.9) than in the later ones (science/medicine, Clowes, 11.9, science/
 Reconstructing a grammar of point of view 277

other 11.1, st.d. 6.6; on diachronic developments and further details about
heterogeneity, see pp. 294-295, 300, 301, 305).
Interesting continua, which can be related not only to point of view but
also to degree of informational density, can be identified. There is for ex-
ample an important difference in the general frequency of open-class ad-
verbs between author-centred narratives such as diaries and auto-biogra-
phies and, on the other hand, biographies, where the author looks at his
subject from outside — as narrator, analyser or commentator. A high fre-
quency of open-class adverbs seems to be a particularly prominent feature
in educational treatises. These texts also contain a high mean frequency
of adjectives and seem to be strikingly uniform in this respect as com-
pared for example with handbooks. It is however science that stands out
in the statistics, with almost 50 adjectives per 1,000 words (cf. the Bible,
trials and autobiographies at the low end; see also section 8).

6. Semantic roles and syntactic functions

6.1. Analysis and classification of the corpus material

Before discussing the semantic properties of adverbs and adjectives and

their relation to extralinguistic variables I would like to illustrate the ma-
terial and the linguistic features studied in it by providing extracts from
a number of texts. Extracts from Roger Ascham's Scholemaster (1563-
1568) illustrate a text where adverbs and adjectives are particularly nu-
merous. The items considered in the analysis have been underlined.

(1) There is a waie, touched in the first booke of Cicero De

Oratore, which, wiselie brought into scholes, truely taught, and
co~stantly vsed, would not onely take wholly away this butcher-
lie feare in making of latines, but would also, with ease and
pleasure, and in short time, as I know by good experience,
worke a true choice and placing of wordes, a right ordering of
sentences, an easie vnderstandyng of the tonge, a readiness to
speake, a facultie to write, a true judgement, both of his owne,
and other mens doinges, what tonge so euer he doth vse.
(Ascham, The Scholemaster 183)
278 Anneli Meurman-Solin

(2) In very deede, if childre- were brought vp, in soch a house, or

soch a Schole, where the latin tonge were properlie and perfitlie
spoken, as Tib. and Ca. Gracci were brought vp, in their mother
Cornelias house, surelie, than the dailie vse of speaking, were
the best and readiest waie, to learne the latin tong. But, now,
commonlie, in the best Scholes in England, for wordes, right
choice is smallie regarded, true proprietie whollie neglected,
confusion is brought in, barbariousnesse is bred vp so in yong
wittes, as afterward they be, not onelie marde for speaking, but
also corrupted in judgement: as with moch adoe, or neuer at all,
they be brought to right frame againe. (Ascham, The Schole-
master 185)

While Ascham's prose is a good example of an abundant use of evalua-

tive adverbs and adjectives especially as expressions of attitude, Lady
Margaret Hoby's diary is marked by the absence of this feature.

(3) Thursday 13
In the morninge, after priuat praier, I wret some thinges touch-
inge Houshould mattres: then I did eate my breakfast, and did
order diuerse thinges in the granirie: sonne after Mr Hoby Came
home, and I kept him Companie tell he went away againe: after
diner I went to Birstall, and se him, who was not well, and his
aples tithed: then I wrought tell almost: 6 :, and praied with Mr
Rhodes, and priuatly in my Closett : after medetation, I went to
supper : after, I had reed of the bible, after to lector, and then to
bed. (Hoby, Diary of Lady Margaret Hoby 71)

The text is organized by means of time relators. Of the total of adverbs

in the diary, 82% are descriptive, such as priuatly in the passage above.
The material attested in the corpus suggested an initial classification of
adverbs based on the principal semantic roles such as space, time, pro-
cess, respect, contingency, modality and degree (cf. Quirk et al. 1985:
479). The role of contingency was not expressed by means of open-class
adverbs. Similarly, it was not possible to identify unambiguous examples
of adverbs in the role of respect. The role of space was rarely realized by
means of open-class adverbs, and items of interest in the reconstruction
of point of view, such as universally, were labelled as descriptive, despite
the semantic components they share with adverbs of generic time refer-
ence (see further section 6.3).8
 Reconstructing a grammar of point of view 279

After setting apart the uses as conjuncts, adverbs in the role of time
were grouped into three subcategories on the basis of type of time refer-
ence: (i) point of time, (ii) habitual, and (iii) generic. Because of their re-
lation to modality, particularly important among time adverbs are those
with generic time reference (see p. 290):

(4) The Rhenish wine that is co~monly drunken in Gentlemens

houses and Citizens houses is commonly a yere old at γ least
before it be drunken, & therfore it is older than ye common
Clared wine, which dureth not commonlye aboue one yeare, and
if Rhennish wyne be drunken within the yeare, it is commonly
racked before it be drunken, therfore for two causes it hath
fewer dregges and lesse terrestritie or grosse earthlynesse than
the Clared wine hath, and therfore breedeth the stone lesse than
Clared wine that is commonly drunke in gentlemens houses
doth. If I can proue this y' I have sayde, and also that Clared
wine is hoter than white Rhennish and white French wines be,
there is nothing to let me but I may conclude without anye
withstanding, that Clared or red wines breede the stone more
than white wines do. (Turner, A New Boke of the Natures and
Properties of All Wines B8R)

In this example, Turner aims at persuasiveness by claiming general valid-

ity for practices and phenomena he describes in his handbook. The tauto-
logous use of commonly(e) in the passage adds to the truth-value of the
generalizations in a way that can be compared with the use of modal ad-
verbs and adjectives.
In the analysis of adverbs in the roles of process, modality and degree,
it was assumed to be particularly useful to try to distinguish descriptive
adverbs, such as predicate-related adverbs depicting concrete processes,
from modal and evaluative ones, and from those in the role of degree.
The latter group, which Halliday (1985: 82) calls mood adjuncts and
comment adjuncts, is related to the macrofunction 'interpersonal' and is
therefore central to a study of point of view. The emphasis is on expres-
sions of attitude by participants, but focusing subjuncts, restricting the at-
tention to the constituent they modify, have also been looked at; their
function as explicit markers of the author's manipulation of his/her text
are particularly interesting.9
The role of process turned out to be realized by predicate-oriented de-
scriptive adverbs such as priuatly in the extract from Lady Hoby's diary
280 Anneli Meurman-Solin

(example (3)), or evaluative adverbs such as properlie in the extract from

Ascham (example (2)). An adverb used as process adjunct in one context
may in other contexts be subject-oriented and refer to state of mind, syn-
tactically functioning as item subjunct implying affect or attitude (see
examples (7)-{10)), or it can be a syntactically peripheral (i.e. non-in-
tegrated) comment of the author or the narrator, such as wiselie in ex-
ample (2).10 With the exception of modal manifestly and the textually
used agreeably and lately, the adverbs underlined in the following pas-
sage have been categorized as predicate-oriented descriptive adverbs func-
tioning as process adjuncts:

(5) And agreeably to this we found, that if, as soon as the sus-
pended and well rubb'd Electric was brought to settle freely, we
applied to the chafed edge, but without touching it, the lately
mention'd Cushion, which, by reason of its rough Superficies
and porosity, was fit for the Electrical Effluvia to fasten upon,
the edge would manifestly be drawn aside by the Cushion
steadily held, and if this were slowly removed, would follow it
a good way; and when this body no longer detain'd it, would re-
turn to the posture wherein it had settled before. (Boyle, Elec-
tricity & Magnetism 19)

In contrast, process adjuncts in the following passage express the author's

evaluation of the action denoted by the predicate verbs the adverbs modi-
fy:

(6) Kynge Philip, whan he harde that his sonne Alexander dyd
singe swetely and properly, he rebuked him gentilly, sayinge,
But, Alexander, be ye nat ashamed that ye can singe so well
and connynglyl (Elyot, The Boke Named the Gouernour 27)

It is however assumed that, in the assessment of stance marking, it is rel-

evant to distinguish between these predicate-oriented evaluative adverbs
and subject-oriented adverbs of the following kind:

(7) ... he had informed the Frenche King and his Cleargie and
realme to the end that..., and that by the magnificence of these
two Kings, co~ncorde mighte be made in the Church, vnto the
 Reconstructing a grammar of point of view 281

whiche businesse the Frenche Kyng hadde gladlye graunted,

and sente messengers vnto Pope Gregorie, who notwythstandyng
was obstinate. (Stow, The Chronicles of England 571)

(8) To which Sir John Perrott answered, Good my Lord (sayd he)
I hope you know you may Command my Life and my Livinge,
yet lay not that Burden on me, but leave me to enjoy my Con-
science, and I will not willingly meddle with other Mens Con-
sciences. (Perrott (?), The History of That Most Eminent States-
man, Sir John Perrot 37-38)

(9) They are only our ofspring that we neglect in this point and
haveing made them ill children we foolishly expect they should
be good men. (Locke, Directions Concerning Education 51)

In the first of the three examples, gladlie has been interpreted to express
the author's evaluation of the king's state of mind. As an attitudinal ex-
pression, however, it may be relatively weak, as it is possible that a more
extensive corpus would give evidence of a collocational status of 'gladly
grant'. In the second example willingly is simultaneously subject- and
speaker-oriented. In the third example foolishly refers to an indefinite
plural subject we, including the author. Whatever variation there is in
their relation to the grammatical subject, the speaker as subject or the
combined subject- and narrator-orientation, the term 'subject-oriented' is
used to refer to all adverbs functioning as item subjuncts, and the adverbs
are also presented as one group in the statistics. Some subject-oriented
adverbs such as humbly in the following extract are undoubtedly con-
ventional expressions in letters:

(10) Right worshipful mother, I humbly recommend mee unto you,

desiring you of your dayly blessing, praing Jesu long to con-
tinew your helth to his pleasur. (Agnes Plumpton, Plumpton
Correspondence 231)

The identification of sentence adverbs is quite problematic, as word or-

der or punctuation cannot be used as conclusive evidence. It is often im-
possible to make a distinction between markers of evaluation that func-
tion as adjuncts or subjuncts and those that function as disjuncts. This is
282 Anneli Meurman-Solin

particularly true in narratives where the author now and then, in more or
less implicit ways, takes the role of the narrator, only to disappear subse-
quently. The adverb wisely in They wisely left the wording unchanged
functions as a marker of point of view, whether it is interpreted to mean
Ί (= the author or the speaker) think it was wise of them to leave the
wording unchanged' or 'They were wise when they left the wording un-
changed'. It is only in cases like Wisely, they left the wording unchanged
or The wording was wisely left unchanged that the disjunct function has
been considered unambiguous. In addition to initial position and/or punc-
tuation, the verb category is of relevance, as sentence adverbs have been
shown to be more frequent when the propositions adjoining them have
stative verbs as predicate verbs (cf. Swan 1988: 376, 385, 407, 423).
Examples of the following kind have been interpreted to function as
disjuncts:

(11) If happily some few had noe inclination that waye, yet, beeing
unable to stay the course of so greate a current, they were con-
tent not to shew a will to affect that which they had no power
to effect. (Hayward, Annals of the First Four Years of the Reign
of Queen Elizabeth 3)

(12) So this Knight spending his Patrimony (as many of his Yeres
and Calling do now-a-dayes, wastfully, and above their Habili-
tie) had lefite but a bare Hope to recover his Estate. (Perrott (?),
The History of That Most Eminent Statesman, Sir John Perrot
32)

(13) I have often heard him say, that he admir'd by what strange in-
spiration he came to talk things so soft, and so passionate, who
never knew love, nor was us'd to the conversation of women;
but (to use his own words) he said, most happily, some new,
and till then, unknown power instructed his heart and tongue in
the language of love, and at the same time, in favour of him,
inspir'd Imoinda with a sense of his passion. (Behn, Oroonoko
156)

(14) The like success I had with the end of an Iron Key, and the like
also with a cold piece of polish'd black Marble; and sometimes
 Reconstructing a grammar of point of view 283

the feathers did so readily and strongly fasten themselves to

these extraneous and unexcited Bodies, that I have been able
(though not easily) to make one of them draw the feather from
the Amber it self. (Boyle, Electricity & Magnetism 32-33)

An attempt has been made to identify examples which could illustrate the
various stages in the development from integrated to peripheral adverbiale
(see pp. 280, 311), but, within the scope of this study, a more detailed
analysis is not possible.
The role of modality is realized by emphasizing subjuncts and disjuncts
commenting on truth-value. It was no less problematic to distinguish the
peripheral from the integrated adverbials here than in the case of evalua-
tives. Only unambiguous cases were included in the category of disjuncts.
As regards the role of modality, examples were further subcategorized
into epistemic, perception, deontic and boulomaic modality. Epistemic
modality is typical of the texts in dialogue form:

(15) L. C. J. You are sure ofthat?

Dunne. Yes, my Lord.
L. C. J. Consider of it, are you sure there was no body else?
Dunne. Truly, my Lord, I did not mind that there was any body
else.
L. C. J. Recollect your self, and consider well of it.
Dunne. Truly, my Lord, I do not know of any body else.
{State Trials, The Trial of the Lady Alice Lisle IV 114.C2)

Perception modality frequently co-occurs with verbs of perception:

(16) ... placing it on a black object Plate, because it was it self a

white body, and casting the light on it with a deep plano-convex
Glass, I could exceeding plainly perceive it to be all perforated
and porous, much like a Honey-comb, but that the pores of it
were not regular; yet it was not unlike a Honey-comb in these
particulars. (Hooke, Micrographia 13 5 113)

(17) In al the 3. bathes a man may evidently se how the water burbe-
lith up from the springes. (Leland, The Itinerary of John Leland
I 142)
284 Anneli Meurman-Solin

No examples were attested of boulomaic modality, and deontic modality

expressed by means of open-class adverbs was found to be relatively rare:

(18) The Husbandman whose occupation is the generali affaires of

the common wealth, as some to the Markets, some to the Citie,
and some to the seates of Iustice; must necessarilie be imploied
almost in continuall trauell: And therefore it is meet that he be
prouided euer of a good and easie trauelling Horse. (Markham,
Countrey Contentments 72)

There are important differences both between genres and between indi-
vidual texts as regards the general frequency of specific types of adverbs
and adjectives in the corpus. Before focusing on the semantic roles of
evaluation, modality and degree, I summarize frequencies and distribu-
tions of adverbs and adjectives in the other principal roles (see pp. 272-
273).

6.2. Descriptive adverbs and adjectives

Descriptive adverbs and adjectives form the most important group of fea-
tures that can be meaningfully contrasted with adverbs and adjectives
functioning as markers of point of view. Descriptive open-class adverbs
functioning as process adjuncts are manifestly infrequent in the material
as compared with descriptive adjectives. The total mean frequency of the
former is 0.7, st.d. 0.7, that of the latter 13.4, st.d. 9.4. There is more uni-
formity in the distribution of adjectives, but also a significant difference
between the frequency of modifiers and complements; standard deviation
is slightly lower in the use of modifiers (mean 9.6, st.d. 7) than in that of
complements (3.8, st.d. 3.1).
In order to allow comparisons and an assessment of the statistical sig-
nificance of the findings, both mean frequencies per 1,000 words of run-
ning text and percentages of the totals of adverbs and adjectives are given
in the following table. A high mean frequency of descriptive adjectives is
assumed to have the highest negative score on the dimension 'high inten-
sity of viewpoint marking' versus 'low intensity of viewpoint marking';
a high frequency of descriptive adverbs can be given a lower negative
weight. The texts in the table have therefore been listed in an order that
highlights, in the third column, the increase in the mean frequencies of
 Reconstructing a grammar of point of view 285

descriptive features in the texts of each period (for abridged names of

genres, see Table 1).

Table 2. Mean frequencies and percentages of descriptive open-class adverbs

and adjectives in the Early Modern English part of the Helsinki Cor-
pus of English Texts. Percentages of descriptive adjectives as comple-
ments in brackets.

Adverbs Adjectives
Mean % Mean %

1500-1570:
XX CORP LET PRIV 0.8 10 4.3 20 (16)
XX BIBLE TYNDALE 0.5 24 4.3 54 (17)
XX COME STEVENSO 0 0 5.9 38 (20)
XX COME UDALL 0 0 6.7 35 (24)
NN BIA MOWNTAYNE 0.3 6 7 42 (24)
IR SERM FISHER 0.7 8 7.4 28 (25)
NN TRAV TORKINGT 0 0 7.5 41 (32)
XX CORO LET NON-PRIV 0.2 2 8.9 30 (20)
NN HIST MORERIC 1.9 18 9.9 35 (19)
NI FICT HARMAN 1.8 20 10 40 (24)
NN DIARY MACHYN 0.1 11 10.3 80 (77)
NI FICT MERRYTAL 0.8 24 10.4 51 (34)
NN HIST FABYAN 0.6 9 10.5 52 (42)
XX PHILO BOETHCO 0.9 12 10.7 27 (17)
XX TRI THROCKM 0.8 10 11.8 51 (35)
IS/EX EDUC ASCH 2.4 13 12.4 36 (32)
STA LAW STAT3 0.4 15 12.6 51 (43)
NN DIARY EDWARD 0.2 8 13.1 65 (60)
IS/EX EDUC ELYOT 0.4 4 13.2 33 (31)
NN BIO ROPER 1.1 9 14.5 45 (31)
IR SERM LATIMER 0.4 12 14.8 48 (39)
NN TRAV LELAND 0.1 3 17.2 63 (53)
IS HANDO FITZH 0.2 4 19.2 71 (46)
EX SCIO RECORD 1.8 46 30.7 87 (63)
EX SCIM VICARY 2.1 46 30.9 71 (37)
IS HANDO TURNER 1.6 16 57.7 85 (64)

Total N=-26 0.8 13.9

Standard deviation 0.7 11.1
286 Anneli Meurman-Solin

Table 2. Continued.

Adverbs Adjectives
Mean % Mean %

1570-1640:
XX BIBLE AUTHOLD 0.4 15 4 56 (27)
XX TRI ESSEX 0.2 5 4.3 25 (18)
XX TRI RALEIGH 0.4 11 6.8 36 (24)
XX CORP LET PRIV 0.3 7 7.1 23 (15)
IS HANDO GIFFORD 0.2 5 7.4 33 (18)
XX COME SHAKESP 0.6 16 7.5 40 (27)
NN ΒΙΑ FORMAN 0.2 7 7.5 51 (30)
IR SERM HOOKER 0.6 11 8.2 38 (19)
EX EDUC BRINSLEY 1.1 7 8.3 26 (19)
XX COME MIDDLET 0.4 8 9.6 34 (23)
NN BIO PERROTT 1 15 9.6 43 (33)
NN HIST STOW 0.4 9 9.8 53 (40)
NI FICT DELONEY 0.3 4 10.4 47 (31)
NN DIARY MADOX 0 0 10.4 54 (36)
XX PHILO BOETHEL 1.2 20 10.5 31 (22)
XX CORO LET NON-PRIV 0.7 11 10.8 38 (30)
IR SERM SMITH 0.2 6 11 61 (22)
EX EDUC BACON 0.2 3 12 31 (26)
STA LAW STAT4 0.3 36 13.5 50 (42)
NN TRAV COVERTE 0.7 9 14.5 61 (43)
NI FICT ARMIN 0 0 15.1 57 (38)
NN HIST HAYWARD 0.9 11 16.1 44 (30)
NN DIARY HOBY 2.8 81 16.4 79 (75)
NN TRAV JOTAYLOR 0.6 12 17.1 51 (36)
EX SCIO BLUNDEV 0.6 15 25.2 64 (52)
EX SCIM CLOWES 1.4 11 26.1 50 (41)
IS HANDO MARKHAM 0.3 4 30.7 62 (35)

Total N-=27 0.6 12.2

Standard deviation 0.6 6.5
 Reconstructing a grammar of point of view 287

Table 2. Continued.

Adverbs Adjectives
Mean % Mean %

1640-1710:
XX TRI LISLE 0.2 6 3.2 29 (25)
XX TRI OATES 0.3 10 3.9 39 (31)
IR SERM TILLOTS 0.6 6 6.8 17 (14)
NN DIARY PEPYS 0 0 7.2 28 (16)
NN BIA FOX 0 0 7.2 47 (34)
XX CORO LET NON-PRIV 0.3 5 7.3 31 (27)
NI FICT PENNY 0.8 10 7.6 35 (22)
XX CORP LET PRIV 0.2 4 7.8 30 (16)
NI FICT BEHN 1.3 20 7.8 30 (25)
XX COME FARQUHAR 0.2 8 8.1 35 (19)
NN HIST BURNETCHA 0.2 4 8.9 30 (22)
XX COME VANBR 0.7 12 9 35 (22)
NN BIO BURNETROC 0.6 6 9.1 31 (18)
IS EDUC LOCKE 0.8 7 9.8 24 (16)
XX PHILO BOETHPR 0.6 6 10 34 (23)
IS HANDO WALTON 0.7 12 12.7 49 (32)
NN HIST MILTON 1.5 22 13.1 47 (36)
IR SERM JETAYLOR 0.7 12 15.5 40 (29)
STA LAW STAT7 2.3 48 16.7 67 (56)
NN DIARY EVELYN 0.5 7 17 48 (37)
IS EDUC HOOLE 0.7 6 18.1 47 (41)
IS HANDO LANGF 0.8 16 24.5 61 (35)
NN TRAV FRYER 0.4 6 26.5 60 (43)
EX SCIO BOYLE 2.5 14 35.1 64 (48)
NN TRAV FIENNES 0 0 37.4 84 (62)
EX SCIO HOOKE 2.1 18 39.6 73 (51)

Total N==26 0.7 14.2

Standard deviation 0.7 10.2

Total N==79 0.7 13.4

Standard deviation 0.7 9.4
288 Anneli Meurman-Solin

In Table 2, the relatively low proportion of adverbs that concretely de-

scribe an action or a process (see example (5)) stresses the dominant role
of attitudinal expressions among adverbs (cf. the proportions of descrip-
tive and attitudinal ajectives). Lady Hoby's diary is the only text in the
Early Modern English part of the Helsinki Corpus where the majority of
open-class adverbs are descriptive. In addition, the highest means have
been attested in scientific treatises (with the exception of Blundevile, the
texts here labelled SCIM and SCIO). There is an interesting trend towards
a higher mean frequency of these adverbs in Statutes of the Realm·, also
their proportion of all open-class adverbs increases from 15% to 48%. No
such straightforward pattern of change is reflected in other genres or text
categories; standard deviation is mostly high (see for example educational
treatises and handbooks, and, among texts representing non-imaginative
narration, histories, biographies and travelogues; for fiction, see Meur-
man-Solin 1995). Despite the obvious functional differences of the genres,
the total absence of descriptive open-class adverbs in a number of trav-
elogues positions them close to diaries and autobiographies.
As regards descriptive adjectives, both their mean frequencies and pro-
portion are exceptionally high in Turner, Vicary, Record, Markham,
Clowes, Blundevile, Langford, Boyle, Hooke, Fiennes and Fryer (see also
Leland and Coverte). The combined effect of high frequency and a pre-
valent role in the total of adverbs in the texts makes it evident that the
feature is genre-specific in science/medicine (28.5, st.d. 3.4) and in
science/other (32.7, st.d. 6.1), perhaps also in handbooks (25.4 st.d. 17.9)
and in travelogues (20, st.d. 10.5). Scientific treatises are strikingly uni-
form in this respect. In addition to these genres, the proportion of de-
scriptive adjectives is rather consistently higher in diaries (see especially
Edward VI 65%, Machyn 80% and Lady Hoby 79%).
From a diachronic perspective the mean frequencies seem to become
lower in texts representing handbooks (1st subperiod: 38.5, st.d. 27.2;
2nd: 19, st.d. 16.4; 3rd: 18.6, st.d. 8.4) and in trials (1st: 11.8; 2nd: 5.6,
st.d. 1.7; 3rd: 3.5, st.d. 0.5), but higher in travelogues (1st: 12.3, st.d. 6.9;
2nd: 15.8., st.d. 1.8; 3rd: 31.9, st.d. 7.7). The last-mentioned development
is part of a differentiation of genres representing non-imaginative narra-
tion. With more texts per period, trends of this kind may turn out to be
less straightforward. Moreover, depending on the features looked at, there
are often major differences between the two texts representing a period.
This is the case for example in handbooks, where, in the period 1500—
1570, Fitzherbert (mean 19.2) and Turner (mean 57.7) differ from each
other to a considerable extent as a result of the very repetitive style of the
 Reconstructing a grammar of point of view 289

latter. In fact, information about the type/token ratio would be relevant

here. Yet the striking uniformity in science seems to provide evidence
diagnostic of text type. For further information on descriptive adjectives
as complements, see section 8.

6.3. The role of time

Time position adverbs represent 15-26% of the total of open-class ad-

verbs in texts representing non-imaginative narration,11 in fiction (A
Hundred Mery Talys), in Middleton's and Farquhar's comedies, Walton's
handbook and private letters, in the last-mentioned only in the third sub-
period. Fabyan's history is at the high end of the scale with 39% of the
total of open-class adverbs referring to a point of time. In their function
as time relators (Quirk et al. 1985: 1451-1453), these adverbials are thus
a generic feature in non-imaginative narration. There is a highly signifi-
cant correlation between adverbs and adjectives in the role of time (see
note 13).
Adverbs with generic time reference (mean=0.2, st.d. 0.4) occur äs a
text-type-specific feature in handbooks with the exception of Gifford and
Walton (cf. however the proportion of semantically-related modal adverbs
in Gifford in Table 4 and of adverbs with habitual time reference (0.6; cf.
total mean=0.2, st.d. 0.3) in Walton). The means of adverbs such as com-
monly and generally are as high as 1.9 in Turner, 1.2 in Fitzherbert, 1 in
Markham and 0.7 in Langford.12 It is also worth noting that adverbs
with generic time reference are very rare in interactive texts, which at
least partly explains their absence in Gifford and Walton, both containing
passages in dialogue form (see p. 322).
The role of generic time is strengthened by adjectives in Fitzherbert
(1.2; cf. total mean 0.2, st.d. 0.4), that of habitual time (total mean 0.5,
st.d. 0.5) especially in Markham (1.8), Clowes (1.2), Boyle (1.3), Brinsley
(2.7), Locke (2.3), Elyot (1.3) and Hoole (1.3). A new feature is the rela-
tively high frequency of these adjectives in texts representing sermons
(generic: Fisher 2.5, Tillotson 0.9; habitual: Taylor 1.5). Educational trea-
tises are marked by the highest mean frequencies of adjectives with ha-
bitual time reference (1.4, st.d. 0.9), and sermons by those with generic
time reference (0.6), but the latter are very heterogeneous in this respect
(st.d. 1). As a whole, evidence provided by adjectives of this kind seems
to be text-specific rather than valid for identifying differences between
text types.
290 Anneli Meurman-Solin

Generic time reference and habitual time reference expressed by means

of open-class adverbs correlate significantly with one another (p=.003).13
Adverbs with generic time reference also correlate significantly with per-
ception modality (p=.017), particularly in Turner and Blundevile (see fur-
ther pp. 300-301). Moreover, adjectives with habitual and generic time
reference correlate highly significantly (p=.000) with evaluative, intensify-
ing and focusing adverbs, and significantly (p=.004) with adverbs in the
role of perception modality. Because of these significant correlations, ex-
pressions in the two subroles of time have been included among markers
of attitude.

6.4. Point-of-view markers

The following sections discuss the central markers of point of view,

grouped into evaluative, modal (including generic and habitual time refer-
ence as mentioned above), intensifying and focusing expressions, first by
looking at their normalized frequencies in texts and genres (sections
6.4.1-6.4.3) and then by commenting on their relative proportion in the
total of open-class adverbs and that of adjectives, in genres, text catego-
ries and texts grouped together on the basis of some other extra-linguistic
variable (sections 10-11). The main emphasis in the discussion is on the
presence of viewpoint markers, although their absence is just as relevant.
Both percentages and mean frequencies are essential; the former indicate
the relative importance of the feature in a text, the latter allow compari-
sons between them.

6.4.1. Evaluative adverbs and adjectives

The texts in the following table are listed in order of increasing frequen-
cies of evaluative adverbs, as adverbs have been assumed to have a high-
er weight in the role of evaluation than adjectives (cf. factor 1 in Biber
1988: 102). On the same rationale, the percentages of complements is
also indicated, as complements are syntactically less integrated and there-
fore more dynamic as attitudinal expressions than attributive adjectives
(see section 8; cf. Biber 1988: 237).
 Reconstructing a grammar of point of view 291

Table 3. Mean frequencies and percentages of evaluative open-class adverbs

and of adjectives in individual texts in the Early Modern English part
of the Helsinki Corpus of English Texts. Percentages of evaluative ad-
jectives as complements in brackets.

Adverbs Adjectives
Mean % Mean %

1500-1570:
XX BIBLE TYNDALE 0.3 16 1.9 24 (15)
EX SCIM VICARY 0.6 14 2.9 7 (4)
STA LAW STAT3 0.7 24 4.2 17 (7)
NI FICT MERRYTAL 0.9 29 3.6 18 (5)
XX COME UDALL 0.9 40 5.4 28 (4)
NN DIARY MACHYN 0.9 67 1.5 11 (0)
NN DIARY EDWARD 1.1 58 2.1 10 (3)
XX COME STEVENSO 1.3 73 3.4 22 (8)
NN HIST FABYAN 1.5 24 2.6 13 (2)
EX SCIO RECORD 1.5 38 0.6 2 (0)
NN TRAV TORKINGT 1.5 46 6.2 34 (0)
NN TRAV LELAND 1.6 34 4.7 17 (5)
XX PHILO BOETHCO 1.9 26 17.9 45 (16)
IR SERM LATIMER 2 59 10.6 35 (16)
IS HANDO FITZH 2.1 44 4.7 17 (5)
IS HANDO TURNER 2.7 26 6 9 (3)
NN BIA MOWNTAYNE 3.5 63 5.4 32 (8)
XX CORP LET PRIV 3.7 45 6 27 (6)
IR SERM FISHER 3.8 45 6.7 26 (3)
XX TRI THROCKM 4.6 54 5.8 25 (7)
NN HIST MORERIC 4.9 47 7.4 26 (8)
NI FICT HARMAN 5.7 64 8.6 35 (10)
XX CORO LET NON-PRIV 6 62 10.3 35 (6)
IS/EX EDUC ELYOT 6.5 67 17.1 44 (7)
NN BIO ROPER 7.4 63 8.8 27 (10)
IS/EX EDUC ASCH 9.8 52 14.7 42 (9)

Total N==26 3 6.5

Standard deviation 2.5 4.5
292 Anneli Meurman-Solin

Table 3. Continued.

Adverbs Adjectives
Mean % Mean %

1570-1640:
NN DIARY HOBY 0.3 10 2.8 13 (11)
XX BIBLE AUTHOLD 0.5 20 1.4 20 (14)
ST LAW STAT4 0.5 55 6.4 24 (3)
EX SCIO BLUNDEV 0.9 23 0.8 2 (2)
IS HANDO GIFFORD 1.1 33 9.5 42 (11)
NN BIA FORMAN 1.2 33 3.6 25 (8)
XX TRI RALEIGH 1.6 48 6.1 32 (12)
NI FICT DELONEY 1.9 31 6.4 29 (ID
NN HIST STOW 1.9 41 3.7 20 (2)
XX COME SHAKESP 1.9 48 4.7 25 (10)
XX TRI ESSEX 2.2 68 6 35 (10)
XX PHILO BOETHEL 2.3 40 14.4 43 (14)
IR SERM SMITH 2.3 67 2.9 16 (11)
IS HANDO MARKHAM 2.5 33 10.2 21 (9)
NN BIO PERROTT 2.5 36 7.5 33 (7)
XX COME MIDDLET 2.5 54 12.4 44 (13)
NN DIARY MADOX 2.5 59 4.9 26 (10)
EX EDUC BACON 2.6 44 18.7 49 (15)
XX CORP LET PRIV 2.7 53 6.2 20 (8)
NN TRAV JOTAYLOR 2.8 60 10.6 32 (6)
IR SERM HOOKER 2.9 54 7.8 36 (15)
XX CORO LET NON-PRIV 3 47 8.5 30 (7)
NI FICT ARMIN 3.1 70 6 23 ( i)
NN TRAV COVERTE 3.5 46 4.7 20 (6)
NN HIST HAYWARD 4.2 47 10.6 29 (16)
EX EDUC BRINSLEY 7.2 49 14.4 46 (19)
EX SCIM CLOWES 7.3 61 15 29 (11)
Total N==27 2.5 7.6
Standard deviation 1.6 4.5

1640-1710:
NN BIA FOX 0.7 67 4.3 28 (8)
XX TRI OATES 0.8 29 0.9 9 (4)
IS HANDO LANGF 1 19 7.3 18 (10)
STA LAW STAT7 1.1 24 3.8 15 (5)
NI FICT BEHN 1.3 20 10.9 42 (17)
 Reconstructing a grammar of point of view 293

Table 3. Continued.

Adverbs Adjectives
Mean % Mean %

XX TRI LISLE 1.3 44 3.3 31 (8)

XX COME FARQUHAR 1.3 58 9.9 42 (15)
XX CORP LET PRIV 1.8 30 7.1 28 (13)
NN DIARY PEPYS 1.8 50 11.3 44 (17)
IS HANDO WALTON 1.9 29 8.2 32 (15)
NN TRAV FIENNES 1.9 48 5.6 13 (7)
NN HIST BURNETCHA 2.1 46 12.5 42 (14)
IR SERM JETAYLOR 2.6 45 14.5 37 (15)
NI FICT PENNY 2.9 39 6.7 31 (10)
NN TRAV FRYER 3 48 11.1 25 (9)
XX COME VANBR 3.1 54 9.9 38 (6)
XX CORO LET NON-PRIV 3.2 50 6.6 28 (16)
IR SERM TILLOTS 3.6 38 19.1 49 (15)
NN DIARY EVELYN 3.6 49 9.9 28 (8)
IS EDUC LOCKE 3.7 34 18.1 44 (20)
EX SCIO HOOKE 4 33 6.8 12 (7)
NN HIST MILTON 4 56 7.6 27 (10)
XX PHILO BOETHPR 4.3 44 11.6 39 (22)
IS EDUC HOOLE 4.9 45 12.4 32 (12)
NN BIO BURNETROC 5.5 49 11.3 38 (12)
EX SCIO BOYLE 7.1 41 9 16 (8)

Total N==26 2.8 9.2

Standard deviation 1.6 4.2

Total N==79 2.8 7.8

Standard deviation 1.9 4.5

In Table 3, mean frequencies and percentages provide information of a

very different kind. There may be about the same proportion of evaluative
adverbs (50-60%) in Ascham's Scholemaster, in the trial of Throckmor-
ton and in the diary of Edward VI, but the first contains twice as many
occurrences per 1,000 words of running text as the second, and ten times
as many as the third. At least twice as high mean frequencies as the total
mean of evaluatives (2.8) have been attested in Ascham, Roper, Elyot,
Harman and official letters in the first subperiod, Brinsley and Clowes in
294 Anne Ii Meurman-Solin

the second, and Boyle and Burnet's biography in the third. As the com-
bined effect of percentage and frequency is significant especially in the
texts of the first subperiod and in Clowes, the presence of evaluative ad-
verbs can be claimed to be particularly manifest in these texts. Evaluative
adverbs are also twice as frequent in Preston's translation of Boethius as
in the two earlier ones.
There is some evidence of genre-specific distributions: four out of six
educational treatises, two out of three biographies, three out of six his-
tories and three out of six scientific treatises show a tendency to use nu-
merous evaluative adverbs, so that both their mean frequency and their
proportion of the total of adverbs are manifestly high. In addition, Tho-
mas Harman's Caveat is shown to differ from other representatives of fic-
tion. In contrast, evaluative adverbs are rare in texts such as the Bible,
statutes, some diaries and Vicary's Anatomy, but also in a number of
comedies and handbooks, in two out of three autobiographies, and Blun-
devile's treatise on geometry.
The mean frequencies remain relatively stable over time. Among hand-
books, the two texts representing a subperiod show some uniformity (1st
subperiod: 2.4, st.d. 0.4; 2nd: 1.8, st.d. 0.9; 3rd: 1.4, st.d. 0.6). Both
frequency- and percentage-wise, Fitzherbert (1500-1570) and Langford
(1640-1710) represent the two extremes on the continuum, so that there
is a trend towards a less significant role of evaluation. There is a manifest
decrease also in educational treatises from the mean frequency of 8.1
(st.d. 2.3), through 4.9 (st.d. 3.2) to 4.3 (st.d. 0.9). The texts in the first
and third subperiods are quite homogeneous as compared with Brinsley
and Bacon, which represent the second subperiod. A similar decrease has
been attested in autobiographies (from 3.5 to 0.7), private letters (from
3.7 to 1.8) and official letters (from 6 to 3.2). Despite important differ-
ences, handbooks and educational treatises share the general function of
being instructive. It is interesting that a decrease of evaluative adverbs
over time should characterize both genres. The trend towards more evalu-
ative adverbs especially in science (medicine: 0.6 < 7.3; other: 1.5 < 5.5,
st.d. 2.2) would have to be studied further using a more extensive corpus,
but the important role of these adverbs in texts such as Clowes and Boyle
is worth noting (cf. Table 8).
More than 40% of all adjectives are evaluative in five out of six educa-
tional treatises and in the Boethius translations (Preston 39%). Middleton
and Farquhar stand out as marked by the role of these evaluatives among
comedies, Gifford among handbooks, Tillotson among sermons, Burnet
among histories, Pepys among diaries and Behn in fiction. These are the
 Reconstructing a grammar of point of view 295

most important texts where adjectives of this kind also occur most dense-
ly; to them should, however, be added Clowes's Treatise for the Artifi-
cial! Cure of Struma (mean 15), which differs from other scientific trea-
tises to a considerable extent (cf. the very low mean frequencies (0.6-2.9)
in the sixteenth-century treatises and the somewhat higher frequencies
(6.8 and 9) in the texts dating from the last subperiod). Uniformity is
quite conspicuous in educational treatises (15.9, st.d. 2.4), philosophy
(14.6, st.d. 3.2) and biographies (9.2, st.d. 2), but also where infrequency
of these adjectives has been attested, for example in law (4.8, st.d. 1.4)
and in private letters (6.4, st.d. 0.6; cf. evaluation of state of mind in let-
ters, p. 311).

6.4.2. Modality

The following table serves to position the texts of the corpus with respect
to the role of modality. Besides mean frequencies and percentages, the
proportion of epistemic, perception and deontic modality in the total of
open-class adverbs and adjectives is reported in each text.14 The texts
are listed per period in order of increasing mean frequencies of modal ad-
verbs. Because of their infrequency, deontic adverbs have been omitted
from the table (see pp. 283-284), but their proportion can be calculated
by counting the difference between the total mean of modals and the
means of epistemic and perception modals. In addition, as the general fre-
quency of modal adjectives is low (see absolute numbers, p. 273), only
texts where at least 5% of the total of adverbs or adjectives are modal
have been included. Whatever the percentage of adjectives, texts with
mean frequencies lower than 0.5/1,000 have also been omitted.15
Texts reflecting higher frequencies typically represent philosophy, ser-
mons, education, trials and late science. Standard deviation is manifestly
higher in the mean frequencies of modal expressions as compared with
those conveying evaluations. The absence of modality characterizes a
number of texts (see note 15) in a way that — positively or negatively —
stresses generic features. In two-thirds of diaries and travelogues modality
is absent or rare. In contrast with other representatives of the genres, Fitz-
herbert and Langford in handbooks, Vicary and Record in scientific trea-
tises and Elyot and Hoole in educational treatises share the feature of not
using modal adjectives.
296 Anneli Meurman-Solin

Table 4. Mean frequencies and percentages of modal adverbs and adjectives in

the Early Modern English section of the Helsinki Corpus of English
Texts.

Adverbs Epistemic Perception

Mean % Mean Mean

1500-1570:
NI FICT MERRYTAL 0 0 0 0
STA LAW STAT3 0.2 6 0.1 0.1
NN TRAV LELAND 0.3 6 0 0.3
XX COME STEVENSO 0.3 18 0.3 0
XX COME UDALL 0.4 20 0.4 0
NN HIST MORERIC 0.5 5 0.5 0
NN BIO ROPER 0.6 5 0 0.6
NN HIST FABYAN 0.6 9 0.2 0.4
IR SERM LATIMER 0.8 24 0.8 0
NN ΒΙΑ MOWNTAYNE 0.9 16 0.7 0.2
XX TRI THROCKM 1.1 13 0.3 0.8
XX BIBLE TYNDALE 1.1 53 1.1 0
IS HANDO TURNER 1.2 12 0 1.2
XX CORO LET NON-PRIV 1.3 13 1 0.3
IR SERM FISHER 1.3 16 0.7 0.7
IS/EX EDUC ASCH 1.6 9 1.4 0.2
XX CORP LET PRIV 1.6 20 1.6 0
XX PHILO BOETHCO 3.3 45 3.1 0.2

Total N=26 0.7

Standard deviation 0.7

1570-1640:
IR SERM SMITH 0 0 0 0
XX TRI ESSEX 0 0 0 0
NN HIST HAYWARD 0 0 0 0
NN BIO PERROTT 0.2 3 0.2 0
NN BIA FORMAN 0.2 7 0.2 0
NN DIARY MADOX 0.3 7 0.3 0
NN HIST STOW 0.4 9 0.4 0
NI FICT ARMIN 0.4 9 0.4 0
XX COME MIDDLET 0.5 12 0.4 0.2
EX SCIM CLOWES 0.6 5 0.3 0.3
 Reconstructing a grammar of point of view 297

Table 4. Continued.

Adverbs Epistemic Perception

Mean % Mean Mean

EX SCIO BLUNDEV 0.6 15 0 0.6

XX TRI RALEIGH 0.6 19 0.6 0
XX CORP LET PRIV 0.8 15 0.7 0.1
XX COME SHAKESP 0.8 20 0.8 0
XX CORO LET NON-PRIV 0.9 14 0.5 0.4
IS HANDO MARKHAM 1 13 1 0
EX EDUC BACON 1.1 18 1.1 0
IS HANDO GIFFORD 1.1 33 1.1 0
XX BIBLE AUTHOLD 1.1 44 1.1 0
XX PHILO BOETHEL 1.3 23 0.6 0.4
IR SERM HOOKER 1.4 25 0.8 0.6
EX EDUC BRINSLEY 1.8 12 1.4 0.4
NI FICT DELONEY 1.8 29 1.8 0

Total N=27 0.6

Standard deviation 0.5

1640-1710:
IR SERM JETAYLOR 0.2 3 0.2 0
NN HIST MILTON 0.3 5 0.2 0.2
XX COME VANBR 0.4 7 0.4 0
NN DIARY PEPYS 0.4 11 0 0.2
NN BIO BURNETROC 0.5 4 0.5 0
STA LAW STAT7 0.5 11 0.5 0
IS EDUC LOCKE 0.6 5 0.6 0
IS HANDO WALTON 0.6 9 0.6 0
NN TRAV FRYER 0.6 9 0.6 0
Nl FICT BEHN 0.7 11 0.7 0
XX CORO LET NON-PRIV 0.9 13 0.7 0
XX CORP LET PRIV 1 17 1 0
XX TRI OATES 1 38 0.9 0.1
XX TRI LISLE 1.2 39 1.2 0
NI FICT PENNY 2.1 29 2.1 0
EX SCIO HOOKE 2.5 21 1 1.5
EX SCIO BOYLE 2.9 17 1.1 1.7
XX PHILO BOETHPR 3.3 34 1.7 0.6
IR SERM TILLOTS 3.3 35 2.6 0.8
298 Anneli Meurman-Solin

Table 4. Continued.

Adverbs Epistemic Perception

Mean % Mean Mean

Total N==26 0.9

Standard deviation 1

Total N==79 0.8 0.6 0.2

Standard deviation 0.8 0.6 0.3

Adjectives Epistemic Perception Deontic

Mean % Mean Mean Mean

1500-1570:
NI FICT MERRYTAL 1.2 6 1.2 0 0
STA LAW STAT3 0.7 3 0.3 0 0.4
NN TRAV LELAND 0 0 0 0 0
XX COME STEVENSO 2 13 1.8 0.2 0
XX COME UDALL 1.1 6 0.9 0.2 0
NN HIST MORERIC 2.1 7 1.9 0 0.2
NN BIO ROPER 1.9 6 1.3 0.4 0.2
NN HIST FABYAN 0.9 5 0.7 0.2 0
IR SERM LATIMER 1.6 5 0 0 1.6
NN ΒΙΑ MOWNTAYNE 1 6 0.7 0 0.3
XX TRI THROCKM 2.2 9 1.4 0.8 0
XX BIBLE TYNDALE 0.5 6 0.5 0 0
IS HANDO TURNER 1.2 2 0.4 0.4 0.4
XX CORO LET NON-PRIV 1.3 4 0.8 0.2 0.3
IR SERM FISHER 1.3 5 0.9 0.4 0
IS/EX EDUC ASCH 1 3 0.6 0 0.4
XX CORP LET PRIV 1.4 6 1.2 0 0.2
XX PHILO BOETHCO 2.8 7 1.4 0.9 0.5

Total N==26 1.1

Standard deviation 0.7

1570-1640:
IR SERM SMITH 1 5 0.8 0 0.2
XX TRI ESSEX 1.5 9 0.8 0.5 0.2
 Reconstructing a grammar of point of view 299

Table 4. Continued.

Adjectives Epistemic Perception Deontic

Mean % Mean Mean Mean

NN HIST HAYWARD 1.7 5 1.1 0.2 0.4

NN BIO PERROTT 1 5 0.4 0.2 0.4
NN BIA FORMAN 0 0 0 0 0
NN DIARY MADOX 0.8 4 0.3 0 0.5
NN HIST STOW 0 0 0 0 0
NI FICT ARMIN 0.8 3 0.4 0 0.4
XX COME MIDDLET 2 7 1.8 0 0.2
EX SCIM CLOWES 2 4 1.1 0.3 0.6
EX SCIO BLUNDEV 2.2 6 1.6 0 0.6
XX TRI RALEIGH 1.8 10 1.2 0.4 0.2
XX CORP LET PRIV 1.6 5 1.6 0 0
XX COME SHAKESP 1.5 8 1.3 0 0.2
XX CORO LET NON-PRIV 0.9 3 0.7 0 0.2
IS HANDO MARKHAM 1.3 3 1.1 0 0.2
EX EDUC BACON 2 5 1.4 0.2 0.4
IS HANDO GIFFORD 1.3 6 0.8 0.5 0
XX BIBLE AUTHOLD 0.8 12 0.7 0.1 0
XX PHILO BOETHEL 2.9 9 1.3 1.2 0.4
IR SERM HOOKER 2 9 1 0 1
EX EDUC BR1NSLEY 1.9 6 1.1 0.4 0.4
NI FICT DELONEY 1.1 5 1 0 0.1

Total N=27 1.3

Standard deviation 0.7

1640-1710:
IR SERM JETAYLOR 1.8 5 1 0.3 0.5
NN HIST MILTON 1.3 5 1 0 0.3
XX COME VANBR 1.1 4 1.1 0 0
NN DIARY PEPYS 2.6 10 1.8 0.2 0.6
NN BIO BURNETROC 3.1 10 2.1 0.2 0.8
STA LAW STAT7 0.5 2 0 0 0.5
IS EDUC LOCKE 4.1 10 3.1 0 1
IS HANDO WALTON 1.9 7 1.9 0 0
300 Anneli Meurman-Solin

Table 4. Continued.

Adjectives Epistemic Perception Deontic

Mean % Mean Mean Mean

NN TRAV FRYER 1.3 3 0.9 0 0.4

NI FICT BEHN 0.6 2 0.4 0 0.2
XX CORO LET NON-PRIV 4.1 17 2.9 0 1.2
XX CORP LET PRIV 1.4 6 1.4 0 0
XX TRI OATES 3.4 34 3.4 0 0
XX TRI LISLE 1.8 17 1.8 0 0
NI FICT PENNY 1.7 8 1.5 0 0.2
EX SCIO HOOKE 3.5 6 2.5 0.8 0.2
EX SCIO BOYLE 3.3 6 2.3 0.6 0.4
XX PHILO BOETHPR 2.4 8 1.1 1.2 0.1
IR SERM TILLOTS 6.4 16 4.2 1.4 0.8

Total N==26 2
Standard deviation 1.5

Total N==79 1.4 1 0.2 0.3

Standard deviation 1.1 0.8 0.3 0.3

About half of open-class adverbs are modal in Colville's Boethius transla-

tion and in the Bible, about one-third in Gifford's handbook, Tillotson's
sermon and the two late-seventeenth-century trials. Modality is not central
to histories, travelogues, diaries, biographies and fiction (excluding Delo-
ney and Penny Merriments·, see Meurman-Solin 1995). This adverbial role
becomes more prominent over time in science, so that Vicary and Record
have almost no modal adverbs, Clowes and Blundevile 0.6 (5 and 15%
respectively), Boyle 2.9 (17%), and Hooke 2.5 (21%).
Table 4 also shows that modality is mostly epistemic in the texts of the
corpus, and that this subrole is most prominent in texts written in dia-
logue form or in texts explicitly addressed to a particular person (letters)
or audience (sermons). Also adjectives of epistemic modality are mani-
festly typical of texts labelled as interactive in the corpus; particularly
uniform practices have been attested in private letters (1.4, st.d. 0.2) and
in philosophy (1.3, st.d. 0.1); somewhat less homogeneous are comedies
(1.3, st.d. 0.5) and trials (1.7, st.d. 1).
 Reconstructing a grammar of point of view 301

Adverbs of perception modality represent 10-15% of the total of open-

class adverbs in Turner's handbook, in Blundevile's, Boyle's and Hooke's
scientific treatises and in Hooker's sermon, and the mean frequencies are
also clearly higher than the total mean. These adverbs (for example plain-
ly, openly, clearly, manifestly and evidently) typically occur with verbs
such as see, perceive, know, prove, show, manifest, demonstrate, discover,
find and distinguish, and they are also used in explicit comments of the
author (as shall plainely appeare hereafter when wee come to describe
the same, Blundevile: 48V), or in a direct address to the reader (as by pe-
rusing the said tables you may plainely see, Blundevile: 50R, Whereby
you may plainely perceive that..., Blundevile: 49V, and as you may evi-
dently discerne it, Brinsley: 43).
Perception modality realised by means of adjectives is a prevalent fea-
ture in the Boethius translations (1.1, st.d. 0.2). In addition, the validity of
perceptions is stressed in the trial of Throckmorton, in the two late-seven-
teenth-century scientific treatises, Boyle and Hooke, and in Tillotson's
sermon.
It is only in educational treatises that there is some similarity in the
mean frequencies of deontic modal adjectives (0.6, st.d. 0.3); other genres
are very heterogeneous in this respect. Yet deontic modality characterizes
three sermons (Latimer, Hooker and Tillotson). Adverb realizations in this
subrole have been attested only in the Boethius translations by Elizabeth
I and Preston, in Pepys's diary and in the late-seventeeth-century official
letters.

6.4.3. Intensifying and focusing adverbs and adjectives

It was further found appropriate to form separate groups of adverbs func-

tioning as intensifying and focusing subjuncts. The two are not always
differentiated (cf. Halliday 1985: 82), but, on the basis of the present
findings, they do not correlate significantly. The former play an important
role as markers of involvement (see Biber 1988: 106, 240); by means of
the latter the author can guide the reader's attention (additionally, espe-
cially or exclusively) to a particular thing. Focusing devices are assumed
to correlate with the degree of interactiveness (cf. Nevalainen 1991: 76-
79). Especially letters contain examples of conventionalized uses of inten-
sifiers:
302 Anneli Meurman-Solin

(19) Right worshipfull Sir, in my most hardest wyse I recomend me

unto you, desiring to here of your prosperytie and welfare, and
good spede in your matters, the which I marvell greatly that I
have no word from you. Sir, I marvell greatly that ye let the
matter rest so long, and labors no better for your selfe, and ye
wold labor it deligently. (Robert Plumpton, Plumpton Corre-
spondence 186)

The distributions of intensifiers are recorded in Table 5:

Table 5. Mean frequencies and percentages of intensifying adverbs in -ly and

adjectives in the Early Modern English part of the Helsinki Corpus
of English Texts.

Intensifying
Adverbs Adjectives
Mean % Mean %

1500-1570:
EX SCIO RECORD 0 0 0 0
NN DIA MACHYN 0 0 0 0
IS HANDO FITZH 0 0 0.2 1
IR SERM LATIMER 0 0 0.4 1
XX BIBLE TYNDALE 0 0 0.2 3
NI FICT HARMAN 0.2 2 0 0
NN DIARY EDWARD 0.2 8 0.5 2
XX COME STEVENSO 0.2 9 0.5 3
NN ΒΙΑ MOWNTAYNE 0.3 6 0 0
NN TRAV TORKINGT 0.4 13 1.2 7
XX TRI THROCKM 0.6 7 0.6 2
XX PHILO BOETHCO 0.6 8 3.2 8
NN HIST FABYAN 0.6 9 0.6 3
STA LAW STAT3 0.7 24 2.9 12
XX CORP LET PRIV 0.8 9 0.8 3
NI FICT MERRYTAL 0.8 24 0.2 1
NN BIO ROPER 0.9 8 1.5 5
IS/EX EDU ELYOT 0.9 9 1.6 4
NN TRAV LELAND 0.9 19 0.1 1
XX COME UDALL 0.9 40 0.2 1
EX SCIM VICARY 1 21 0.2 0
XX CORO LET NON-PRIV 1.1 11 1.1 4
NN HIST MORERIC 1.2 12 0.5 2
 Reconstructing a grammar of point of view 303

Table 5. Continued.

Intensifying
Adverbs Adjectives
Mean % Mean %

IS HANDO TURNER 1.4 14 0.6 1

IR SERM FISHER 1.6 18 4 16
IS/EX EDU ASCH 1.8 10 0.6 2

Total N=26 0.7 0.8

Standard deviation 0.5 1.1

1570-1640:
IR SERM SMITH 0 0 0 0
NN DIARY HOBY 0 0 0 0
XX COME MIDDLET 0 0 0.4 1
STA LAW STAT4 0.1 9 0.5 2
NN HIST HAYWARD 0.2 2 0.6 2
NI FICT ARMIN 0.2 4 0 0
IR SERM HOOKER 0.2 4 0.4 2
XX TRI ESSEX 0.2 5 0.5 3
XX TRI RALEIGH 0.2 7 0.2 1
NN ΒΙΑ FORMAN 0.2 7 0.5 3
NN TRAV COVERTE 0.3 4 0.3 1
XX CORP LET PRIV 0.3 5 1.4 4
EX SCIO BLUNDEV 0.3 8 0 0
XX COME SHAKESP 0.3 8 0.6 4
XX BIB AUTHOLD 0.3 13 0 1
XX PHILO BOETHEL 0.4 8 0.6 2
XX CORO LET NON-PRIV 0.5 8 0.2 1
EX EDU BACON 0.5 9 0.2 0
NN DIARY MADOX 0.5 11 0 0
EX SCIM CLOWES 0.6 5 0.3 1
NN TRAV JOTAYLOR 0.6 12 0.1 0
IS HANDO GIFFORD 0.6 19 0.5 2
EX EDU BRINSLEY 0.7 5 0.2 1
NN BIO PERROTT 0.8 12 0 0
NN HIST STOW 0.8 18 0.2 1
NI FICT DELONEY 1 16 0.7 3
IS HANDO MARKHAM 1.5 20 0.8 2
304 Anneli Meurman-Solin

Table 5. Continued.

Intensifying
Adverbs Adjectives
Mean % Mean %

Total N=21 0.4 0.3

Standard deviation 0.3 0.3

1640-1710:
XX TRI OATES 0 0 0 0
IS HANDO WALTON 0 0 0.2 1
NN BIA FOX 0 0 0.4 2
NN HIST MILTON 0 0 0.5 2
XX TRI LISLE 0 0 0.5 5
STA LAW STAT7 0.1 2 0.2 1
NI FICT PENNY 0.2 2 0 0
NN TRAV FRYER 0.2 3 1.1 3
XX COME FARQUHAR 0.2 8 0.5 2
XX COME VANBR 0.6 10 0.1 1
NN TRAV FIENNES 0.6 14 0.2 0
IS EDUC HOOLE 0.7 6 0 0
XX CORO LET NON-PRIV 0.7 11 0.7 3
IS HANDO LANGF 0.8 16 0.3 1
NN DIARY PEPYS 0.8 22 0.2 1
IR SERM TILLOTS 0.9 10 1.4 3
XX CORP LET PRIV 0.9 16 0.3 1
NN HIST BURNETCHA 0.9 19 0.2 1
EX SCIO HOOKE 1 8 0.3 1
IS EDUC LOCKE 1 9 1.2 3
XX PHILO BOETHPR 1 10 1.7 6
IR SERM JETAYLOR 1.2 21 0.7 2
NN DIARY EVELYN 1.3 18 0.8 2
EX SCIO BOYLE 1.7 10 1.1 2
NI FICT BEHN 1.8 29 0.9 3
NN BIO BURNETROC 2.3 20 0.8 3

Total N:=26 0.7 0.6

Standard deviation 0.6 0.5

Total N==79 0.6 0.6

Standard deviation 0.5 0.7
 Reconstructing a grammar of point of view 305

The table shows that, in a higher degree than in other texts, intensification
is realized by means of adverbs in Aphra Behn's fiction, in Burnet's bio-
graphy and in Fisher's sermon (both frequency- and percentage-wise), but
also in Turner's and Markham's handbooks, in Ascham and in Boyle. In
Behn the proportion of intensifiers exceeds that of evaluative expressions,
which is exceptional in the corpus.
As regards normalized mean frequencies of intensifying adverbs in
genres, there is a higher frequency of them particularly in early-sixteenth-
century education (1st subperiod: 1.4, st.d. 0.7; cf. 2nd: 0.6, st.d. 0.1; 3rd:
0.8, st.d. 0.2) and in contemporary official letters (1st subperiod: 1.1; cf.
2nd: 0.5 and 3rd: 0.7). While in sixteenth-century non-medical science the
feature is virtually absent, the texts representing the last subperiod have
a mean frequency of 1.4 (st.d. 0.5; cf. evaluation in science pp. 294-295).
Intensifying adverbs and adjectives correlate highly significantly (p=
.001). Adjective realizations do however serve to identify a number of
texts where intensification is integrated in noun phrases, and adverbs in
this subrole are less prominent (for example, early Statutes of the Realm
and Colville's and Preston's Boethius translations). Fisher's sermon best
illustrates this correlation pattern. Moreover, in contrast to the attested
correlation between adjective and adverb realizations, the proportion of
adverbs, as compared with that of adjectives, is striking in Vicary (adv.
21%, adj. 0%) and in four of the six handbooks (Turner: 14% versus 1%,
Gifford: 19% versus 2%, Markham: 20% versus 2% and Langford 16%
versus 1%). A similar difference can be illustrated by numerous texts re-
presenting non-imaginative narration such as travelogues, histories, bio-
graphies and diaries, where there are 15-20 percentage units more of ad-
verbs than of adjectives (of the totals of the two word-classes). These dif-
ferences reflect the tendency, attested in a number of genres, to integrate
descriptive information in noun phrase structures and to express the
author's or the speaker's comments by means of adverbs (see further pp.
313-315, 328-329).
Intensifying and focusing subjuncts do not correlate significantly (p=
.057). The following table positions the texts on a continuum on the basis
of the mean frequency and percentage of focusing adverbs (in the first
two columns):
306 Anneli Meurman-Solin

Table 6. Mean frequencies and percentages of focusing adverbs in -ly and ad-
jectives in the Early Modern English part of the Helsinki Corpus of
English Texts.

Focusing
Adverbs Adjectives
Mean % Mean %

1500-1570:
XX COME STEVENSO 0 0 0 0
XX COME UDALL 0 0 0 0
XX BIB TYNDALE 0 0 0.3 4
NN DIA MACHYN 0 0 0.6 5
NN DIARY EDWARD 0 0 2.1 10
EX SCIO RECORD 0.1 4 1 3
XX CORP LET PRIV 0.2 2 1 5
NN HIST MORERIC 0.2 2 1.1 4
NN HIST FABYAN 0.2 3 1.1 6
NI FICT MERRYTAL 0.2 5 0.9 5
IR SERM LATIMER 0.2 6 0.6 2
STA LAW STAT3 0.2 6 0.8 3
XX CORO LET NON-PRIV 0.3 3 1.6 5
XX PHILO BOETHCO 0.3 4 0.9 2
NN ΒΙΑ MOWNTAYNE 0.3 6 0 0
EX SCIM VICARY 0.3 7 1.8 4
IS/EX EDUC ASCH 0.4 2 0.4 1
IS HANDO TURNER 0.4 4 0.4 1
NI FICT HARMAN 0.4 4 0.6 2
NN BIO ROPER 0.6 5 2 6
NN TRAV LELAND 0.6 13 1.6 6
IR SERM FISHER 0.7 8 1.8 7
IS HANDO FITZH 0.8 16 0.4 1
NN TRAV TORKINGT 0.8 25 0.7 4
IS/EX EDU ELYOT 1.1 11 1.6 4
XX TRI THROCKM 1.1 13 0.9 4

Total N=26 0.4 0.9

Standard deviation 0.3 0.6
 Reconstructing a grammar of point of view 307

Table 6. Continued.

Focusing
Adverbs Adjectives
Mean % Mean %

1570-1640:
NN DIARY HOBY 0 0 0 0
XX BIB AUTHOLD 0 0 0.2 3
XX TRI RALEIGH 0 0 0.8 5
STA LAW STAT4 0 0 1.1 4
IR SERM HOOKER 0 0 1.2 5
IS HANDO MARKHAM 0 0 1.6 3
NI FICT DELONEY 0.1 2 0.3 1
NN TRAV JOTAYLOR 0.1 2 0.6 2
XX PHILO BOETHEL 0.1 3 0.4 1
NN TRAV COVERTE 0.2 2 1 4
XX CORO LET NON-PRIV 0.2 3 1.6 6
NI FICT ARMIN 0.2 4 0.2 1
XX COME MIDDLET 0.2 4 0.2 1
EX SCIO BLUNDEV 0.2 4 4.4 11
XX TRI ESSEX 0.2 5 0.7 4
IS HANDO GIFFORD 0.2 5 1 4
NN HIST STOW 0.2 5 1 6
IR SERM SMITH 0.2 6 0.2 1
XX CORP LET PRIV 0.3 5 0.6 2
NN BIO PERROTT 0.4 6 1.2 6
NN DIARY MADOX 0.5 11 0.2 1
EX EDUC BACON 0.7 12 1.2 3
EX SCIM CLOWES 0.9 8 2.7 5
NN HIST HAYWARD 0.9 11 1.9 5
NN BIA FORMAN 1 27 1.2 8
EX EDUC BRINSLEY 1.6 11 2.5 8

Total N==27 0.3 1

Standard deviation 0.4 1
308 Anneli Meurman-Solin

Table 6. Continued.

Focusing
Adverbs Adjectives
Mean % Mean %

1640-1710:
NN ΒΙΑ FOX 0 0 0 0
XX PHILO BOETHPR 0 0 0.3 1
XX COME VANBR 0 0 0.7 3
IR SERM TILLOTS 0 0 0.8 2
IR SERM JETAYLOR 0.2 3 0.7 2
STA LAW STAT7 0.2 5 1.8 7
NN DIARY PEP YS 0.2 6 0.2 1
XX TRI LISLE 0.2 6 0.5 5
XX COME FARQUHAR 0.2 8 0 0
EX SCIO HOOKE 0.3 3 1.7 3
XX TRI OATES 0.3 10 0.6 6
NN BIO BURNETROC 0.5 4 0.5 2
NI FICT PENNY 0.5 6 0.2 1
NN HIST MILTON 0.5 7 0.9 3
XX CORO LET NON-PRIV 0.5 8 1.2 5
NN HIST BURNETCHA 0.5 12 0.7 2
NN TRAV FRYER 0.6 9 0.6 1
XX CORP LET PRIV 0.6 11 0.2 1
NN TRAV FIENNES 0.6 14 0 0
EX SCIO BOYLE 0.8 4 2.7 5
NN DIARY EVELYN 0.8 11 0.3 1
NI FICT BEHN 0.9 14 0.5 2
IS EDUC LOCKE 1.2 11 0.8 2
IS HANDO LANGF 1.2 23 0.8 2
IS HANDO WALTON 1.3 21 0.6 2
IS EDUC HOOLE 1.8 17 1.6 4

Total N==26 0.5 0.7

Standard deviation 0.5 0.6

Total N==79 0.4 0.9

Standard deviation 0.4 0.8
 Reconstructing a grammar of point of view 309

More than 20% of the total of open-class adverbs are focusing in Tor-
kington's travelogue, Forman's autobiography and in Walton and Lang-
ford, the two late-seventeenth-century handbooks. The proportion of fo-
cusing subjuncts may be a genre-related feature, as higher percentages
have been attested also in Leland's and Fiennes' travelogues and in Fitz-
herbert's handbook. Educational treatises share the feature of manipulat-
ing the reader's attention by using focusing subjuncts, so that their pro-
portion ranges from 11% to 17%, except in Ascham (only 2%). Also the
mean frequencies are usually high, and there is evidence of some degree
of uniformity (1.1, st.d. 0.5). Another rather homogeneous genre is bio-
graphies (mean of focusing adverbs 0.5, st.d. 0.07). In addition, the fea-
ture is quite prominent in the trial of Throckmorton, and well above the
total mean in Clowes, Boyle, Evelyn, Hayward and Behn.
From a diachronic perspective, there is however a manifest increase of
focusing subjuncts and a trend towards a higher degree of uniformity in
educational treatises (1st subperiod: 0.7, st.d. 0.5; 2nd: 1.2, st.d. 0.7; 3rd:
1.5, st.d. 0.5) and in handbooks (1st: 0.6, st.d. 0.3; 2nd: 0.1, st.d. 0.1; 3rd:
1.2, st.d. 0.1). In contrast, the trial of Throckmorton contains considerably
more focusing subjuncts than the later trials (1.1; cf. 2nd: 0.1, st.d. 0.1;
3rd: 0.2, st.d. 0.1). While the role of modality is dominant especially in
later trials, evaluative expressions and focusing subjuncts (see Tables 3
and 6) characterize the trial of Throckmorton, which dates from 1554.
As regards adjective realizations, the focus-marking function is clearly
prevalent in science, more uniformly in science/medicine (2.2, st.d. 0.7)
than in science/other (2.4, st.d. 1.5). Also official letters are homogeneous
in this respect (1.5, st.d. 0.2; cf. private letters: 0.6, st.d. 0.4). Mean fre-
quencies are particularly high in Blundevile (4.4; 11% of all adjectives),
in Clowes (2.7; 5%) and Boyle (2.7; 5%). As compared with adjectives,
the proportion of adverb realizations of the total of adverbs is consid-
erably higher in Fitzherbert (adv. 16%, adj. 1%), Torkington (adv. 25%,
adj. 4%), Walton (adv. 21%, adj. 2%), Fiennes (adv. 14%, adj. 0%) and
Langford (adv. 23%, adj. 2%) (cf. intensifying adverbs and adjectives, p.
305). It seems that handbooks and travelogues tend to mark the focus by
means of adverbs, reserving modifier functions in noun phrases for de-
scriptive adjectives (cf. p. 288). Also the findings for Behn (adv. 14%,
adj. 2%), Bacon (adv. 12%, adj. 3%), Locke (adv. 11%, adj. 2%) and
Hoole (adv. 17%, adj. 4%) reflect the importance of adverb realizations.
Among educational treatises, Elyot, Ascham, Brinsley and Hoole show
higher mean frequencies of focusing adjectives than the other representa-
tives of the genre.
310 Anneli Meurman-Solin

Highly significant correlations have been attested between intensifying,

focusing and evaluative adverbs (p=.000). This evidence stresses the
relevance of co-occurrence patterns; the cumulative effect of features adds
to the statistical significance of the findings, but, at the same time, no in-
formation is lost, as idiolectal preferences in the marking of point of view
are retrievable.

7. Orientation and position of evaluative adverbs

In the analysis of evaluative adverbs, information was recorded about

orientation, scope and position (pp. 271-272). The following findings are
very preliminary, as the corpus is not always large enough for conclu-
sions on points of this kind. The majority of these adverbs were found to
be predicate-oriented, which demonstrates their syntactic affinity to pro-
cess adjuncts that are semantically descriptive. The correlation between
predicate-oriented evaluative adverbs and descriptive adverbs is signifi-
cant (p=.002). The mean frequency of predicate-oriented evaluative ad-
verbs is 1.5, that of subject-oriented evaluative adverbs 1 and of disjuncts
0.3.
In addition to our earlier discussion of evaluative adverbs in texts (pp.
290-295), the distributions of subject-oriented evaluative adverbs and dis-
juncts across genres seem particularly relevant. Subject-oriented adverbs
are typical of biographies (2.4, st.d. 1.3), official letters (2, st.d. 1.4), later
science/medicine (i.e. Clowes, 2.1) and private letters (1.8, st.d. 0.8), the
most uniform genre in this respect. In the first two genres there are more
subject-oriented than predicate-oriented evaluative adverbs. Despite the
relatively high mean frequency in education (1.4), standard deviation is
too high (1.2) to interpret the frequency as being conditioned by genre.
Deviation in biographies and official letters can be explained by diachron-
ic developments, texts dating from 1500-1570 showing considerably
higher mean frequencies than the later texts (3.9 in Roper's biography,
3.5 in official letters). Further study will be necessary to confirm a similar
decrease in private letters (1st subperiod: 2.5; 2nd: 1.8; 3rd: 0.9). Also in
trials, the earliest (i.e. Throckmorton, mean 2.6) manifestly differs from
later ones (2nd subperiod: 0.9, st.d. 0.4; 3rd: 0.8, st.d. 0.2).
Subject-oriented evaluative adverbs correlate highly significantly (p=
.000) with adjectives referring to state of mind (e.g. happy, humble,
 Reconstructing a grammar of point of view 311

eager). In contrast, there is only an almost significant correlation between

the latter and the total of evaluative adjectives (p=040). Adjectives refer-
ring to state of mind (total mean 2, st.d. 1.9) turned out to be a weighty
diagnostic feature in the identification of similarities between interactive
genres (private letters: 7.3, st.d. 3.1; official letters: 3.5, st.d. 1; and
comedies: 3.7, st.d. 1.1). Biography (2.6, st.d. 0.1) is the most uniform
genre in this respect. These adjectives also usefully reveal important dif-
ferences within genres. The point of view in More's and Burnet's his-
tories seems to be very much on the feelings of the protagonists in the
course of events, and we find other protagonist-centred texts sharing this
feature (Mowntayne's autobiography and the three biographies in the cor-
pus). Of other non-interactive texts, Aphra Behn's fiction stands out (see
also Meurman-Solin 1995).16
It is in science that evaluative adverbs most conspicuously take the
function of an independent proposition. It is significant that it should be
the same text, Clowes's medical treatise, where both subject-oriented
evaluative adverbs and peripheral uses of these adverbs co-occur in ex-
ceptionally high frequencies; the mean of disjuncts in Clowes is 1.2 (cf.
the total mean 0.3), that of subject-oriented adverbs 2.1 (cf. the total
mean 1). Also in Record, where viewpoint marking is very marginal, the
mean of disjuncts is 0.6. Another early text, Turner's handbook, marked
by generic time reference and perception modality (see pp. 289-290 and
p. 300), contains examples of evaluative adverbs as disjuncts (mean 0.8).
Otherwise, texts where the author's comments are expressed by means of
sentence adverbs are mostly those where point of view is more explicitly
expressed than in the average (for example Ascham 1, Roper 0.9, Hay-
ward 0.8; cf. however the positioning of disjuncts, p. 312). The uniform-
ity in official letters (0.5, st.d. 0.1) perhaps suggests that the feature is
genre-specific (see also p. 319 on orientation and text category).
Initially-positioned evaluative adverbs do not correlate with those posi-
tioned medially, but both initial and medial positions correlate almost sig-
nificantly with final position (p=.029 and .019 respectively). Predicate-
oriented evaluative adverbs are mostly positioned medially (0.9), but they
also occur in post-verb positions (0.5) and very rarely in initial position
(0.1). Similar distributions were attested in subject-oriented adverbs (me-
dial 0.6, final 0.3, initial 0.05) and in disjuncts (medial 0.2, final 0.02,
initial 0.04).17 Initially-positioned predicate-oriented evaluative adverbs
were mostly found in sermons, philosophy, Clowes and Record. Medial
and final positions of predicate- and subject-oriented evaluative adverbs
compare as follows: among texts where the normalized mean frequency
312 Anneli Meurman-Solin

is at least 1, the proportion of final position is 63% in Latimer's and 90%

in Smith's sermon and 64% in Brinsley. The final position in Elyot (54%)
and Ascham (40%) relates to their high frequency of intransitive verbs
(see p. 316).
In the light of general frequencies, the textual role of expressing the
author's or the speaker's evaluation does not seem to lead to a preference
of initial position in sentence structure. However, when we look at the
distributions of initially-positioned evaluative adverbs in genres, we notice
that initially-positioned disjuncts are strikingly frequent in Clowes's medi-
cal treatise (0.5, total mean 0.04), and well above the average also in
philosophy, autobiography, education and history (all of them 0.1). The
development of adverbial positions in science is interesting. In Vicary's
medical treatise no evaluative adverbs functioning as disjuncts occur in
initial position; instead, he prefers medial (0.3) or final (0.2). Similarly
Record always puts disjuncts in medial position (0.6), whereas later texts
representing science/other sometimes choose the initial position (0.1; cf.
medial 0.4, final 0). There is yet another distinctive feature in Clowes's
grammar: predicate-oriented evaluative adverbs are put in final position
more often (1.2; total mean 0.5) than in the other texts of the corpus, with
the exception of educational treatises (1.7).
The position of adverbs of epistemic modality is somewhat more fre-
quently initial (0.3) than medial (0.2) but rarely after the verb (0.05). The
initial position is most clearly preferred in the Bible (1, cf. medial 0.1)
and comedies (0.3; medial 0.03). The initial position is chosen somewhat
more often than the medial position also in private letters (0.7; medial
0.3) in philosophy (0.9; medial 0.7) and in fiction (0.6; medial 0.2). In
contrast, medial position is preferred in science/other (0.5; initial 0.2), in
sermons (0.4; initial 0.3) and in trials (0.4; initial 0.2). Even though the
frequencies are low, it is perhaps interesting that all epistemic modals are
positioned medially in law. The Boethius translations provide useful ma-
terial for assessing developments over time (see Table 7):

Table 7. Adverbs of epistemic modality in three translations of Boethius' De

Consolatione Philosophiae. Means of normalized frequencies.

Text Initial Medial Final Mean of the subperiod

Colville 2.2 0.6 0.3 0.4 0.1 0.04

Elizabeth I 0.1 0.4 0 0.3 0.2 0.06
Preston 0.5 1.1 0.1 0.2 0.4 0.05
 Reconstructing a grammar of point of view 313

The direction of change in these idiolectal practices is from initially-posi-

tioned to medially-positioned epistemic adverbs. A similar change has
been attested in sermons, where the initial and medial positions represent
mean frequencies of 0.5 and 0.2 respectively in the period 1500-1570, 0.3
and 1 in the period 1640-1710.
Perception modals prefer the medial position in the three time periods
(mean 0.1; initial 0.01, final 0.03). The initial position is very rare except
when fronted because of a comparative element (but more plainly in my
booke of measuring you may see it, Record: B4R), whereas the post-verb
position dominates in Colville's and Preston's Boethius translation (both
0.2), in Vicary (0.2; the only position of perception modals in Vicary), in
the trial of Throckmorton (0.4), and in Fisher's and Hooker's sermons.
This is however due to the frequency of the expression 'it appears that ...'
in these texts (e.g. Whereby it appereth clerely, Boethius tr. by Colville:
77, it appeareth openly, Vicary: 68, it may appeare yet more euidently,
Throckmorton: I,69.C2). Deontic modals are positioned medially, except
in one example: Wherfore necessarily it must be confest, that..., Boethius
tr. by Elizabeth I: 58). As stated, the evidence is not quantitatively re-
presentative, but may shed light on how to expand the corpus for further
study.

8. Adjectives as modifiers and complements

The total mean frequencies of attributive adjectives are consistently higher

than those of complements (for example evaluative adjectives: modifiers,
mean 5.1, st.d. 3.2; complements, mean 2.7, st.d. 1.8, and descriptive ad-
jectives: modifiers, 9.6, st.d. 7; complements, 3.8, st.d. 3.1). Only in
Preston's Boethius translation (mean of complements 6.3), the Bible (1.2
and 1), Smith's sermon (1.9), Vicary (1.9), Hayward (6.1), Hoby (2.3),
Langford (4.2), Hooke (4) and the late-seventeenth-century official letters
(3.7) are there more evaluative adjectives as complements than as modi-
fiers (cf. Table 3). It is interesting that the very few evaluative adjectives
attested in Lady Hoby's diary mostly function as complements (see also
note 18).
The prevalence of adjectives as complements and the relative infre-
quency of nominalizations are related to a more fragmented way of pro-
cessing information, which results in low informational density in texts.
314 Anneli Meurman-Solin

Evaluation and modality integrated in noun phrase structures are com-

municatively less dynamic (cf. Quirk et al. 1985: 1239-1243 and Biber
1988: 237) than the expression of attitudes in complement positions. In
the function of adnominal modifier, an adjective tends to be interpreted as
non-restrictive. Consequently, an attitudinal expression in the premodifi-
cation (an easy task) can take a classifying or descriptive function in con-
trast with adjectives as complements (The task is easy), which convey the
point of the message and function as direct expressions of the speaker's
opinion. To use an example from the Helsinki Scots Corpus (Ireland, Me-
roure ofWyssdome·. 14), in "xij nobile pillaris [the xij artiklis of the faith]
mare precius nore guide are precius stanis" precius in complement posi-
tion conveys the writer's attitude, whereas the attributive use of precius
functions as a classifier in an established compound. In fact, standard de-
viation is lower in the frequencies of descriptive adjectives as modifiers
(see above), whereas the diagnosticity of complements is stressed by the
attested polarization between texts with uniformly low frequencies of
complements and those with exceptionally high frequencies represented
by handbooks (8.9, st.d. 4.7), science/medicine (9.7, st.d. 7.3) and
science/other (8.5, st.d. 2.9). Further study will be necessary to provide
more information about degrees of structural complexity in texts.
As illustrated in Table 2, the use of descriptive adjectives as comple-
ments suggests, in the period 1500-1570, a grouping of a number of
interactive texts, namely letters and sermons, and non-imaginative narra-
tion represented by diaries and Fabyan's history. In addition, the two
early educational treatises and, less manifestly, Turner and Record, are
shown to reflect usages of a particular text type. In the second period,
Stow's history, Perrott's biography and Lady Hoby's diary clearly belong
to the same text type; this is true also of letters and Blundevile's and
Clowes's scientific treatises. The preference of complements can also be
illustrated by the representatives of these genres in the last subperiod. Of
other texts in 1640-1710, Hoole, the two trials and Behn stand out, to
mention just a few examples. In all three periods also statutes reflect a
tendency to use descriptive adjectives as complements. Time- and space-
related texts such as histories, travelogues and diaries vary in information-
al density (Meurman-Solin 1994), but it is important to notice that there
is evidence for establishing a text type where descriptive elements are
typically not integrated in noun phrase structures.
Adjectives expressing epistemic modality occur chiefly as complements,
with the exception of official letters, some sermons and Blundevile's and
Hooke's scientific treatises, where the use as prenominal modifier domi-
 Reconstructing a grammar of point of view 315

nates. The prevalence of complements is also true of perception and de-

ontic modality; it is only in the Bible, especially in Tyndale, that we find
a higher proportion of adjectives in these two subroles as prenominal
modifiers than as complements.
In the case of adjectives referring to state of mind, complements are
prevalent especially in fiction, comedies, philosophy and sermons, modi-
fiers in letters. 18 The dominance of modifiers in letters can be explained
by the numerous epistolary formulae where adjectives of this kind have
become petrified expressions in polite language use, so that their role as
viewpoint markers has been weakened.

9. Evaluative adverbs as related to voice and transitivity

As stated earlier (see p. 272), a classification of verbs co-occurring with

evaluative adverbs mostly resulted in too low frequencies per verb catego-
ry to give statistically relevant evidence. Despite the obvious problems re-
lated to the interpretation of evidence selected solely by quantitative cri-
teria, an analysis was performed on the basis of the most numerously re-
presented category, verbs of action. 19 All verbs of action co-occurring
with evaluative adverbs were in the passive in Vicary's and Blundevile's
scientific treatises, in Stevenson's comedy, Forman's autobiography, and
the trial of Oates. Only in law is the mean frequency of passive verbs
higher than that of active verbs. Passive verbs are particularly frequent in
Ascham, Boyle and Locke; nineteen out of a total of 79 texts contain
more passive than active verbs with evaluative adverbs. 20 In contrast
with the passive voice, there is a highly significant correlation (p=.000)
between the active voice and subject-oriented evaluative adverbs.
Diachronic developments in the mean frequency of passive verbs are
perhaps of interest in cases where they cannot be explained only with
reference to similar changes in the frequency of evaluative adverbs. The
highest mean frequencies of passive verbs were attested in educational
treatises and handbooks in the period 1500-1570; the decrease over time
is chiefly due to a high frequency of passives in Ascham in the former,
and to the avoidance of passives in Gifford and Langford in the latter. 21
Further study will be necessary to explain the change over time in the
frequency of passive verbs in travelogues (0.6, st.d. 0.1; 0.6, st.d. 0.7; 1.2,
316 A nneli Meurman-Solin

st.d. 0.6). As regards text categories, passive verbs are more frequent in
secular instruction (Bacon, Locke, Turner and Walton), non-imaginative
narration (Leland, Fiennes, Fryer, Stow and Edward VI), and in exposi-
tory (Vicary, Record and Blundevile) and statutory texts (statutes dating
from the first two subperiods).
In the dimension transitive versus intransitive, Elyot's Gouernour stands
out as the only text where more than half (58%) of the verbs adjoining
evaluative adverbs are intransitive (2.5 against a total mean of 0.4).22 As
regards genre, mean frequencies reflect no homogeneous patterns in the
use of intransitive verbs in this specific context. From a diachronic per-
spective, the most relevant finding is the importance of intransitive verbs
in early educational treatises as compared with later representatives of the
text type (2.1, st.d. 0.6; 0.4, st.d. 0.5; 0.5, st.d. 0.4). The difference
between Vicary and Clowes is highlighted by the exceptionally high fre-
quency of intransitive verbs in the latter (Vicary 0, Clowes 1.2). Transi-
tive verbs are more uniformly prevalent in biographies (2.5, st.d. 0.8),
education (3, st.d. 1.4) and official letters (1.9, st.d. 0.6) than in other
genres.
Subject-oriented evaluative adverbs are usually positioned medially with
transitive verbs except in the trial of Throckmorton. A case in point is of-
ficial letters, where in 1500-1570 medial position with active transitive
verbs gives a mean of 1.7, as compared with final position 0.2. Medially-
positioned disjuncts occur more frequently with passive than with active
transitive verbs. There is also evidence that predicate-oriented evaluative
adverbs favour medial position with passive verbs. This can be illustrated
for example by Clowes, where the mean for predicate-oriented evaluative
adverbs both in medial and in final position is 0.5 with active transitive
verbs; in contrast, all evaluative adverbs are positioned medially (1.4)
with passive transitive verbs.

10. Genre and text category as conditioning factors

10.1. Point of view in genres

The clustering of viewpoint markers, evaluative, modal (including habitu-

al and generic time reference), intensifying and focusing adverbs and ad-
jectives becomes evident when we study the correlations between the fea-
 Reconstructing a grammar of point of view 317

tures. Adverbs and adjectives in the same semantic role correlate highly
significantly, with the exception of focusing adverbs and adjectives. High-
ly significant correlations have also been attested for example between
evaluative adverbs and adjectives with habitual and generic time refer-
ence, focusing adverbs and intensifying adverbs, between modal adverbs
and intensifying adjectives, between evaluative and modal adjectives, be-
tween locative and descriptive adjectives, between adverbs of perception
modality and descriptive and focusing adjectives, and between subject-
oriented evaluative adverbs and adjectives referring to state of mind (for
these and other correlations, see Appendix 2). Table 8 illustrates the pro-
portion of attitudinal expressions in genres:

Table 8. Percentages of markers of point of view in genres in the Early Modern

English part of the Helsinki Corpus. Standard deviation in brackets.
N=number of texts.

Genre Ν Markers of point of view

Science/other 4 28 (14)
Science/medicine 2 34 (22)
Law 3 36 (5)
Diaries 6 40 (20)
Travelogue 6 40 (13)
Handbooks 6 42 (15)
Bible 2 50 (2)
History 6 52 (13)
Autobiography 3 53 (12)
Fiction 6 57 (10)
Biography 3 63 (6)
Trial proceedings 5 63 (5)
Sermons 6 64 (14)
Comedies 6 64 (2)
Official letters 3 67 (8)
Education 6 68 (7)
Private letters 3 69 (9)
Philosophy 3 69 (2)

Total 79 53 (16)
318 A nneli Meurman-Solin

Despite the high general frequency of adverbs and adjectives in scientific

treatises (see Table 1), the proportion of markers of point of view is the
lowest in the corpus, but the heterogeneity of the genre is evident, for
example as compared with law, another genre with a relatively low per-
centage. Handbooks and education, which share the feature of being in-
structive, are shown to be very wide apart on the continuum; the latter
reflect a conspicuously high degree of uniformity. The varying degrees of
viewpoint marking in non-imaginative narration put diaries and trav-
elogues at the lower end, history and autobiographies close to the mean
and biographies at the high end on the scale. Explicitly interactive texts,
including texts completely in dialogue form (see p. 322) and letters and
sermons, are all among genres where markers of point of view are at least
60% of the total of adverbs and adjectives. The uniformity of genres is
quite striking, particularly at the high end of the scale. However, it is
necessary to stress once more that the number of texts per genre is too
low and the evidence must be considered only preliminary.

10.2. Point of view in text categories

There are altogether eight texts classified as expository, eight as secular

instruction and twenty-four as non-imaginative narration in the Early
Modern English part of the Helsinki Corpus, and it is in these three cat-
egories that the evidence is more conclusive. Table 9 (see p. 319) illus-
trates the mean frequencies of adverbs and adjectives in text categories
and gives the percentages of the total of point-of-view markers among
them.
The variable 'text category' clearly conditions the mean frequency of
adjectives in texts, so that 'expository' and 'secular instruction' or their
combination show the highest mean frequencies, the two types of 'narra-
tion' and 'statutory' the lowest. Statutory and expository texts and imagi-
native narration are shown to be the most uniform, but, despite the high
number of texts, there is evidence of some degree of homogeneity also in
the category non-imaginative narration. Adverbs are distributed more un-
evenly. Similarity in secular instruction/expository is conditioned by date.
The frequency of both adverbs and adjectives is highest in expository
texts and in secular instruction.
 Reconstructing a grammar of point of view 319

Table 9. Mean frequencies of open-class adverbs and adjectives and percent-

ages of markers of point of view in text categories in the Early Mod-
ern English part of the Helsinki Corpus. Standard deviation in brack-
ets. N=number of texts.

Adverbs Adjectives Viewpoint

Text category Ν Mean Mean markers (%)

Statutory 3 2.8 (1.9) 25.5 (1) 36 (5)

Expository 8 9.3 (5.3) 43.7 (9) 40 (23)
Narration, non-imaginative 24 5.6 (2.9) 25.8 (8.8) 47 (16)
Instruction, secular 8 7.4 (2.9) 38.9 (14.8) 48 (18)
Narration, imaginative 6 6.1 (2) 23.5 (2.5) 57 (10)
Instruction, religious 6 6 (2.6) 29.1 (8.8) 64 (14)
Unspecified 22 5 (2.5) 22.2 (8.6) 64 (7)
Instruction, secular/Exposit. 2 14.3 (6.4) 37.1 (3.1) 68 (2)

Total 79 6.2 (3.5) 28.3 (11.3) 53 (16)

Besides the two early texts representing secular instruction/expository, it

is in religious instruction and, slightly more uniformly, in imaginative
narration that the presence of viewpoint markers is dominant. Statutory
and expository texts show similar percentages, but the latter are clearly
more heterogeneous. The differences in the distributions of individual fea-
tures, discussed in earlier sections, are covered up by the equal propor-
tions in non-imaginative narration and secular instruction. The table also
shows that point of view is more manifestly prevalent in religious than in
secular instruction, and that there is a similar difference between imagina-
tive and non-imaginative narration.
No specific direction of change has been attested in the total mean fre-
quencies of evaluative adverbs in the different text categories over time.
In contrast, in individual text categories a striking increase in the fre-
quency and uniformity in the occurrence of predicate-oriented evaluative
adverbs is seen in expository texts (from 0.5, st.d. 0.5 to 4.5, st.d. 1). The
most manifest change in subject-orientation is reflected in the decrease of
such evaluative adverbs in the two types of narrative (from 1.9, st.d. 1.4
in imaginative narration and 1.4, st.d. 1.5 in non-imaginative narration to
0.7, st.d. 0.3 and 0.7, st.d. 0.8 respectively). As regards descriptive ad-
verbs (total mean 0.7), expository texts (1.5, st.d. 0.8) are well above the
mean and relatively homogeneous.
320 Anneli Meurman-Solin

Diachronically, there is an increase in the mean frequency and uniform-

ity of focusing subjuncts in secular instruction (from 0.6, st.d. 0.3, to 1.4,
st.d. 0.3), and a less significant one in imaginative narration (1st sub-
period: 0.3, st.d. 0.2; 2nd: 0.2, st.d. 0; 3rd: 0.7, st.d. 0.3). As regards in-
tensifiers, a change may have taken place in their frequency and uniform-
ity of usage in expository texts (from 0.5, st.d. 0.7, to 1.4, st.d. 0.5). An
increase of descriptive adverbs has been attested in statutory texts (from
0.3-0.4 to 2.3) (cf. the change in the frequency of descriptive adjectives
from 12.6 to 16.7 in this category). Three text categories show a pattern
where the texts dating from 1570-1640 contain lower frequencies of de-
scriptive adverbs than the often remarkably uniform earlier and/or later
texts (expository: 1.9, st.d. 0.2; 0.8, st.d. 0.5; 2.3, st.d. 0.2) secular in-
struction: 0.9, st.d. 1; 0.2, st.d. 0.1; 0.8 , st.d. 0.1) and imaginative nar-
ration: 1.3, st.d. 0.7; 0.1, st.d. 0.2; 1, st.d. 0.4).
Diachronic developments in the mean frequencies of adjectives in the
various text categories are well documented. As regards descriptive adjec-
tives, there is a decrease over time in secular instruction (from 38.5, st.d.
27.2 to 16.3, st.d. 6.5); expository texts remain very homogeneous, par-
ticularly in the first (mean 30.8, st.d. 0.1) and the third subperiod (37.3,
st.d. 3.2; cf. 2nd subperiod: 17.9, st.d. 9.1). The same trend is seen in
imaginative narration (1st subperiod: 10.2, st.d. 0.3; 2nd: 12.7, st.d. 3.3;
3rd: 7.7, st.d. 0.2). As a reflection of a differentiation process, standard
deviation increases over time in non-imaginative narration (from 11.2,
st.d. 3.5 to 15.8, st.d. 10.9). Evaluative adjectives show an increase in fre-
quency and uniformity of usage in expository texts (from 1.8, st.d. 1.6, to
12.2, st.d. 7.8, and finally to 7.9, st.d. 1.6), in religious instruction (from
8.7, st.d. 2.7, to 5.4, st.d. 3.5, and finally to 16.8, st.d. 3.3). This is true
also of secular instruction, which remains relatively homogeneous over
time (5.3, st.d. 0.9; 9.8, st.d. 0.4; 11.5, st.d. 4.9). No change seems to
have taken place in the categories 'non-imaginative narration' and 'sta-
tutory'.
An increase in the frequency of adjectives in the role of epistemic mo-
dality has been attested in expository texts (from 0.4, st.d. 0.1, to 1.3, st.d.
0.2, and finally to 2.4, st.d. 0.1). A similar increase is seen in instruction
(religious: from 0.4 to 2.6, and secular: from 0.2 to 1.4), but the standard
deviation is high. Focusing adjectives characterize expository (2.2, st.d.
1.1) and statutory texts (1.3, st.d. 0.5) and instruction (secular: 0.9, st.d.
0.5; religious: 0.9, st.d. 0.6).The frequency of focusing adjectives in ex-
pository texts also increases over time (1.4, st.d. 0.5; 2.7, st.d. 1.3; 2.2.,
st.d. 0.7). There is no significant uniformity of practice in the use of in-
 Reconstructing a grammar of point of view 321

tensifying adjectives in terms of text categories. However, the difference

between sixteenth- and seventeenth-century statutes is striking, as the for-
mer use intensifying adjectives much more frequently (2.9 as compared
with 0.5 and 0.2). Adjectives referring to state of mind are clearly a fea-
ture typical of imaginative narration (3.2, st.d. 1), but also of religious
instruction (2.3, st.d. 1.1) and secular instruction/expository (2.5, st.d.
0.5).

11. On the conditioning effects of other extralinguistic

variables

Further evidence of text- or genre-specific characteristics can be provided

by relating the findings discussed above to six extralinguistic variables
coded into the Helsinki Corpus: audience description, setting, interaction,
relationship to spoken language, and the author's rank and sex. The audi-
ence of all handbooks and Elyot's and Locke's educational treatises is
labelled as 'non-professional', while the audience of scientific treatises
and Ascham's, Bacon's, Brinsley's and Hoole's educational treatises is
assumed to have been professional. A t-test showed an almost significant
difference suggesting a more prominent use of focusing adjectives
(p=.019) and descriptive adverbs (p=.026) in texts written to a profes-
sional audience than in those written to a non-professional audience
(mean frequency 1.9, st.d. 1.1, and 0.9, st.d. 0.5, respectively). Instead of
being conditioned by audience, these features could, however, partly re-
flect differences between handbooks and science (see Tables 2 and 6).
In Elyot's Gouernour ('non-professional') 67% of open-class adverbs
are evaluative, and only 2% modal, while in Ascham's Scholemaster
('professional') 52% are evaluative, 9% modal. The mean of descriptive
adverbs is only 0.4 in Elyot, 2.4 in Ascham, the mean of focusing adjec-
tives 1.6 in Elyot, 0.4 in Ascham. Descriptive adverbs are more typical
of scientific treatises (professional audience) than of instructive hand-
books (non-professional audience), there being almost no occurrences in
three of the six handbooks (cf. also the slightly higher frequency of de-
scriptive adjectives in the former (28.5, st.d. 3.4, and 32.7, st.d. 6.1, as
compared with 25.3, st.d. 17.9, and the fact that scientific treatises are
much more uniform in this respect).
322 Anneli Meurman-Solin

All comedies, letters and trials are labelled as 'interactive' in the cor-
pus. In addition, Colville's Boethius translation, fiction by Deloney and
Penny Merriments, Gifford's handbook and Brinsley's educational treatise
are in dialogue form. The other two Boethius translations, Harman, Armin
and Walton are only partly dialogue but they have been included among
interactive texts in the statistics.
Tested by t-test, the differences between interactive and non-interactive
texts are as follows. Interactive texts are marked by the use of modal ad-
verbs (p=.015) and particularly by the frequency of adverbs and adjec-
tives of epistemic modality (p=.002 and .008 respectively). The frequency
of adjectives referring to state of mind is statistically highly significant
(p=.001). In contrast, non-interactive texts are characterized by a statisti-
cally almost significant probability of a more prevalent use of descriptive
(p=.030), intensifying (p=.039) and time adverbs (p=.031), and of focus-
ing and deontic modal adjectives (p=.024 and .021 respectively). The
highly significant feature (p=.000) diagnostic of non-interactive texts is
the mean frequency of descriptive adjectives (16.1, st.d. 10.5) as com-
pared with interactive texts (8.4, st.d. 2.8). This also reflects the lower
general frequency of adjectives in the latter group of texts (see also Table
1)·
The variable 'setting' has been given parameter values 'formal', 'in-
formal' and 'unspecified'. Law, the Bible, all sermons, trials and official
letters have been labelled as 'formal', diaries, fiction (excluding Behn),
comedies and private letters as 'informal'. A statistically significant (p=
.006) difference between the groups of texts 'formal' and 'informal' was
attested in the higher mean frequency in the former of focusing adjectives
(formal: 0.9, st.d. 0.5; informal: 0.5, st.d. 0.5). The evidence is weakened
by the relatively low number of occurrences and a high standard devia-
tion. Interestingly, adverbs of perception modality (p=.017) are a distinc-
tive feature in texts labelled 'formal' (cf. adjectives of perception mo-
dality, p=.058). In contrast, adjectives referring to state of mind are more
frequent in informal texts (p=.016; formal: mean 1.7, st.d. 1.3, informal:
mean 3.3, st.d. 2.4).
The variable 'relationship to spoken language' was looked at by com-
paring 'script', represented by all sermons and comedies, and 'speech-
based' texts, all trials and Fox's autobiography, with the rest of the texts
of the corpus. As regards the total of adverbs, both script and speech-
based texts are towards the lower end of the scale (normalized mean fre-
quency 3-6).' In the use of adverbs, texts in the two groups resemble one
another. Only Fox's autobiography, known to have been dictated, differs
 Reconstructing a grammar of point of view 323

from 'written' texts by a very low general frequency of adverbs (only six
instances in the extract of 5,560 words). As regards adjectives, the mean
frequencies are consistently very much lower than in 'written' texts or
'script' (e.g. evaluative adjectives 4, total mean 8; descriptive adjectives
7, total mean 13).
That the author's social rank (for parameter values, see Kytö 1996: 53)
may affect a text's point of view is borne out by comparisons of contem-
porary representatives of the same genre. For example in science, 61% of
adverbs in Clowes ('high rank') are evaluative and only 9% are modal or
time adverbs with habitual or generic time reference. In contrast, his
contemporary Blundevile, of the 'professional' rank, uses only 23% of
evaluatives, but altogether 42% of modals and of the above-mentioned
time adverbs, and also slightly more descriptive adverbs. There is a some-
what less conspicuous difference between Boyle, 'high' and Hooke, 'pro-
fessional' in the latter half of the seventeenth century. Boyle uses fewer
modal adverbs, Hooke fewer evaluatives (see Tables 3 and 4); the total of
point-of-view adverbs is also lower in Hooke. In educational treatises
Hoole, 'professional', uses more evaluative adverbs, adverbs with generic
time reference and focusing subjuncts than Locke, 'high': 45% versus
34%, 14% versus 7% and 17% versus 11% respectively. Focusing sub-
juncts are a prominent feature in Madox's ('professional') diary com-
pared with Edward VI ('high'). However, the corpus is not representative
enough for the overall assessment of the conditioning effect of the
author's rank.
This is also true as regards the variable 'author's sex', and it is there-
fore not possible to draw any conclusions about the possible distinctive
features of women's language (cf. The Corpus of Early English Corre-
spondence, Nevalainen—Raumolin-Brunberg 1994). Lady Hoby's diary
differs from all the other texts by using open-class adverbs that are chief-
ly descriptive. About 48% of these adverbs are evaluative in both Fryer's
and Celia Fiennes's travelogues, but there are no modals in Fiennes; in
contrast the proportion of intensifiers (3% versus 14%) is clearly higher
in the latter. The important role of intensifiers may be related to women's
language, as also in Aphra Behn's style they represent the majority of
open-class adverbs (29%; cf. evaluative 20%, modal 11 %, descriptive
20% and focusing 14%). In Oates and Alice Lisle, the two late-seven-
teenth-century trials, the main difference is in the frequency of evalua-
tives, the latter trial of a woman containing 44% of evaluatives compared
with 29% in the former. Yet I am rather reluctant to consider these find-
ings as evidence of anything else but idiolectal practices.
324 Anneli Meurman-Solin

12. Conclusions

The problem of how point-of-view marking relates to different narrator

roles remains unsolved, as further study will be necessary to cope with
the complexity of participant networks in early genres. As early texts are
often amalgamations of different sources, the topic is exceptionally chal-
lenging. The present study gives information about the general frequency
of markers of point of view in texts, but the identification of the various
speakers whose attitudes they articulate would shed more light on genre-
and category-based features. The structural features of each of the genres
included in the corpus should be described in more detail before the
various narrator roles in them can be referred to in an analysis of point of
view. Ideally, such a description should however be based on entire texts
(for the sampling principles adopted for the Helsinki Corpus, see Kytö
1996: 1).
Important differences have been attested in the general frequencies of
adjectives and open-class adverbs and their co-occurrence patterns in
genres and text categories in the Early Modern English part of the Hel-
sinki Corpus. Markers of attitude as expressed by means of these two fea-
tures have been found to show variation that can be considered meaning-
ful in the identification of similarities and differences between texts and
various groups of texts. The study stresses that the evidence provided is
not conclusive because of the low number of texts in the various classes
and subclasses. However, the high degree of uniformity in the frequencies
and co-occurrence patterns of a number of features, and the significant
correlation patterns between them, show that some groupings into text
types can be suggested.
In the reconstruction of point of view, the features looked at and their
subcategorization turned out to be relevant and useful. For example, the
grouping of modal adverbs and adverbs with habitual and generic time
reference in the assessment of modality was proved justified by the highly
significant correlation attested between these adverbs and also the cor-
responding adjectives, whether grouped together or considered separately.
The significant correlation between perception modality and generic time
reference is particularly interesting. Realizations of types of modality are
differently distributed across texts: subtypes of modality do not correlate
with one another. While modality as a whole shows the highest mean fre-
quencies in the Bible, the Boethius translations and the two late trials,
 Reconstructing a grammar of point of view 325

epistemic modality characterizes interactive texts, such as letters, com-

edies and trials, perception modality some handbooks, scientific treatises
and sermons, and deontic modality educational treatises.
The methodological problem of a feature's relative infrequency, such as
that of open-class adverbs in the different semantic subcategories, was
solved by considering both the percentages and the mean frequencies of
items in texts. A high proportion of a particular feature provides signifi-
cant information for the description of idiolects, but mean frequencies and
the standard deviations between them are essential when we compare
texts, genres and text categories. By using both ways of assessing the
importance of point-of-view markers in texts, it has been possible to posi-
tion them on a bipolar dimension depicting degrees of intensity in view-
point marking. The findings of the study stress the relevance of co-occur-
rence patterns of features, so that a text with a high mean frequency of
items in a cluster of various point-of-view markers, preferably realized by
means of adverbs or predicative adjectives in the roles of evaluation, mo-
dality or degree, can be positioned at the high end of the scale of stance
marking, particularly if the relative proportions of the items in the total of
these adverbs and adjectives are also high.
Despite the low frequencies of evaluative and modal adverbs in a num-
ber of texts, some evidence has been provided about the positioning of
these adverbs in sentence structure. Adverbs of perception modality and
deontic modality are typically put in medial position (there is a highly
significant correlation between them and medial position), whereas no
such preference can be seen in the positioning of epistemic adverbs. On
the other hand, only epistemic modal adverbs correlate with initial posi-
tion, which reflects their more established use as sentence adverbs. That
perception modality also correlates sigificantly with final position is due
to their use with intransitive verbs. Final position seems to be the un-
marked position: there is an almost significant correlation with initial and
medial positions, but initial and medial positions do not correlate. This
may be due to differences in transitivity. The preferred position of inten-
sifiers is established to be medial; they correlate significantly with modal
adverbs, typically the medially-positioned modal adverbs, in contrast with
the other two positions. A highly significant correlation was also attested
between intensifying adverbs and predicate-oriented evaluative adverbs,
which also applies to focusing subjuncts. The highest frequencies of ini-
tially-positioned disjuncts were attested in scientific treatises.
The very high number of statistically significant correlations between
various subcategories of viewpoint markers and, on the other hand, the
326 Anneli Meurman-Solin

different distributions of descriptive elements reflect a pattern of comple-

mentary distribution which suggests the existence of bipolar dimensional-
ity. For example the expression of attitudes is chiefly realized by means
of modal and intensifying adverbs in texts reflecting a high degree of in-
formational density (cf. Meurman-Solin 1994: 111). This finding is also
in accordance with Biber's claim that a density of attributive adjectives
can be meaningfully contrasted with a high frequency of amplifiers.
The texts containing a high frequency of attributive adjectives seem to
be those where the author's presence is dominant because of his authori-
tative status as regards the topic discussed, i.e. handbooks. In these texts
two aims coincide: one objective is to give all the necessary information
to the reader, the other is to convince the reader of the reliability of that
information. The first aim is achieved for example by using numerous ad-
jectives, the second aim by means of adverbs of epistemic and perception
modality and intensifying adverbs. The general frequency of the former
is much lower in texts where intensifiers are a significant feature, higher
in texts where epistemic modality is prevalent (see subfactors 2 and 3 in
Meurman-Solin 1994: 99). That there is a highly significant correlation
also between perception modals and focusing adjectives indicates that in
texts with high mean frequencies of these features, information is pro-
cessed by stressing the truth-value of perceptions and by restricting or ex-
panding the focus. Adverbs of perception modality also correlate signifi-
cantly with adverbs of habitual and generic time reference, as exemplified
by Turner's handbook.
In addition to the bipolar dimension 'low degree of viewpoint marking'
versus 'high degree of viewpoint marking', a network of dimensions has
been created by means of a further subcategorization of point-of-view
markers. For example certain texts (Ascham, Clowes, Roper, Burnet's
biography and Harman) have been shown to be at the high end on a scale
assessing the importance of the role of evaluation as expressed by means
of adverbs. Ascham and Clowes are also among the texts that have ex-
ceptionally high frequencies of evaluative adjectives, but in the former
their proportion of the total of adjectives is higher; by contrast, in Clo-
wes, the proportion of adjectives as complements is higher than in As-
cham. Modal expressions have been found to be less frequent than evalu-
ative ones, but their significance in a grammar of point of view is height-
ened, on the one hand, by the tendency to position epistemic modal ad-
verbs initially and use them as independent propositions, and on the other
hand, by the higher frequency of modal adjectives as complements. This
hierarchical system of assessing the communicative dynamism of view-
 Reconstructing a grammar of point of view 327

point markers makes it possible to identify, for example from among

these five texts, Clowes as the one where the choices between the various
realizations stress the dominance of point of view.
The clustering of texts on the basis of correlation patterns provides evi-
dence for a typology. Despite the highly significant correlation between
evaluative adverbs and adjectives, the former co-occur prevalently with
intensifying and focusing subjuncts; the prototypical representatives of
texts illustrating this pattern are diaries (especially those by Madox and
Evelyn), private letters and Celia Fiennes's travelogue. The latter, i.e.
evaluative adjectives, co-occur typically with modal and intensifying ad-
jectives, so that this pattern of clustering features characterizes especially
Fisher and Hooker among sermons, Roper and Perrott among biographies,
Brinsley among educational treatises, Raleigh among trials and Hayward
among histories. Private letters of the period 1500-1570 and official let-
ters of the period 1640-1710 could also be mentioned as examples. A
third cluster can be formed on the basis of the significant correlation be-
tween modal and intensifying adverbs, attested in handbooks (Turner, Gif-
ford and Markham), sermons (Fisher and Tillotson), Udall's comedy, let-
ters and Boethius translations, the last two only in the period 1640-1710.
A fourth cluster consists of co-occurrence patterns of modal and evalua-
tive adjectives, which correlate highly significantly, and of adjectives with
habitual and generic time reference; Locke is a prototype in this respect.
While the assumed multidimensional pattern is borne out by the evi-
dence of the above-mentioned correlations, bipolarity is seen in the lack
of significant correlation between descriptive adjectives and adverbs and
those categorized as markers of point of view. The only important excep-
tion is that adverbs of perception modality correlate with descriptive ad-
jectives highly significantly. Evidence of this kind speaks for a typology
with core texts illustrating prototypes and other texts representing vari-
eties, which are less uniform and reflect overlappings with representatives
of other varieties, grouped together around other core texts.
The weight of text category as a conditioning factor is quite significant.
The degree of inner homogeneity is higher when assessed by means of
adjectives, whether evaluative or descriptive; adverbs serve better to iden-
tify differences between texts. Yet the differences between the mean fre-
quencies of the categories are conspicuous as regards both adjectives and
adverbs. The most uniform category characterized by a consistently high-
er-than-mean frequency of viewpoint markers is secular instruction; this
uniformity is however slightly flawed by a relatively high standard devia-
tion in the frequency of one feature, i.e. descriptive adjectives. This cat-
328 Anneli Meurman-Solin

egory could be compared with 'expository', another category with regu-

larly high mean frequencies, but also regularly high deviation in the fre-
quency of point-of-view markers, and more uniformity in a generally very
high frequency of descriptive adjectives. In addition, while imaginative
narration has been seen to develop towards greater uniformity, non-imagi-
native narration tends to become more diversified; the author-centred
diaries and autobiographies manifestly differ from travelogues and biogra-
phies particularly in the third subperiod.
As regards the conditioning effect of other extralinguistic variables, the
influence of audience seems subordinate to that of text category, so that
scientific treatises and certain educational treatises differ from one an-
other as texts belonging to expository or secular instruction categories,
despite the fact that they share the feature of being written to a 'pro-
fessional' audience. The variable 'interaction' is one of the most impor-
tant diagnostic features in the identification of specific genre styles. Inter-
action tends to increase the frequency of modal adverbs and adjectives,
particularly those in the role of epistemic modality, and the frequency of
adjectives referring to state of mind. In contrast, lack of interaction seems
to lead to the inclusion of more information, which is reflected in the
higher frequency of not only descriptive but also focusing and deontic
modal adjectives and, in addition, in a more frequent use of descriptive,
intensifying and time adverbs.
The two parts of speech mostly correlate at least within the same se-
mantic role, but often also within the total of viewpoint markers. How-
ever, the study provides information about the preference of either of the
two realizations, so that for example point of view is expressed by means
of intensifying and modal adverbs (rather than adjectives) in informa-
tionally dense texts where descriptive adjectives are exceptionally fre-
quent (e.g. handbooks, travelogues, histories and biographies).
Also the distinction between the evaluation of state of mind and other
evaluative comments has been shown to be relevant. There is no correla-
tion between the totals of the two groups (on subject-oriented evaluative
adverbs, see below) and their different distributions usefully reveal dis-
tinctive features in genres. Letters, comedies, fiction, and biographies
share the feature of containing subject-oriented evaluative adverbs and ad-
jectives describing state of mind; these two realizations correlate highly
significantly.
The mean frequency of adjectives as complements is a more powerfully
diagnostic feature in the assessment of stance marking than adjectives
integrated in noun phrase structures, which, in principle, tend to be non-
 Reconstructing a grammar of point of view 329

restrictive and refer to permanent characteristics. Also standard deviation

has been found to be higher in the frequencies of descriptive adjectives as
complements than as modifiers; complements are conspicuously frequent
in science and handbooks as compared with other genres, science/other
being particularly uniform.
Besides adding to the informational density, attributive adjectives inte-
grate generally valid evaluative or modal comments rather than give ex-
pression to the author's or the speaker's direct statements of opinion. In
their communicative function they convey generalizing or classifying
comments which create a framework of given information for the intro-
duction of more specific, more directly participant-related comments in
complement positions. Another interesting point is that it is in modifier
positions that for example evaluative adjectives referring to state of mind
have been shown to develop into genre-specific conventions. This is part
of a general tendency in which an increasing degree of nominalization
and the resulting informational density can be contrasted with a prominent
role of viewpoint marking as a text-organizing device, chiefly expressed
by means of predicative adjectives and adverbs. The former process
creates a context for the conventionalization and semantic and syntactic
lexicalization of adjectives, whereas the latter reflects the degree in which
the different participant voices are present in texts. Moreover, it has been
shown that complements dominate texts written in dialogue form such as
comedies and 'scripts' of texts addressing an audience such as sermons,
whereas modifiers are more frequent in letters, in which epistolary con-
ventions are often noun phrase structures containing adjectives referring
to state of mind.
As regards similar developments in adverbs, the study has identified
contexts where the varying positions of evaluative adverbs reflect changes
in their syntactic function and meaning. Besides disjuncts, for example
subject-oriented adverbs referring to state of mind may function as cour-
tesy subjuncts modifying speech act verbs. However, a larger corpus will
be necessary in order to provide sufficient evidence for tracing develop-
ments towards collocational uses of adverbs and verbs.
Integration of point-of-view markers is a source for generic conven-
tions, in which the communicative efficiency of these markers is usually
weakened. Fragmentation can be illustrated by the development of predi-
cate-oriented process adjuncts into expressions of the author's evaluation
that can take the function of an independent proposition in sentence struc-
ture. The dimension 'integration' versus 'fragmentation' is thus central in
a grammar of point of view. On the basis of the present findings, the di-
330 Anne Ii Meurman-Solin

mension can be directly related to the rise of collocations with inherent

semantic lexicalization and, in the case of sentence adverbs, also to gram-
maticalization processes.

Appendix 1.

Extralinguistic variables of audience, interaction and setting.

Ρ period
SUB subperiod
CAT text category
TT genre
Q abbreviated title
υ description of audience
SET setting
E l , E2, E3 = EModEl (1500-1570), EModE2 (1570-1640),
EModE3 = (1640-1710)
PROF professional
NON-PROF = non-professional
INT interactive
FOR formal
INF informal
X, XX unspecified

Ρ SUB CAT TT Q U INT SET

Ε ΕΙ STA LAW STAT3 PROF X FOR

Ε ΕΙ IS HANDO FITZH NON-PROF X X
Ε ΕΙ IS HANDO TURNER NON-PROF X X
Ε ΕΙ EX SCIM VICARY PROF X X
Ε ΕΙ EX SCIO RECORD PROF X X
Ε ΕΙ IS/EX EDUC ELYOT NON-PROF X X
Ε ΕΙ IS/EX EDUC ASCH PROF X X
Ε ΕΙ XX PHILO BOETHCO X INT X
Ε ΕΙ IR SERM FISHER X X FOR
Ε ΕΙ IR SERM LATIMER X X FOR
Ε ΕΙ XX TRI THROCKM X INT FOR
Ε ΕΙ NN HIST MORERIC X X X
Ε ΕΙ NN HIST FABYAN X X X
 Reconstructing a grammar of point of view 331

Appendix 1. Continued.

Ρ SUB CAT TT Q U INT SET

Ε El NN TRAV LELAND X X X
Ε El NN TRAV TORKINGT X X X
Ε El NN DIARY MACHYN X X INF
Ε El NN DIARY EDWARD X X INF
Ε El NN ΒΙΑ MOWNTAYNE X X X
Ε El NN BIO ROPER X X X
Ε El NI FICT MERRYTAL X X INF
Ε El NI FICT HARMAN X INT INF
Ε El XX COME UDALL X INT INF
Ε El XX COME STEVENSO X INT INF
Ε El XX CORP PRIVLET X INT INF
Ε El XX CORO OFFLET X INT FOR
Ε El XX BIBLE TYND X X FOR
Ε E2 STA LAW STAT4 X X FOR
Ε E2 IS HANDO GIFFORD NON-PROF INT X
Ε E2 IS HANDO MARKHAM NON-PROF X X
Ε E2 EX SCM CLOWES PROF X X
Ε E2 EX SCIO BLUNDEV PROF X X
Ε E2 EX EDUC BRINSLEY PROF INT X
Ε E2 EX EDUC BACON PROF X X
Ε E2 XX PHILO BOETHEL X INT X
Ε E2 IR SERM HOOKER X X FOR
Ε E2 IR SERM SMITH X X FOR
Ε E2 XX TRI ESSEX X INT FOR
Ε E2 XX TRI RALEIGH X INT FOR
Ε E2 NN HIST STOW X X X
Ε E2 NN HIST HAYWARD X X X
Ε E2 NN TRAV JOTAYLOR X X X
Ε E2 NN TRAV COVERTE X X X
Ε E2 NN DIARY MADOX X X INF
Ε E2 NN DIARY HOBY X X INF
Ε E2 NN BIA FORMAN X X X
Ε E2 NN BIO PERROTT X X X
Ε E2 NI FICT ARMIN X INT INF
Ε E2 NI FICT DELONEY X INT INF
Ε E2 XX COME SHAKESP X INT INF
Ε E2 XX COME MIDDLET X INT INF
332 Anneli Meurman-Solin

Appendix 1. Continued.

Ρ SUB CAT TT Q U INT SET

Ε Ε2 XX CORP PRIVLET X INT INF

Ε Ε2 XX CORO OFFLET X INT FOR
Ε Ε2 XX BIBLE AUTH X X FOR
Ε Ε3 STA LAW STAT7 X X FOR
Ε Ε3 IS HANDO WALTON NON-PROF INT X
Ε Ε3 IS HANDO LANGF NON-PROF X X
Ε Ε3 EX SCIO HOOKE PROF X X
Ε Ε3 EX SCIO BOYLE PROF X X
Ε Ε3 IS EDUC LOCKE NON-PROF X X
Ε Ε3 IS EDUC HOOLE PROF X X
Ε Ε3 XX PHILO BOETHPR X INT X
Ε Ε3 IR SERM TILLOTS X X FOR
Ε Ε3 IR SERM JETAYLOR X X FOR
Ε Ε3 XX TRI OATES X INT FOR
Ε Ε3 XX TRI LISLE X INT FOR
Ε Ε3 NN HIST BURNETCHA X X X
Ε Ε3 NN HIST MILTON X X X
Ε Ε3 NN TRAV FIENNES X X X
Ε Ε3 NN TRAV FRYER X X X
Ε Ε3 NN DIARY PEPYS X X INF
Ε Ε3 NN DIARY EVELYN X X INF
Ε Ε3 NN ΒΙΑ FOX X X X
Ε Ε3 NN BIO BURNETROC X X X
Ε Ε3 NI FICT PENNY X INT INF
Ε Ε3 NI FICT BEHN X X X
Ε Ε3 XX COME VANBR X INT INF
Ε Ε3 XX COME FARQUHAR X INT INF
Ε Ε3 XX CORP PRIVLET X INT INF
Ε Ε3 XX CORO OFFLET X INT FOR
 Reconstructing a grammar of point of view 333

Appendix 2.

Correlations

1. Open-class adverbs in the EModE section of the Helsinki Corpus of English

Texts. N=79.

EVA = evaluative; SOEVA = subject-oriented with reference to state of mind

MOD = modal; EPI=modal/epistemic, PER=modal/perception
INT = intensifying
FOC = focusing
TH = habitual time reference; TG = generic time reference
DES = descriptive

EVA MOD EPI PER INT

EVA 1.000 .326 .196 .379 .464

P= .003 .084 .001 .000
MOD .326 1.000 .888 .579 .303
P= .003 .000 .000 .007
EPI .196 .888 1.000 .171 .180
P= .084 .000 .133 .113
PER .379 .000 .171 1.000 .349
P= .001 .579 .133 .002
INT .464 .303 .180 .349 1.000
P= .000 .007 .113 .002
FOC .416 -.041 -.076 .094 .215
P= .000 .720 .503 .410 .057
DES .367 .227 .071 .417 .261
P= .001 .045 .537 .000 .020
THTG .491 .266 .176 .322 .401
P= .000 .018 .120 .004 .000
SOEVA .719 .034 .053 -.010 .225
P= .000 .769 .646 .927 .046

FOC DES THTG SOEVA

EVA .416 .367 .491 .719

P= .000 .001 .000 .000
MOD -.041 .227 .266 .034
P= .720 .045 .018 .769
EPI -.076 .071 .176 .053
P= .503 .537 .120 .646
334 Anneli Meurman-Solin

Appendix 2. Continued.

FOC DES THTG SOEVA

PER .094 .417 .322 -.010

P= .410 .000 .004 .927
INT .215 .261 .401 .225
P= .057 .020 .000 .046
FOC 1.000 .100 .539 .132
P= .382 .000 .246
DES .100 1.000 .263 .144
P= .382 .019 .207
SOEVA .132 .144 .069 1.000
P= .246 .207 .549

2. Adjectives in the EModE section of the Helsinki Corpus. N=79.

AEVA = evaluative; ASOE = with reference to state of mind

AMOD = modal; ΑΕΡΙ = modal/epistemic, APER = perception, ADEO =
modal/deontic
AINT = intensifying; AFOC = focusing
ΑΤΗ = habitual time reference; ATG = generic time reference
ADES = descriptive

AEVA AMOD ASOE AINT AFOC ATHTG

AEVA 1.000 .483 .231 .312 .064 .527

P= .000 .040 .005 .576 .000
AMOD .483 1.000 .140 .240 .143 .298
P= .000 .219 .033 .209 .008
ASOE .231 .140 1.000 .125 -.255 -.043
P= .040 .219 .272 .023 .709
AINT .312 .240 .125 1.000 .115 .334
P= .005 .033 .272 .315 .003
AFOC .064 .143 -.255 .115 1.000 .374
P= .576 .209 .023 .315 .001
ATHTG .527 .298 -.043 .334 .374 1.000
P= .000 .008 .709 .003 .001
ADES -.026 -.045 -.375 -.066 .303 .258
P= .819 .695 .001 .562 .007 .022
 Reconstructing a grammar of point of view 335

Appendix 2. Continued.

ADES

AEVA -.026
P= .819
AMOD -.045
P= .695
ASOE -.375
P= .001
AINT -.066
P= .562

ADES

AFOC .303
P= .007
ATHTG .258
P= .022
ADES 1.000
P=

3. Open-class adverbs and andjectives in the EModE section of the Helsinki

Corpus. N=79.

AEVA AMOD ΑΕΡΙ APER ADEO AINT

EVA .539 .294 .197 .241 .246 .191

P= .000 .009 .082 .033 .029 .091
MOD .350 .590 .493 .658 .093 .377
P= .002 .000 .000 .000 .415 .001
EPI .358 .479 .439 .430 .072 .316
P= .001 .000 .000 .000 .531 .005
PER .107 .427 .315 .579 .078 .243
P= .347 .000 .005 .000 .492 .031
INT .394 .240 .169 .242 .141 .371
P= .000 .033 .137 .032 .215 .001
FOC .361 .066 -.022 -.022 .288 .070
P= .001 .565 .848 .846 .010 .540
DES .064 .084 .031 .157 .057 -.012
P= .578 .460 .788 .167 .620 .920
336 Anneli Meurman-Solin

Appendix 2. Continued.

AFOC ADES ATHTG

EVA .314 .115 .491

P= .005 .313 .000
MOD .111 .023 .266
P= .330 .843 .018
EPI -.059 -.227 .176
P= .604 .044 .120

AFOC ADES ATHTG

PER .414 .505 .322

P= .000 .000 .004
INT .174 .234 .401
P= .125 .038 .000
FOC .281 .158 .539
P= .012 .165 .000
DES .284 .429 .263
P= .011 .000 .019

Notes

1. Meurman-Solin (forthcoming) increases the representativeness of the ma-

terial by including texts available in the Helsinki Scots Corpus. Compari-
sons are also made between the two geographical varieties. The Scots cor-
pus contains sixty-two texts dating from 1450-1700 and representing four-
teen different prose genres, some of them illustrated by 7-9 specimens. The
size of the first version is 600,000 words of running text. The revised and
expanded version was made internationally available in 1995.
2. Printing seems to have had a dual function in the history of Scots. In the
sixteenth century it served to establish the so-called Scottish Standard,
printed texts often following the Scottish practices even more regularly than
unprinted texts. In the seventeenth-century texts its role as a means to reach
a wider audience becomes evident, and, particularly in the latter half of the
century, many printed texts, in contrast with texts that remained in manu-
script, are completely anglicized.
3. In this study, the terms 'genre' and 'text type' are distinguished, so that the
former refers to texts grouped together by extralinguistic criteria, the latter
by linguistic criteria (see for example Biber 1988: 68-70, 206-207).
 Reconstructing a grammar of point of view 337

4. The influence of changes in ideals and ways of thinking can be illustrated

for example by tracing how new ideas about scholarliness and objectivity
are reflected in scientific writing (cf. Taavitsainen 1994).
5. A case in point is the diachronic development of adverbs of epistemic mo-
dality (see Hanson 1987).
6. Some participles were included when the interpretation as adjective was
supported by evidence such as a co-ordinate non-participial adjective or an
adverb modifier (e.g. very).
7. In matters related to the statistical analyses in this study, I gratefully
acknowledge the expertise of Ms Leena Sadeniemi at the University Com-
puting Centre, University of Helsinki.
8. Adjectives in the role of space (total mean 0.8) were found to characterize
science/medicine (4.5, st.d. 1.9) and philosophy (1.4, st.d. 0.2); science/
other (2.8, st.d. 2.3) is heterogeneous, as Blundevile uses locative adjectives
much more frequently (6.2) than the other representatives of the genre. Also
history (1.1, st.d. 0.8) and travelogues (1, st.d. 0.7) vary, Fabyan and Hay-
ward showing higher mean frequencies (2.2 and 2.1 respectively) in his-
tories, and especially Leland (2.3) in travelogues. Locative adjectives corre-
late highly significantly with descriptive adjectives (see note 13).
9. In their function as logical connectors in information processing, also con-
juncts would provide interesting material for comparisons.
10. In the syntactic categorization of adverbials I am using the terminology and
criteria of Quirk et al. (1985: chapter 8). A predicate-oriented process ad-
junct modifies the predicate verb, as the descriptive slowly in He ran slowly
and the evaluative beautifully in She sang beautifully. A subject-oriented
item subjunct does not evaluate the action or process denoted by the predi-
cate verb, but serves to express the author's or the speaker's evaluation of
the state of mind of the person marked as subject in the utterance, as in Bit-
terly, he resigned, which can be paraphrased by 'he was/felt bitter when he
resigned'.
11. A relatively high proportion of time position adverbs was attested for ex-
ample in More's and Hayward's histories, Torkington's, Coverte's and
Fiennes' travelogues, Perrott's biography, Fox's autobiography and Mac-
hyn's and Edward's diaries.
12. Adverbs with generic time reference may be diagnostic of some travelogues
(e.g. Leland 0.6 and Fryer 0.8). Both generic and habitual time reference
are prominent in the educational treatises by Ascham (1.2 and 0.6 respec-
tively), Brinsley (0.7 and 0.9), Locke (0.8 and 1.3) and Hoole (1.5 and 0.5),
and, to a lesser extent, in the scientific treatises by Blundevile (0.6 and 0.5)
and Boyle (0.4 and 0.8). While absent in later Statutes of the Realm, ad-
verbs with generic time reference have been attested in early-sixteenth-cen-
tury Statutes (0.4).
338 Anneli Meurman-Solin

13. In this study the statistical significance of ρ values are assessed as follows:
almost significant ρ < 0.05, significant ρ < 0.01, highly significant ρ <
0.001.
14. Among closed-class adverbs of epistemic modality, certes is used only in
Colville's Boethius translation (28 instances). Neither no doubt (10 in-
stances) or doubtless (21 instances) show any text- or genre-specific distri-
butions. Perhaps occurs 49 times in the corpus; among these, 37 have been
attested in the texts of the period 1640-1710, mostly in science, educational
treatises and comedies. There are 151 instances of indeed and variants in
the corpus. This adverb is rare in the first subperiod (10 instances: 1 in a
private letter, 4 in Roper and 5 in Stevenson). In 1570-1640 (77 instances),
Gifford (12), Bacon (7), Armin (11), Shakespeare (6) and the Authorized
Version of the Bible (7) could be mentioned; in 1640-1710 (64 instances),
Preston's Boethius translation (10) and Vanbrugh's comedy (7) stand out.
The deontic needs and variants (85 instances) is rare in the third subperiod
(12 instances). In addition to the Boethius translations, the typical contexts
for its use in earlier subperiods are handbooks, Records's Geometry and
Brinsley.
15. There were less than 0.5/1 000 instances of modal adverbs or adjectives in
science and diaries, Fitzherbert, Elyot, Torkington and Harman in the pe-
riod 1500-1570, in law, travelogues and Lady Hoby's diary in the period
1570-1640, and in Burnet's history, Fiennes, Fox, Evelyn, Farquhar, Lang-
ford and Hoole in the period 1640-1710. Despite the general infrequency
of modal adjectives in the last two, deontic modality shows relatively high
frequencies in them (1.3 and 1 respectively) as compared with other texts.
16. In Scottish prose, also first-person pronouns and private verbs, that dele-
tion, sentence relatives and amplifiers have high factor scores in genres
such as letters, autobiographies and diaries. This finding seems to provide
converging evidence as compared with frequencies and distributions of ad-
jectives referring to state of mind (for further information see Meurman-
Solin 1994: 99, 135 and forthcoming).
17. Subject-oriented evaluative adverbs occur in initial position in later science/
medicine (Clowes) and comedies (both 0.2) and in philosophy, private
letters and sermons (0.1), in medial position chiefly in biographies and
Clowes (both 2) and in private and official letters (1.4 and 1.7 respec-
tively), in final position particularly in trials (0.6), education (0.5) and fic-
tion (0.5).
18. Adjectives referring to state of mind are slightly less frequent as modifiers
(0.9, st.d. 1.3) than as complements (1.1, st.d. 1), but the differences
between individual genres are conspicuous. Most importantly, complements
dominate especially in fiction (2.1, st.d. 0.8; cf. modifiers 1.1, st.d. 0.6), in
comedies (2.3, st.d. 1.1; cf. modifiers 1.4, st.d. 0.7), in philosophy (1.6,
st.d. 0.6; cf. modifiers 0.3, st.d. 0.03) and in sermons (1.6, st.d. 0.9; cf.
 Reconstructing a grammar of point of view 339

modifiers 0.7, st.d. 0.6). In contrast, modifiers are strikingly frequent in pri-
vate letters (5.7, st.d. 2.8; cf. complements 1.6, st.d. 0.3) and also in official
letters (2.6, st.d. 0.5; cf. complements 0.9, st.d. 0.5).
19. In a grammar of point of view, co-occurrence patterns of evaluative adverbs
and verbs of cognition and inert perception would be particularly interesting
(cf. Taavitsainen in the present volume).
20. In texts with relatively high frequencies of evaluative adverbs, the fre-
quency of passive verbs exceeds that of active verbs in two handbooks
(Turner and Walton), three travelogues (Leland, Fiennes and Fryer), two
educational treatises (Bacon and Locke) and in Stow's history, Colville's
Boethius translation and the trial of Raleigh. Of course passives attested in
a specific context do not provide evidence for any grouping of texts (cf.
Biber 1988: 111-113 on passives as diagnostic of abstract and formal
styles); nor can a straightforward correlation be seen in the dominance of
adjectives as complements (see Table 3) and verbs in the passive (cf. pas-
sives and predicative adjectives in factor 5 in Biber 1988: 103).
21. A lowering of mean frequencies of passive verbs in this specific context
has also been attested in trials (from 0.8 to 0.1). In contrast, an increase
characterizes scientific writing (science/medicine from 0.5 to 2; science/
other from 0.3 to 2.3). The former may be due to idiolectal differences
between Vicary and Clowes, as the two texts have been shown to differ
from one another also in other respects, but the latter pattern of change may
be of more general relevance.
22. Machyn's diary and the Authorized Version contain higher percentages of
intransitive verbs with evaluative adverbs, but the number of occurrences is
low. Other texts with clearly higher mean frequencies seem to belong par-
ticularly to the following two categories: secular instruction (e.g. Ascham
1.6) and non-imaginative narration (Roper's and Burnet's biographies 1.1,
Milton's history 1.2, Madox's and Evelyn's diaries 1-1.1, Torkington's and
Fryer's travelogues 0.8-1).

References

Adamson, Sylvia—Vivien Law—Nigel Vincent—Susan Wright (eds.)

1990 Papers from the 5 th International Conference on English Historical
Linguistics. (Current Issues in Linguistic Theory 65.) Amsterdam—
Philadelpia: John Benjamins.
Allerton, D. J.—Edward Carney—David Holdcroft (eds.)
1979 Function and context in linguistic analysis. A Festschrift for Wil-
liam Haas. Cambridge: Cambridge University Press.
340 Anneli Meurman-Solin

Axmaker, Shelley—Annie Jaisser—Helen Singmaster (eds.)

1988 Proceedings of the Fourteenth Annual Meeting of Berkeley the Lin-
guistics Society. Berkeley: Berkeley Linguistics Society, University
of California.
Bhatia, Vijay K.
1980 Analysing genre: Language use in professional settings. New York:
Longman.
Biber, Douglas
1988 Variation across speech and writing. Cambridge, New York, etc.:
Cambridge University Press.
Biber, Douglas—Edward Finegan
1989 "Styles of stance in English: Lexical and grammatical marking of
evidentiality and affect", TEXT 9/1: 93-124.
1992 "The linguistic evolution of five written and speech-based English
genres from the 17th to the 20th centuries", in: Matti Rissanen—
Ossi Ihalainen—Terttu Nevalainen—Irma Taavitsainen (eds.), 688-
704.
Bolinger, Dwight
1972 Degree words. (Janua Linguarum, Series Maior 53.) The Hague:
Mouton.
Chafe, Wallace L.
1982 "Integration and involvement in speaking, writing, and oral litera-
ture", in: Deborah Tannen (ed.), 35-53.
1985 "Linguistic differences produced by differences between speaking
and writing", in: David R. Olson—Nancy Torrance—Angela Hil-
yard (eds.), 105-123.
Duncan-Rose, Caroline—Theo Vennemann (eds.)
1988 On language, rhetorica, phonologica, syntactica. A Festschrift for
Robert P. Stockwell. London—New York: Routledge.
Ehrlich, Susan
1990 Point of view. A linguistic analysis of literary style. London—New
York: Routledge.
Enkvist, Nils Erik
1988 "Styles as parameters in text strategy", in: Willie van Peer (ed.),
125-151.
Fernändez, Francisco—Miguel Fuster—Juan Jose Calvo (eds.)
1994 English historical linguistics 1992. (Current Issues in Linguistic
Theory 113.) Amsterdam—Philadelphia: John Benjamins.
Halliday, Μ. A. K.
1979 "Modes and meaning and modes of expression: Types of gram-
matical structure, and their determination by different semantic
functions", in: D. J. Allerton—Edward Carney—David Holdcrofit
(eds.), 57-79.
 Reconstructing a grammar of point of view 341

1985 An introduction to functional grammar. London: Edward Arnold.

Halliday, Μ. A. K.—Ruqaiya Hasan
1976 Cohesion in English. London: Longman.
1985 Language, context, and text: Aspects of language in social-semiotic
[1990] perspective. Oxford: Oxford University Press.
Hanson, Kristin
1987 "On subjectivity and the history of epistemic expressions in Eng-
lish", in: Barbara Need—Eric Schiller—Anna Bosch, 133-147.
Helsinki Corpus
1991 The Helsinki Corpus of English Texts. Helsinki: Department of
English, University of Helsinki.
Hopper, Paul J.—Elizabeth Closs Traugott
1993 Grammaticalization. Cambridge: Cambridge University Press.
Kress, Gunther
1989 "History and language: Towards a social account of linguistic
change", Journal of Pragmatics 13/3: 445-466.
Kytö, Merja (comp.)
1996 Manual to the diachronic part of the Helsinki Corpus of English
Texts: Coding conventions and lists of source texts. (3rd edition.)
Helsinki: Department of English, University of Helsinki.
Langacker, R.
1990 "Subjectification", Cognitive Linguistics 1/1: 5-38.
Lehmann, Christian
1985 "Grammaticalization: Synchronic variation and diachronic change",
Lingua e Stile 20: 303-318.
Lehmann, Winfred P.—Yakov Malkiel (eds.)
1982 Perspectives on historical linguistics. (Current Issues in Linguistic
Theory 24.) Amsterdam—Philadelphia: John Benjamins.
Meurman-Solin, Anneli
1993a Variation and change in early Scottish prose. Studies based on the
Helsinki Corpus of Older Scots. (Annales Academiae Scientiarum
Fennicae 65.) Helsinki: Suomalainen Tiedeakatemia.
1993b "The author-addressee relationship and the marking of stance in the
characterization of sixteenth- and seventeenth-century genre styles",
New Horizons in Stylistics? (Belgisch Tijdschrift voor Filologie en
Geschiedenis) 71/3: 733-745.
1994 "On the evolution of prose genres in Older Scots", Nowele, 23: 91—
138.
1995 "Marking of stance in Early Modem English imaginative narration",
in: Wolfgang Görtschacher—Holger Klein (eds.), Narrative strat-
egies in early English fiction. Lewiston, NY—Salzburg: Edwin
Mellen, 25-52.
342 Anne Ii Meurman-Solin

forthc. "Point of view in Scottish and English genre styles in 1500-1700".

A paper read at the 8th International Conference of English Histori-
cal Linguistics. Edinburgh, September 1994.
Nevalainen, Terttu
1991 BUT, ONLY, JUST. Focusing adverbial change in Modern English
1500-1900. (Memoires de la Societ6 Neophilologique de Helsinki
51.) Helsinki: Societe Neophilologique.
Nevalainen, Terttu—Helena Raumolin-Brunberg
1994 "Sociolinguistics and language history: The Helsinki Corpus of
Early English Correspondence", Hermes 13: 135-143.
Nikiforidou, Vassiliki—Mary VanClay—Mary Niepokuj—Deborah Feder (eds.)
1986 Proceedings of the Twelfth Annual Meeting of the Berkeley Linguis-
tics Society, Berkeley: Berkeley Linguistics Society, University of
California.
Olson, David R.—Nancy Torrance—Angela Hilyard (eds.)
1985 Literacy, language and learning·. The nature and consequences of
reading and writing. Cambridge: Cambridge University Press.
Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik
1985 A comprehensive grammar of the English language. London—New
York: Longman.
Rissanen, Matti—Ossi Ihalainen—Terttu Nevalainen—Irma Taavitsainen (eds.)
1992 History of Englishes: New methods and interpretations in historical
linguistics. Berlin: Mouton de Gruyter.
Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.)
1993 Early English in the computer age: Explorations through the Hel-
sinki Corpus. Berlin—New York: Mouton de Gruyter.
Simpson, Paul
1993 Language, ideology and point of view. London—New York: Rout-
ledge.
Swan, Toril
1988 Sentence adverbials in English: A synchronic and diachronic inves-
tigation. Diss. University of Tromso.
Taavitsainen, Irma
1993 "Genre/subgenre styles in Late Middle English", in: Rissanen et al.
(eds.), 171-200.
1994 "On the evolution of scientific writings from 1375 to 1675: Reper-
toire of emotive features", in: Francisco Fernändez—Miguel Fus-
ter—Juan Jos6 Calvo (eds.), 329-342.
Tannen, Deborah (ed.)
1982 Spoken and written language: Exploring orality and literacy. (Ad-
vances in Discourse Processes 9.) Norwood, New Jersey: Ablex
Publishing Corporation.
 Reconstructing a grammar of point of view 343

Traugott, Elizabeth Closs

1982 "From propositional to textual and expressive meanings: Some se-
mantic-pragmatic aspects of grammaticalization", in: Winfred P.
Lehmann—Yakov Malkiel (eds.), 245-271.
1986 "From polysemy to internal semantic reconstruction", Vassiliki Niki
foridou—Mary VanClay—Mary Niepokuj—Deborah Feder (eds.),
539-550.
1988a "Is internal semantic-pragmatic reconstruction possible?", in: Caro-
line Duncan-Rose—Theo Vennemann (eds.), 128-144.
1988b "Pragmatic strengthening and grammaticalization", in: Shelley Ax-
maker—Annie Jaisser—Helen Singmaster (eds.), 406-416.
1990 "From less to more situated in language: The unidirectionality of
semantic change", in: Sylvia Adamson—Vivien Law—Nigel Vin-
cent—Susan Wright (eds.), 497-517.
Traugott, Elizabeth Closs—Suzanne Romaine
1985 "Some questions for the definition of 'style' in socio-historical lin-
guistics", Folia Linguistica Historica 6: 7-39.
van Peer, Willie (ed.)
1988 The taming of the text. Explorations in language, literature and cul-
ture. London—New York: Routledge.
Bibliography

This short title list presents bibliographical details of the Helsinki Corpus
texts or groups of texts referred to in the present study. The short titles
are followed, in parentheses, by the abbreviated titles used for these texts
in the Manual to the Diachronic Part of the Helsinki Corpus of English
Texts (see Introduction to the present volume) and by the bibliographical
details of the editions used. Dates of possible previous editions are given
in square brackets. For more information on the extracts sampled (page
numbers, manuscripts followed, more exact composition dates, etc.) the
reader is referred to Parts Two and Three of the Manual.

JElfric's Letter to Wulfsige; ALlfric's Letter to Wulfstan

From: Die Hirtenbriefe sElfrics in altenglischer und lateinischer
Fassung. (Bibliothek der angelsächsischen Prosa IX.) Ed. Bernhard
Fehr. Hamburg: Verlag von Henri Grand, 1914:
Mfric 's Letter to Wulfsige (LWSIGE)
Mfric's Letters to Wulfstan (LWSTAN1; LWSTAN2)
Aelred of Rievaulx's De Institutione Inclusarum (AELR3; AELR4)
1984 Aelred of Rievaulx's De Institutione Inclusarum. (Early English Text
Society 287.) Eds. John Ayto—Alexandra Barratt. London.
Alfred's Boethius (BOETHAL)
1899 King Alfred's Old English Version of Boethius De Consolatione
Philosophiae. Ed. Walter John Sedgefield. Oxford: The Clarendon
Press.
Alfred's Curα Pastoralis (CP)
1958 King Alfred's West-Saxon Version of Gregory's Pastoral Care.
[1871] Parts MI. (Early English Text Society, Original Series 45, 50.) Ed.
Henry Sweet. London.
Ancrene Wisse (ANCR)
1962 Ancrene Wisse. (Early English Text Society 249.) Ed. John Ronald
Reuel Tolkien. London.
Armin, Robert, A Nest of Ninnies (ARMIN)
1842 Fools and Jesters: With a Reprint of Robert Armin 's Nest of Ninnies.
[1608] London: The Shakespeare Society.
346 Bibliography

The Arraignment of the Earles of Essex and Southampton (ESSEX)

1873 "The Arraignment of the Earles of Essex and Southampton in West-
minster Hall on Thursdaye ye XIX. of Februarie 1600", in The Dr.
Farmer Chetham MS. Being a Commonplace-book in the Chetham
Library, Manchester. Temp. Elizabeth, James I. and Charles I. Con-
sisting of Verse and Prose, Mostly Hitherto Unpublished. (Chetham
Society LXXXIX.) Ed. Alexander B. Grosart. Manchester: The Chet-
ham Society.
Ascham, Roger, The Scholemaster (ASCH)
1870 The Scholemaster. (English Reprints.) Ed. Edward Arber. London.
The Authorized Version (AUTHNEW; AUTHOLD)
1911 The Holy Bible. An Exact Reprint in Roman Type, Page for Page of
the Authorized Version Published in the Year 1611. With an intro-
duction by Alfred W. Pollard. London—Oxford—New York: Henry
Frowde and Oxford University Press.
Bacon, Francis, Advancement of Learning (BACON)
1970 The Twoo Bookes of the Proficience and Advancement of Learning
(1605). (English Experience 218.) Amsterdam: Theatrvm Orbis Ter-
rarvm Ltd.—New York: Da Capo Press. [Facsimile.]
Barrington Family Letters (EVERARD; JBARRING; MASHAM; TBARRING)
1983 Barrington Family Letters, 1628-1632. (Camden Fourth Series 28.)
Ed. Arthur Searle. London.
Beaumont Papers (EBEAUM)
1884 Beaumont Papers. Letters Relating to the Family of Beaumont, of
Whitley, Yorkshire, from the Fifteenth to the Seventeenth Centuries.
Ed. W. D. Macray. London: Nichols.
Bede's Ecclesiastical History (BEDEHE)
1959 The Old English Version of 'Bede's Ecclesiastical History of the Eng-
[1890; lish People'. Parts 1,1; 1,2. (Early English Text Society, Original Se-
1891] ries 95, 96.) Ed. Thomas Miller. London.
Behn, Aphra, Oroonoko (BEHN)
1960 Shorter Novels: Seventeenth Century. Ornatus & Artesia, Oroonoko,
Isle of Pines, Incognita. Ed. Philip Henderson. London: Dent—New
York: Dutton.
The Benedictine Rule (ME) (BENEDME)
1902 Three Middle-English Versions of the Rule of St. Benet and Two
Contemporary Rituals for the Ordination of Nuns. (Early English
Text Society, Original Series 120.) Ed. Ernst A. Kock. London.
Beowulf (BEOW)
1953 Beowulf and Judith. (ASPR IV.) Ed. Elliott Van Kirk Dobbie. New
York: Columbia University Press.
 Bibliography 347

Bible See The Authorized Version; Rolle, The Psalter, Rushworth Gos-
pels', Tyndale, Five Books of Moses·, Tyndale, The New Testament·,
Wycliffe, The Old Testament (Genesis and Numbers)·, Wycliffe and
Purvey, The New Testament (John).
The Blickling Homilies (BLICK2; BLICK6; BLICK12; BLICK10; BLICK 17)
1967 The Blickling Homilies. (Early English Text Society, Original Series
[1874- 58, 63, 73.) Ed. Richard Morris. London.
1880]
Blundevile, Thomas, The Tables of the Three Speciall Right Lines Belonging to
a Circle (BLUNDEV)
1597 Α Briefe Description of the Tables of the Three Speciall Right Lines
Belonging to a Circle, Called Signes, Lines Tangent, and Lines Se-
cant. London: John Windet.
Bodley Homilies (BOD5; BODIO; BOD12)
1962 Twelfth-century Homilies in MS. Bodley 343. Part I. (Early English
[1909] Text Society, Original Series 137.) Ed. A. O. Belfour.
Boethius (De Consolatione Philosophiae)
See Alfred's Boethius·, Chaucer, Boethius·, Colville, Boethius', Eliza-
beth I, Queen Elizabeth's Englishings of Boethius, Plutarch, &c;
Preston, Boethius.
The Book of Margery Kempe
See Kempe, Margery.
The Book of Vices and Virtues (VICES4)
1942 The Book of Vices and Virtues. A Fourteenth Century English Trans-
lation of the Somme le Roi of Lorens d Orleans. (Early English Text
Society 217.) Ed. W. Nelson Francis. London.
Boyle, Robert, Electricity & Magnetism (BOYLE)
1927 Electricity & Magnetism, 1675-6. (Old Ashmolean Reprints 7.) Se-
ries ed. R. W. T. Gunther. [Facsimile.] Oxford: University of Ox-
ford.
Brinsley, John, Ludus Literarius, or the Grammar Schoole (BRINSLEY)
1917 Ludus Literarius or the Grammar Schoole (1627). Ed. Ε. T. Cam-
pagnac. Liverpool: University Press—London: Constable.
The Brut or the Chronicles of England (BRUT)
1960 The Brut or the Chronicles of England. Part I. (Early English Text
[1906] Society, Original Series 131.) Ed. Friedrich W. D. Brie. London.
Burnet, Gilbert, Burnet's History of My Own Time (BURNETCHA)
1897, Burnet's History of My Own Time. Part I: The Reign of Charles the
1900 Second. Vols. I—II. Ed. Osmund Airy. Oxford: Clarendon Press.
348 Bibliography

Burnet, Gilbert, The Life and Death of John Earl of Rochester (BURNETROC)
1972 Some Passages of the Life and Death of the Right Honourable John,
Earl of Rochester, Who Died the 26th of July, 1680. Written by His
Own Direction on His Death-bed, by Gilbert Burnet, D. D. (London,
1680). Menston: The Scolar Press. [Facsimile.]
Capgrave, John, Abbreuiacion of Cronicles (CAPCHR)
1983 John Capgrave's Abbreuiacion of Cronicles. (Early English Text
Society 285.) Ed. Peter J. Lucas. Oxford.
Capgrave's Sermon (CAPSERM)
1971 John Capgrave's Lives of St. Augustine and St. Gilbert ofSempring-
[1910] ham, and a Sermon. (Early English Text Society, Original Series
140.) Ed. J. J. Munro. New York.
Caxton, William, The History of Reynard the Fox (REYNARD)
1970 The History of Reynard the Fox. Translated from the Dutch Origi-
nal by William Caxton. (Early English Text Society 263.) Ed. Nor-
man Blake. London.
Caxton, William, The Prologues and Epilogues (CAXTON)
1956 The Prologues and Epilogues of William Caxton. (Early English Text
[1928] Society 176.) Ed. W. J. B. Crotch. London.
Cecil, Robert, The Edmondes Papers (RCECIL)
1913 The Edmondes Papers. A Selection from the Correspondence of Sir
Thomas Edmondes, Envoy from Queen Elizabeth at the French
Court. Ed. Geoffrey G. Butler. London: J. B. Nichols.
The Cely Letters (GCELY; RCELY)
1975 The Cely Letters 1472-1488. (Early English Text Society 273.) Ed.
Alison Hanham. London.
Charles II, Essex Papers (CHARLES)
1890 Essex papers. Vol. I (1672-1679). (Camden Society, New Series
XLVII.) Ed. Osmund Airy. Westminster.
Chaucer, Geoffrey
From: The Riverside Chaucer. General editor Larry D. Benson.
[Based on The Works of Geoffrey Chaucer, edited by F. N. Robin-
son.] (3rd edition.) Boston: Houghton Mifflin, 1987:
Astrolabe (ASTR)
Boethius (BOETHCH)
The Canterbury Tales (CTBATH; CTMEL; CTMERCH; CTPARS;
CTPROL; CTSUMM)
The General Prologue (CTPROL)
The Merchant's Tale (CTMERCH)
The Parson's Tale (CTPARS)
The Summoner's Tale (CTSUMM)
The Tale of Melibee (CTMEL)
The Wife of Bath's Prologue (CTBATH)
 Bibliography 349

Clifford Letters of the Sixteenth Century (ECUMBERL; KSCROPE)

1962 Clifford Letters of the Sixteenth Century. (Surtees Society CLXXII.)
Ed. A. G. Dickens. Durham—London.
The Cloud of Unknowing (CLOUD)
1958 The Cloud of Unknowing and the Book of Privy Counselling. (Early
[1944] English Text Society 218.) Ed. Phyllis Hodgson. London.
Clowes, William, Treatise for the ArtificialI Cure of Struma (CLOWES)
1970 Treatise for the Artificiall Cure of Struma, 1602. (The English Ex-
perience 238.) Amsterdam: Theatrvm Orbis Terrarvm—New York:
Da Capo Press. [Facsimile.]
Colville, George, Boethius (BOETHCO)
1897 Boethius' Consolation of Philosophy, Translated from the Latin by
George Colville, 1556. (The Tudor Library V.) Ed. Ε. B. Bax. Lon-
don: David Nutt.
Confessio A mantis
See Gower, John, Confessio Amantis.
Correspondence of the Family of Haddock (NHADD; RHADDJR; RHADDSR)
1965 The Camden Miscellany, Volume the Eighth: Containing ... Corre-
[1883] spondence of the Family of Haddock, 1657-1719. (Camden Society,
New Series XXXI.) Ed. Edward Maunde Thompson. London.
Correspondence of the Family ofHatton (ALHATTON; ANHATTON; CHATTON;
EHATTON; FHATTON)
1878 Correspondence of the Family of Hatton Being Chiefly Letters Ad-
dressed to Christopher First Viscount Hatton, A. D. 1601-1704. Vol.
I. (Camden Society, New Series XXII.) Ed. Edward Maunde
Thompson. Westminster.
Correspondence of Lady Katherine Paston (KPASTON; WPASTON2)
1941 The Correspondence of Lady Katherine Past on, 1603-1627. (Nor-
folk Record Society XIV.) Ed. Ruth Hughey. London: Norfolk Re-
cord Society.
The Correspondence of Sir Thomas More (MORELET; MROPER)
1947 The Correspondence of Sir Thomas More. Ed. Elizabeth Frances
Rogers. Princeton: Princeton University Press.
Coverte, Robert, A Trve and Almost Incredible Report of an Englishman
(COVERTE)
1971 A Trve and Almost Incredible Report of an Englishman, 1612. (The
English Experience 302.) Amsterdam: Theatrvm Orbis Terrarvm
Ltd.—New York: Da Capo Press. [Facsimile.]
Cursor Mundi (CURSOR)
1874- Cursor Mundi. (Early English Text Society, Original Series 57, 59,
1878 62, 66, 68.) Ed. Richard Morris. London.
350 Bibliography

The Cyrurgie of Guy de Chauliac (CHAUL)

1971 The Cyrurgie of Guy de Chauliac. (Early English Text Society 265.)
Ed. Margaret Sinclair Ogden. London.
Deloney, Thomas, The Pleasaunt History of.Jack ofNewberie (DELONEY)
1961 The Novels of Thomas Deloney. Ed. Merritt E. Lawlis. Blooming-
ton: Indiana University Press.
Depositions (DEPOSC)
1984 An Anthology of Chancery English. Eds. John H. Fisher—Malcolm
Richardson—Jane L. Fisher. Knoxville: The University of Tennes-
see Press.
Diaries and Letters of Philip Henry (PHENRY)
1882 Diaries and Letters of Philip Henry, M. A. of Broad Oak, Flintshire,
A. D. 1631-1696. Ed. Matthew Henry Lee. London: Kegan Paul,
Trench & Co.
Diets ofCato (DICTS)
1972 "The Old English Diets ofCato", Anglia 90: 1-42. Ed. R. S. Cox.
Digby Plays (DIGBY)
1982 The Late Medieval Religious Plays of Bodleian MSS Digby 133 and
Ε Museo 160. (Early English Text Society 283.) Eds. Donald C.
Baker—John L. Murphy—Louis B. Hall, Jr. Oxford.
Earl of Essex
See The Arraignment of the Earles of Essex and Southampton.
The Edmondes Papers (EDMONDES; ELIZ; RCECIL)
1913 The Edmondes Papers. A Selection from the Correspondence of Sir
Thomas Edmondes, Envoy from Queen Elizabeth at the French
Court. Ed. Geoffrey G. Butler. London: J. B. Nichols.
Edmund
See The Life of St. Edmund (in Middle English Religious Prose).
Edward VI, Journal (EDWARD)
1963 Literary Remains of King Edward the Sixth. Vol. II. (Burt Franklin
[1857] Research & Source Works Series 51.) Ed. John Gough Nichols. New
York.
Elizabeth I, Queen Elizabeth's Englishings ofBoethius, Plutarch, &c (BOETHEL)
1899 Queen Elizabeth's Englishings of Boethius, De Consolatione Philo-
sophiae, A. D. 1593, Plutarch, De Curios itate, Horace, De Arte
Poetica (Part), A. D. 1598. (Early English Text Society, Original
Series 113.) Ed. Caroline Pemberton. London.
Elyot, Thomas, The Boke Named the Gouernour (ELYOT)
1907 The Boke Named the Gouernour (1531). (Everyman's Library edited
by Ernest Rhys.) With an introduction by Foster Watson. London—
New York: Dent—Dutton.
 Bibliography 351

The Equatorie of the Planetis (EPLANETS)

1955 The Equatorie of the Planetis. Ed. Derek J. Price. Cambridge: Cam-
bridge University Press.
Essex Papers (CHARLES; COMMISS)
1890 Essex Papers. Vol. I (1672-1679). (Camden Society, New Series
XLVII.) Ed. Osmund Airy. Westminster.
Essex, trial
See The Arraignment of the Earles of Essex and Southampton.
Evelyn, John, The Diary of John Evelyn (EVELYN)
1959 The Diary of John Evelyn. Ed. E. S. de Beer. London—New York—
Toronto: Oxford University Press.
Fabyan, Robert, The New Chronicles of England and France (FABYAN)
1516 The New Chronicles of England and France. London: Pynson.
Farquhar, George, The Beaux Stratagem (FARQUHAR)
1972 The Beaux Stratagem, 1707. Menston: The Scolar Press. [Facsimile.]
The Ferrar Papers (NFERRAR; RFERRAR)
1938 The Ferrar Papers, Containing a Life of Nicholas Ferrar, the Wind-
ing-sheet (an Ascetic Dialogue), a Collection of Short Moral His-
tories, a Selection of Family Letters. Ed. B. Blackstone. Cambridge:
Cambridge University Press.
Fiennes, Celia, The Journeys of Celia Fiennes (FIENNES)
1947 The Journeys of Celia Fiennes. Ed. Christopher Morris. London: The
Cresset Press.
Fisher, John, The English Works of John Fisher (FISHER)
1935 The English Works of John Fisher, Bishop of Rochester. Part I. (Ear-
[1876] ly English Text Society, Extra Series 27.) Ed. John Ε. B. Mayor.
London.
Fitzherbert, Anthony, The Book of Husbandry (FITZH)
1965 The Book of Husbandry (1534). (English Dialect Society 37.) Ed.
[1882] Walter W. Skeat. Vaduz: Kraus.
Fitzjames, Richard, Sermo Die Lüne (FITZJ)
1907 Sermo Die Lüne in Ebdomada Pasche. (Westminster: Wynkyn de
Wörde (1495?).) Ed. Francis Jenkinson. Cambridge: Cambridge Uni-
versity Press. [Facsimile.]
Forman, Simon, The Autobiography and Personal Diary of Dr. Simon Forman
(FORMAN)
1849 The Autobiography and Personal Diary of Dr. Simon Forman, the
Celebrated Astrologer, from A. D. 1552, to A. D. 1602. Ed. James
Orchard Halliwell. London, privately printed.
Fox, George, The Journal of George Fox (FOX)
1911 The Journal of George Fox. Ed. Norman Penney, with an introduc-
tion by T. Edmund Harvey. Cambridge: Cambridge University Press.
352 Bibliography

Fryer, John, A New Account of East India and Persia (FRYER)

1909, A New Account of East India and Persia, Being Nine Years' Travels,
1912 1672-1681. Vols. I, II. Ed. William Crooke. London: The Hakluyt
Society.
Gaytryge, Dan Jon, Sermon (GAYTR)
1969 Religious Pieces in Prose and Verse. (Early English Text Society,
[1914] Original Series 26.) Ed. George G. Perry. New York.
Gifford, George, A Dialogue Concerning Witches and Witchcraftes (GIFFORD)
1931 A Dialogue Concerning Witches and Witchcraftes, 1593. (Shake-
speare Association Facsimiles 1.) With an introduction by Beatrice
White. London: Humphrey Milford—Oxford University Press.
Gower, John, Confessio Amantis (GOWER)
1957 The English Works of John Gower. Vols. I, II. (Early English Text
[1900, Society, Extra Series 81, 82.) Ed. G. C. Macaulay. London.
1901]
Gregory, William, The Historical Collections of a Citizen of London (CHRLOND)
1876 The Historical Collections of a Citizen of London in the Fifteenth
Century. (Camden Society, New Series XVII.) Ed. James Gairdner.
Westminster.
Gregory's Dialogues', MS C; MS Η
From: Bischofs Wcerferth von Worcester Ubersetzung der Dialoge
Gregors des Grossen. (Bibliothek der angelsächsischen Prosa V.) Ed.
Hans Hecht. Leipzig: Georg H. Wigand's Verlag, 1900:
Gregory 's Dialogues MS C (GDC)
Gregory 's Dialogues MS Η (GDH)
Hali Meidhad (HMAID)
1977 The Katherine Group. Edited from MS. Bodley 34. (Bibliothdque de
la Facultö de Philosophie et Lettres de l'Universit6 de Liöge CCXV.)
Ed. Simonne T. R. O. D'Ardenne. Paris: Societ6 d'Edition "Les
Beiles Lettres".
Handlyng Synne
See Mannyng, Robert.
Harley, Brilliana, Letters of the Lady Brilliana Harley (HARLEY)
1854 Letters of the Lady Brilliana Harley, Wife of Sir Robert Harley, of
Brampton Bryan, Knight of the Bath. (Camden Society LVIII.) Ed.
Thomas Taylor Lewis. London.
Harman, Thomas, A Caveat or Warening for Commen Cursetors (HARMAN)
1937 A Caveat or Warening for Commen Cursetors Vulgarely Called Va-
[1869, gabones. [From the 3rd edition of 1567... Collated with the 2nd
1898] edition of 1567 in the Bodleian Library, Oxford, and with the reprint
of the 4th edition of 1573.] (Early English Text Society, Extra Series
9.) Eds. Edward Viles—Frederick J. Furnivall. London.
 Bibliography 353

Hayward, John, Annals of the First Four Years of the Reign of Queen Elizabeth
(HAYWARD)
1840 Annals of the First Four Years of the Reign of Queen Elizabeth.
(Camden Society VII.) Ed. John Bruce. London.
Henry V, Letter(s) (HENRY5C; HENRY5L)
1984 An Anthology of Chancery English. Eds. John H. Fisher—Malcolm
Richardson—Jane L. Fisher. Knoxville: The University of Tennes-
see Press.
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
Hilton's Eight Chapters (HILTON)
1967 Walter Hilton's Eight Chapters on Perfection. Ed. Fumio Kuriya-
gawa. Tokyo: The Keio Institute of Cultural and Linguistic Studies.
Hoby, Margaret, Diary of Lady Margaret Hoby (HOBY)
1930 Diary of Lady Margaret Hoby, 1599-1605. Ed. Dorothy M. Meads.
London: George Routledge.
Holy Rood-tree (HROOD)
1894 History of the Holy Rood-tree. (Early English Text Society, Original
Series 103.) Ed. Arthur S. Napier. London.
Hooke, Robert, Micrographia (HOOKE)
1968 Early Science in Oxford. Vol. XIII: The Life and Work of Robert
[1938] Hooke. Part V. Ed. R. T. Gunther. London: Dawsons of Pall Mall.
[Facsimile.]
Hooker, Richard, Two Sermons upon Part of S. Judes Epistle (HOOKER)
1969 Two Sermons upon Part ofS. Judes Epistle, 1614. (The English Ex-
perience 195.) Amsterdam: Theatrvm Orbis Terrarvm—New York:
Da Capo Press. [Facsimile.]
Hoole, Charles, A New Discovery of the Old Art of Teaching Schoole (HOOLE)
1969 A New Discovery of the Old Art of Teaching Schoole (1660). (Eng-
lish Linguistics 1500-1800 133.) Ed. Robin Carfrae Alston. Mens-
ton: The Scolar Press. [Facsimile.]
A Hundred Mery Talys (MERRYTAL)
1866 A Hundred Mery Talys, from the Only Perfect Copy Known. (Shake-
speare's Jest Book.) Ed. Herman Oesterley. London: John Russell
Smith.
In Die Innocencium (INNOC)
1875 "Two Sermons Preached by the Boy Bishop, at St. Paul's Temp.
Henry VII, and at Gloucester, Temp. Mary", Camden Society Mis-
cellany VII. (Camden Society, New Series XIV.) Ed. John Gough
Nichols. London.
354 Bibliography

Indenture (INDENT)
1984 An Anthology of Chancery English. Eds. John H. Fisher—Malcolm
Richardson—Jane L. Fisher. Knoxville: The University of Tennes-
see Press.
Judgements (JUDG)
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
Julian of Norwich's Revelations (JULNOR)
1978 Julian of Norwich's Revelations of Divine Love. The Shorter Ver-
sion Ed. from B.L. Add. MS 37790. (Middle English Texts.) Ed.
Frances Beer. Heidelberg: Carl Winter Universitätsverlag.
Juliane (JULME)
1977 The Katherine Group. Edited from MS. Bodley 34. (Bibliothöque de
la Facult6 de Philosophie et Lettres de l'Universite de Liöge CCXV.)
Ed. Simonne T. R. O. D'Ardenne. Paris: Societe d'Edition "Les
Beiles Lettres".
Katherine (KATH); Katherine Group (HMAID; JULME; ΚΑΤΗ; MARGME;
SWARD)
1977) The Katherine Group. Edited from MS. Bodley 34. (Bibliotheque de
la Faculte de Philosophie et Lettres de l'Universite de Lifcge CCXV.)
Ed. Simonne T. R. O. D'Ardenne. Paris: Societe d'Edition "Les
Belies Lettres".
Kempe, Margery, The Book of Margery Kempe (KEMPE)
1940 The Book of Margery Kempe. Vol. I. (Early English Text Society
212.) Eds. Sanford Brown Meech—Hope Emily Allen. London.
The Knyvett Letters (KNYVETT)
1949 The Knyvett Letters (1620-1644). Ed. Bertram Schofield. London:
Constable.
Lambeth Homilies (LAMB1; LAMB3; LAMB4; LAMB7; LAMB8; LAMB 14;
LAMB 15)
1969 Old English Homilies and Homiletic Treatises... of the Twelfth and
[1868] Thirteenth Centuries. [First series.] (Early English Text Society,
Original Series 29, 34.) Ed. Richard Morris. New York.
Langford, Τ., Plain and Full Instructions to Raise All Sorts of Fruit-trees
(LANGF)
1699 Plain and Full Instructions to Raise All Sorts of Fruit-trees That
Prosper in England. London: R. Chiswell.
A Late Middle English Treatise on Horses (HORSES)
1978 A Late Middle English Treatise on Horses. (Stockholm Studies in
English XLVII.) Ed. Anne Charlotte Svinhufvud. Stockholm: Alm-
qvist & Wiksell International.
 Bibliography 355

Latimer, Hugh, Sermons = A Sermon on the Ploughers; Seven Sermons Before

Edward VI (LATIMER)
1868, Sermon on the Ploughers, 18 January 1549; Seven Sermons Before
1869 Edward VI, on Each Friday in Lent, 1549. (English Reprints.) Ed.
Edward Arber. London: Alex. Murray.
A Latin Technical Phlebotomy (PHLEB)
1984 A Latin Technical Phlebotomy and Its Middle English Translation.
(Transactions of the American Philosophical Society 74, Part 2.) Eds.
Linda E. Voigts—Michael R. McVaugh. Philadelphia.
Laws (Early Modern English)
See The Statutes of the Realm (III, IV, VII).
Laws (Middle English)
See The Statutes of the Realm (II).
Layamon's Brut (LAYBR)
1963, Layamon's Brut. Vols. I, II. (Early English Text Society 250, 277.)
1978 Eds. George Leslie Brook—R. F. Leslie. London.
Leland, John, The Itinerary of John Leland (LELAND)
1964 The Itinerary of John Leland in or About the Years 1535-1543. Vol.
I, Parts I to III. Ed. Lucy Toulmin Smith. London: Centaur Press.
Letters of John Pinney (JOPINNEY; JPINNEY)
1939 Letters of John Pinney 1679-1699. Ed. Geoffrey F. Nuttall. Lon-
don—New York—Toronto: Oxford University Press.
Letters of the Lady Brilliana Harley (HARLEY)
1854 Letters of the Lady Brilliana Harley, Wife of Sir Robert Harley, of
Brampton Bryan, Knight of the Bath. (Camden Society LVIII.) Ed.
Thomas Taylor Lewis. London.
Letters of Philip Gawdy (GAWDY)
1906 Letters of Philip Gawdy of West Harling, Norfolk, and of London to
Various Members of His Family, 1579-1616. Ed. Isaac Herbert
Jeayes. London: J. B. Nichols.
The 'Liber de Diversis Medicinis' in the Thornton Manuscript (THORN)
1938 The 'Liber De Diversis Medicinis' in the Thornton Manuscript (MS.
Lincoln Cathedral A.5.2.). (Early English Text Society 207.) Ed.
Margaret Sinclair Ogden. London.
The Life of St. Edmund (in Middle English Religious Prose) (EDMUND)
1972 Middle English Religious Prose. (York Medieval Texts.) Ed. Norman
Blake. London: Edward Arnold.
Locke, John, Directions Concerning Education (LOCKE)
1933 Directions Concerning Education. Ed. Frederic George Kenyon. Ox-
ford: The Roxburghe Club.
London Letters (LLETT)
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
356 Bibliography

Machyn, Henry, The Diary of Henry Machyn (MACHYN)

1848 The Diary of Henry Machyn, Citizen and Merchant-taylor of Lon-
don, from A. D. 1550 to A. D. 1563. (Camden Society XLII.) Ed.
John Gough Nichols. London: J. B. Nichols.
Madox, Richard, The Diary of Richard Madox (MADOX)
1976 An Elizabethan in 1582: The Diary of Richard Madox, Fellow of All
Souls. Ed. Elizabeth Story Donno. London: Hakluyt Society.
Malory, Thomas, Morte Darthur (MALORY)
1954 The Works of Sir Thomas Malory. Ed. Eug£n Vinaver. London: Ox-
ford University Press.
Mandeville's Travels (MAND)
1919 Mandeville's Travels, Translated from the French of Jean d'Outre-
meuse. Vol. I. (Early English Text Society, Original Series 153.) Ed.
P. Hamelius. London.
Mankind (MANK)
1969 The Macro Plays. (Early English Text Society 262.) Ed. Mark
Eccles. London.
Margarete (MARGME)
1977 The Katherine Group. Edited from MS. Bodley 34. (Bibliotheque de
la Faculte de Philosophie et Lettres de l'Universit6 de Lifcge CCXV.)
Ed. Simonne T. R. O. D'Ardenne. Paris: Societ6 d'Edition "Les
Belles Lettres".
Markham, Gervase, Countrey Contentments (MARKHAM)
1973 Countrey Contentments, 1615. (The English Experience 613.) Ams-
terdam: Theatrvm Orbis Terrarvm Ltd.—New York: Da Capo Press.
[Facsimile.]
Metham, John
From: The Works of John Metham Including the Romance of Amor-
yus and Cleopes. (Early English Text Society, Original Series 132.)
Ed. Hardin Craig. London, 1916:
Days of the Moon (METHD)
Physiognomy (METHP)
Middle English Sermons (MS Royal) (ROYAL)
1940 Middle English Sermons, Edited from British Museum MS. Royal 18
Β. XXIII. (Early English Text Society 209.) Ed. Woodburn O. Ross.
London.
Middleton, Thomas, A Chaste Maid in Cheapside (MIDDLET)
1969 A Chaste Maid in Cheapside, 1630. Menston: The Scolar Press.
[Facsimile.]
Milton, John, The History of Britain (MILTON)
1932 The History of Britain, That Part Especially Now Call'd England
(1670). The Works of John Milton. Vol. X. Ed. George Philip Krapp.
New York: Columbia University Press.
 Bibliography 357

Mirk's Festial (MIRK)

1905 Mirk's Festial: A Collection of Homilies, by Johannes Mirkus (John
Mirk). Part I. (Early English Text Society, Extra Series 96.) Ed.
Theodor Erbe. London.
More, Thomas, The History of King Richard III (MORERIC)
1963 The Complete Works of St. Thomas More. Vol. II. Ed. Richard S.
Sylvester. New Haven—London: Yale University Press.
Mowntayne, Thomas, Narratives of the Days of the Reformation (MOWNTAYNE)
1859 Narratives of the Days of the Reformation, Chiefly from the Manu-
scripts of John Foxe the Martyrologist. (Camden Society LXXVII.)
Ed. John Gough Nichols. London.
The Northern Homily Cycle (the Expanded Version) (NHOM)
1973, The Northern Homily Cycle. Parts II, III. [The expanded version in
1984 MSS Harley 4196 and Cotton Tiberius Ε VII.] (Societe Neophilo-
logique de Helsinki 41, 43.) Ed. Saara Nevanlinna. Helsinki: Societe
Neophilologique.
Original Letters (Illustrative of English History) (BEDYLL; CONWAY; COUNC;
CROMWELL; GCROMW; HENRY; LORDS; MOREWOL; SOMERS; SPENCER;
THOWARD; TRINCOLL; TUNSTALL; WCECIL; WOLSEY)
1846 Original Letters, Illustrative of English History; Including Numer-
ous Royal Letters. Ed. Henry Ellis. London: Richard Bentley.
Original Letters of Eminent Literary Men (STRYPE)
1843 Original Letters of Eminent Literary Men of the Sixteenth, Seven-
teenth, and Eighteenth Centuries. (Camden Society XXIII.) Ed. Hen-
ry Ellis. London.
The Ormulum (ORM)
1878 The Ormulum. Vols. I—II. With the notes and glossary of R. M.
White. Ed. Robert Holt. Oxford: The Clarendon Press.
Oxinden, Henry
See The Oxinden and Peyton Letters below (HOXINDEN).
The Oxinden and Peyton Letters (EOXINDEN; HOXINDEN)
1937 The Oxinden and Peyton Letters, 1642-1670. Being the Corre-
spondence of Henry Oxinden of Barham, Sir Thomas Peyton of
Knowlton and Their Circle. Ed. Dorothy Gardiner. London: The
Sheldon Press—New York: The Macmillan Company.
The Oxinden Letters (KOXINDEN; PETTIT; PEYTON; PROUD; ROXINDEN)
1933 The Oxinden Letters 1607-1642. Being the Correspondence of Hen-
ry Oxinden of Barham and His Circle. Ed. Dorothy Gardiner. Lon-
don: Constable & Co. Ltd.
Past on Letters (CPASTON; JPASTON; MPASTON; WPASTON1)
1971 Paston Letters and Papers of the Fifteenth Century. Part I. Ed. Nor-
man Davis. Oxford: Clarendon Press.
358 Bibliography

Penny Merriments (= Samuel Pepys' Penny Merriments) (PENNY)

1976 Samuel Pepys' Penny Merriments. Ed. Roger Thompson. London:
Constable.
Pepys, Samuel, The Diary of Samuel Pepys (PEPYS)
1972, The Diary of Samuel Pepys. Vol. VII (1666); Vol. VIII (1667). Eds.
1974 Robert Latham—William Matthews. London: G. Bell.
Peri Didaxeon (PERI)
1866 Leechdoms, Wortcunning, and Starcraft of Early England. Vol. III.
(Rolls Series 35.) Ed. Oswald Cockayne. London.
Perrott (?), John, The History of That Most Eminent Statesman, Sir John Perrot
(PERROTT)
1728 The History of That Most Eminent Statesman, Sir John Perrott,
Knight of the Bath, and Lord Lieutenant of Ireland. Ed. Richard
Rawlinson. London.
The Peterborough Chronicle (PETERB)
1958 The Peterborough Chronicle, 1070-1154. Ed. Cecily Clark. London:
Oxford University Press.
Petitions (PET3; PET4)
1984 An Anthology of Chancery English. Eds. John H. Fisher—Malcolm
Richardson—Jane L. Fisher. Knoxville: The University of Tennes-
see Press.
Phlebotomy
See A Latin Technical Phlebotomy.
Plumpton Correspondence (APLUMPT; DPLUMPT; IPLUMPT; RPLUMPT;
WPLUMPT)
1839 Plumpton Correspondence. A Series of Letters, Chiefly Domestick,
Written in the Reigns of Edward IV. Richard HI. Henry VII. and
Henry VIII. (Camden Society IV.) Ed. Thomas Stapleton. London.
Polychronicon
See Trevisa, John.
Preston, Richard Lord Viscount, Boethius (BOETHPR)
1695 Anicius Manlius Severinus Boetius, of the Consolation of Philo-
sophy, in Five Books. Made English and Illustrated with Notes, by
the Right Honourable Richard Lord Viscount Preston. London, print-
ed by J. D. for Awnsham and John Churchill, at the Black Swan in
Pater-Noster Row; and Francis Hildyard Bookseller in York.
The Pricke of Conscience (PRICK)
1863 The Pricke of Conscience (Stimulus Conscientiae). Ed. Richard Mor-
ris. Berlin: A. Asher.
Proclamations (LPROCL)
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
 Bibliography 359

Purvey, John, The Prologue to the Bible (WYCPROL)

1850 The Holy Bible, Containing the Old and New Testaments, with the
Apocryphal Books, in the Earliest English Versions Made from the
Latin Vulgate by John Wycliffe and His Followers. Vol. I. Eds.
Josiah Forshall—Frederic Madden. Oxford: University Press.
Record, Robert, The First Principles of Geometrie (RECORD)
1974 The Path-way to Knowledg, Containing the First Principles of Geo-
metrie, 1551. (The English Experience 687.) Amsterdam: Theatrum
Orbis Terrarum—Norwood, N. J.: W. J. Johnson, Inc. [Facsimile.]
Returns (RET)
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
Rolle, Richard, Prose Treatises (RTREAT)
1921 English Prose Treatises of Richard Rolle de Hampole. (Early Eng-
[1866] lish Text Society, Original Series 20.) Ed. George G. Perry. London.
Rolle, Richard, The Psalter (RPSAL)
1884 The Psalter or Psalms of David and Certain Canticles with a Trans-
lation and Exposition in English by Richard Rolle of Hampole. Ed.
H. R. Bramley. Oxford: Clarendon Press.
Roper, William, The Lyfe of Sir Thomas Moore (ROPER)
1958 The Lyfe of Sir Thomas Moore, Knighte, Written by William Roper,
[1935] Esquire, Whiche Maried Margreat, Daughter of the Sayed Thomas
Moore. (Early English Text Society 197.) Ed. Elsie Vaughan Hitch-
cock. London.
Rushworth Gospels (RUSHW)
1871- The Holy Gospels in Anglo-Saxon, Northumbrian and Old Mercian
1887 Versions. Ed. Walter W. Skeat. Cambridge: Cambridge University
Press.
Sawles Warde (SWARD)
1977 The Katherine Group. Edited from MS. Bodley 34. (Bibliothöque de
la Faculte de Philosophie et Lettres de l'Universite de Liege CCXV.)
Ed. Simonne T. R. O. D'Ardenne. Paris: Societe d'Edition "Les
Beiles Lettres".
Selections from the Correspondence of Arthur Capel (AUNGIER; CAPEL;
OSBORNE)
1913 Selections from the Correspondence of Arthur Capel Earl of Essex
1675-1677. (Camden Third Series XXIV.) Ed. Clement Edwards
Pike. London.
Shakespeare, William, The Merry Wiues of Windsor (SHAKESP)
1955 Mr. William Shakespeares Comedies, Histories, & Tragedies. A fac-
simile edition prepared by Helge Kökeritz, with an introduction by
Charles Tyler Prouty. London: Geoffrey Cumberlege—Oxford Uni-
versity Press and New Haven: Yale University Press.
360 Bibliography

Shillingford, John
Letters and Papers of John Shillingford, Mayor of Exeter 1447-50.
(Camden Society, New Series II.) Ed. Stuart A. Moore. New York,
1965 [1871]:
Documents) (SHILLDOC)
Letter(s) (SHILLET)
The Siege of Jerusalem (SJERUS)
1969 The Siege of Jerusalem in Prose. (Memoires de la Sociöte Nöophilo-
logique de Helsinki XXXIV.) Ed. Auvo Kurvinen. Helsinki: Societe
N6ophiloIogique.
Smith, Henry, Sermons = A Preparative to Mariage; Of the Lords Supper, Of
Usurie (SMITH)
1975 A Preparative to Mariage; Of the Lords Supper; Of Usurie, 1591.
(The English Experience 762.) Amsterdam: Theatrum Orbis Terra-
rum—Norwood, N. J.: Walter J. Johnson. [Facsimile.]
State Trials
From: A Complete Collection of State-trials and Proceedings for
High-treason, and Other Crimes and Misdemeanours; from the
Reign of King Richard II. to the End of the Reign of King George I.
Ed. Francis Hargrave. (2nd edition.) London: Printed for J. Walthoe
Sen. Etc., 1730:
Vol. I:
The Trial of Sir Nicholas Throckmorton (THROCKM)
The Trial of Sir Walter Raleigh (RALEIGH)
Vol. IV:
The Trial of Titus Oates (OATES)
The Trial of the Lady Alice Lisle (LISLE)
The Statutes of the Realm (II) (STAT2)
1963 The Statutes of the Realm. Printed by Command of His Majesty King
[1816] George the Third in Pursuance of an Address of the House of Com-
mons of Great Britain. Vol. II. London: Dawsons of Pall Mall.
The Statutes of the Realm (III) (STAT3)
1963 The Statutes of the Realm. Printed by Command of His Majesty King
[1817] George the Third in Pursuance of an Address of the House of Com-
mons of Great Britain. Vol. III. London: Dawsons of Pall Mall.
The Statutes of the Realm (IV) (STAT4)
1963 The Statutes of the Realm. Printed by Command of His Majesty King
[1819] George the Third in Pursuance of an Address of the House of Com-
mons of Great Britain. Vol. IV. London: Dawsons of Pall Mall.
The Statutes of the Realm (VII) (STAT7)
1963 The Statutes of the Realm. Printed by Command of His Majesty King
[1820] George the Third in Pursuance of an Address of the House of Com-
mons of Great Britain. Vol. VII. London: Dawsons of Pall Mall.
 Bibliography 361

Stevenson (?), William, Gammer Gvrtons Nedle (STEVENSO)

1920 Gammer Gvrtons Nedle. [By Mr. S. Mr. of Art.] (The Percy Reprints
2.) Ed. H. F. B. Brett-Smith. Oxford: Basil Blackwell.
The Stonor Letters (BETSON; ESTONOR; TMULL)
1919 The Stonor Letters and Papers, 1290-1483. Vols. I—II. (Camden So-
ciety, Third Series XXIX-XXX.) Ed. Charles Lethbridge Kingsford.
London.
Stow, John, The Chronicles of England (STOW)
1580 The Chronicles of England from Brute unto This Present Yeare of
Christ. London: Printed by Ralphe Newberie.
Taylor, Jeremy, The Marriage Ring (JETAYLOR)
1907 The Marriage Ring (1673). Ed. Francis Coutts. London—New York:
John Lane.
Taylor, John, Pennyles Pilgrimage (JOTAYLOR)
1977 All the Workes of John Taylor the Water Poet, 1630. With an intro-
ductory note by V. E. Neuburg. London: The Scolar Press. [Facsimi-
le.]
Testaments and Wills (TEST)
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
Tillotson, John, Sermons = The Folly of Scoffing at Religion; Of the Tryall of the
Spirits (TILLOTS)
1976 Three Restoration Divines: Barrow, South, Tillotson: Selected Ser-
mons. Vol. II:ii. (Bibliotheque de la Facultö de Philosophie et Lettres
de l'Universite de Liege CCXIII.) Ed. Irfcne Simon. Paris: Soci6t6
d'Editions "Les Beiles Lettres".
Torkington, Richard, Γ Oldest Diarie of Englysshe Travell (TORKINGT)
1884 }* Oldest Diarie of Englysshe Travell: Being the Hitherto Unpublish-
ed Narrative of the Pilgrimage of Sir Richard Torkington to Jeru-
salem in 1517. (The Vellum-Parchment Shilling Series of Miscel-
laneous Literature VI.) Ed. W. J. Loftie. London: Field & Tuer, Y e
Leadenhalle Presse.
Treatise on Horses (HORSES)
See A Late Middle English Treatise on Horses.
Trevisa, John, Polychronicon Ranulphi Higden (TREVISA)
1876, Polychronicon Ranulphi Higden, Monachi Cestrensis. Vols. VI, VIII.
1882 [English translations of John Trevisa and of an unknown writer of
the fifteenth century.] (Rolls Series 41.) Ed. J. Rawson Lumby. Lon-
don.
Trial See The Arraignment of the Earles of Essex and Southampton; see
also State Trials.
362 Bibliography

Trinity Homilies (TRIN5; TRIN12; TRIN20; TRIN23)

1873 Old English Homilies of the Twelfth Century. Second Series. (Early
English Text Society, Original Series 53.) Ed. Richard Morris. Lon-
don.
Tunstall, Cuthbert, Original Letters, Illustrative of English History (TUNSTALL)
1846 Original Letters, Illustrative of English History; Including Numer-
ous Royal Letters. Ed. Henry Ellis. London: Richard Bentley.Turner,
William, A New Boke of the Natures and Properties of All Wines
(TURNER)
1941 Properties of All Wines (1568). A Book of Wines. Eds. Sanford V.
Larkey—Philip M. Wagner. New York: Scholars' Facsimiles & Re-
prints. [Facsimile.]
Turner, William, A New Boke of the Natures and Properties of All Wines
(TURNER)
1941 Properties of All Wines (1568). A Book of Wines. Eds. Sanford V.
Larkey—Philip M. Wagner. New York: Scholars' Facsimiles & Re-
prints. [Facsimile.]
Tyndale, William, Five Books of Moses (TYNDOLD)
1967 William Tyndale's Five Books of Moses Called the Pentateuch, Be-
ing a Verbatim Reprint of the Edition of M.CCCCC.XXX. Comps. J. I.
Mombert—F. F. Bruce. Fontwell—Sussex: Centaur Press Ltd.
Tyndale, William, The New Testament (TYNDNEW)
1938 The New Testament. Translated by William Tyndale, 1534. Ed. N.
Hardy Wallis, with an introduction by Isaac Foot. Cambridge: Cam-
bridge University Press.
Udall, Nicholas, Roister Doister (UDALL)
1934 Roister Doister. (The Malone Society Reprints.) London: John John-
[1935] son—Oxford University Press. [Facsimile.]
Usk, Thomas, Appeal(s) (USK)
1967 A Book of London English 1384-1425. Eds. Raymond Wilson Cham-
[1931] bers—Marjorie Daunt. Oxford: Clarendon Press.
Vanbrugh, John, The Relapse or Virtue in Danger (VANBR)
1927 The Complete Works of Sir John Vanbrugh. Vol. I. Eds. Bonamy
Dobrde—Geoffrey Webb. Bloomsbury: The Nonesuch Press.
Vespasian Homilies (VESPD32; VESPD33; VESPD42; VESPD48; VESPD3)
1971 Early English Homilies from the Twelfth Century MS. Vesp. D. XIV.
[1917] (Early English Text Society, Original Series 152.) Ed. Rubie D-N.
Warner. New York.
Vicary, Thomas, The Anatomie of the Bodie of Man (VICARY)
1888 The Anatomie of the Bodie of Man (1548). Part I. (Early English
Text Society, Extra Series 53.) Eds. Frederick J. Furnivall—Percy
Furnivall. London.
 Bibliography 363

Waerferth
See Gregory's Dialogues.
The Wakefield Pageants in the Towneley Cycle (TOWN)
1958 The Wakefield Pageants in the Towneley Cycle. (Old and Middle
English Texts.) Ed. Arthur Clare Cawley. Manchester: The Man-
chester University Press.
Walter Hilton's Eight Chapters
See Hilton's Eight Chapters.
Walton, Izaak, The Compleat Angler (WALTON)
1983 The Compleat Angler, 1653-1676. Ed. Jonquil Bevan. Oxford: Cla-
rendon Press.
Wycliffe, John, The Old Testament (Genesis and Numbers) (WYCOLD)
1850 The Holy Bible, Containing the Old and New Testaments, with the
Apocryphal Books, in the Earliest English Versions Made from the
Latin Vulgate by John Wycliffe and His Followers. Vol. I. Eds.
Josiah Forshall—Frederic Madden. Oxford: University Press.
Wycliffe, John and John Purvey, The New Testament (John) (WYCNEW)
1879 The New Testament in English According to the Version by John
Wycliffe About A. D. 1380 and Revised by John Purvey About A. D.
1388. Ed. Josiah Forshall—Frederic Madden. Oxford: Clarendon
Press.
Wycliffite Sermons (WESERM1; WESERM11; WESERM27; WESERM40; WSERM4;
WSERM16; WSERM23; WSERM28; WSERM32; WSERM36; WSERM43)
1983 English Wycliffite Sermons. Vol. I. Ed. Anne Hudson. Oxford: Cla-
rendon Press.
The York Plays (YORK)
1982 The York Plays. Ed. Richard Bedle. London: Edward Arnold.
Index of subjects

adjective 8, 10-11, 130, 272-273, - evaluative 272, 280, 290, 290-

275-276, 327 295, 310, 310-313, 315-316,
- complement 125, 127, 273, 284- 317, 319, 321, 323, 325, 327,
287, 290-293, 313-315, 326, 328, 329, 333-336, 339
328-329, 338-339 - focusing 272, 279, 290, 301-310,
- descriptive 273, 284-289, 305, 316-317, 320, 323, 325, 327,
313, 314, 317, 320, 322, 323, 333-336
327, 328, 329, 333-336 - generic time 273, 278, 279, 289,
- evaluative 273, 279, 290-295, 290, 311, 316-317, 323, 324,
313, 317, 320, 323, 326-329, 326, 333-336, 337
333-336 - habitual time 273, 279, 290, 316-
- focusing 273, 301-310, 316-317, 317, 323, 324, 326, 333-336, 337
320, 321, 322, 326, 328, 333-336 - integrated against peripheral 271,
- generic time 289, 290, 316-317, 280, 283, 311
324, 327, 333-336 - intensifying 272, 290, 301-310,
- habitual time 290, 316-317, 324, 316-317, 322, 325-328, 333-336
327, 333-336 - modal 272, 273, 279, 280, 283,
- intensifying 273, 290, 301-310, 289, 295-301, 317, 321-328,
316-317, 327, 333-336 333-336
- locative 273, 317 - open class 271
- modal 273, 279, 295-301, 314, - orientation 272, 281, 310-313,
317, 326, 327, 328, 333-336 337
- modifier 273, 284, 313-315, 338- -position 270, 271, 272, 282, 310-
339 313, 316, 325, 326, 329, 338
- participial 337 -predicate-oriented 279, 280, 310-
- state of mind 273, 295, 310-311, 313, 316, 319, 325
315, 317, 321, 322, 328, 329, - process adjunct 280, 284, 310,
333-336, 338 329
- temporal 273, 289-290 - space 278, 337
adverb 10, 271-272, 275-276, 327, - state of mind 280, 281, 295, 329,
337, see also conjunct, disjunct, 333-336, 337
subjunct - subject-oriented 271, 280-281,
- closed class 338 310-311, 315, 316, 317, 328,
- descriptive 272, 278, 279, 280, 329, 333, 338
284-289, 310, 319-323, 327, - temporal 272, 279, 289-290, 337
328, 333-336
366 Index of subjects

adverbial 125, 127, 133, see also - Brown University Standard Cor-
conjunct, disjunct, subjunct pus of Present-day American
- scope 272, 310-313 English 172
affect 8-10, 185-258, 280 - Century of Prose 20, 129, 173
anglicization 268, 336 - Early American English 11
apposition 7-8, 121-174 - Early English Correspondence 11
- exemplifying 128, 169 - Early English Medical Writing 11
- expository 7-8, 121-174 - Helsinki 3-5, 9, 11-12 and pas-
attitude 142, 160, 193, 201, 232, sim
240, 245, 269, 270, 278, 279, - London-Lund Corpus of Spoken
280, 290, 314, 324, 326, see also British English 172
involvement, point of view, - Older Scots 9, 11, 202, 270, 314
stance, subjectivity - Penn-Helsinki Parsed Corpus of
audience 142, 150, 154, 158, 168, Middle English 11
171, 188-191, 193, 195, 197, - Survey of English Usage Corpus
202, 234, 235, 253, 256, 268, of Written British English 172
270, 271, 300, 321, 328, 329, correlation 269, 272, 274, 289,
330-332, 336 290, 301, 305, 310-311, 315,
- non-professional 321 316-317, 324-328, 333-336,
- professional 321, 323 337, 338
author-centred 277
auxiliary, see verb: be/have, modal deixis 195, 198, 230
auxiliary, passive auxiliary, - proximity 199, 206, 210, 252,
tense auxiliary 253, 256," 257
dialect, see also Great Scandinavian
be 5-6, 17-80, 87-118 Belt
binomial construction - Old English 91
Boethius 294, 300, 301, 305, 312- - Middle English 92-100
313, 322, 324, 327 - social 2
dialectal variation 7, 166, 171
carnivalism 221, 256 dialogue 161, 188, 190, 196, 197,
chi-square test 32 213, 214, 216, 217, 226, 241,
complement 59-63, see also 242, 249, 283, 289, 318, 322, 329
object(-like) complement dimension 2, 8-11, 186, 193, 194,
conjunct 337 198, 203, 204, 232, 244, 251,
contraction 131, 199, 200, 213 252, 256, 269, 279, 284, 305,
conventions, see genre conventions 325-327, see also textual
corpus 3-5, 11-12 dimension
-ARCHER 20, 21 - descriptive versus attitudinal 269,
- Brooklyn-Geneva-Amsterdam- 279, 284, 305, 325-327
Helsinki Parsed Corpus of Old - multi-feature 269-270
English 12 direct speech 196, 221, 226, 235,
236, 252, 257
 Index of subjects 367

discourse 143, 155, 189, 196, 199, - informal language 31, 41, 49, 80,
234, see also genre conventions 197, 322, 330-332
- field 268 f-test 232, 258
- mode 268 functional load 18, 27, 28, 112-113
- tenor 268
disjunct 271, 281-283, 310, 311- gender 31, 50-51
312, 316, 325, 329 generic sentence 91
- sentence-initial 271, 282, 312 genre 1-2, 3-5, 8-11 and passim,
distinctive feature 2, 190, 194, 197, see also text type
230 genre conventions 186, 188-190,
durative, see verb 193, 195, 230, 256, 258
- discourse properties 185, 186,
egocentricity 195, see also focus: 189, 267-270, 301, 329
egocentric - features 203
evaluative, see adjective, adverb -politeness strategies 188, 197, 202
evidentials 201 - rhetorical 157-158
evidentiality 203, 212 - style 191
exclamation 200, 202, 204, 206, genre styles 4-5, 8-11, 188, 190,
210, 214, 223, 231, 232-237, 202-204, 256, 267, 269, 270,
242, 251, 256, 257 272, 328
existential sentence 103 Glim program 68, 80
expletive 10 grammar 122, 123, 267, 273, 326,
327, 339
fabliau 235, 236, 256 - prescriptive 18, 70
factor 1-2, 8-11, 121, 150, 170 grammatical model 2-3
- extralinguistic 1-2, 8-11, 17, 19, grammaticalization 18, 270, 271,
31-32, 38-51, 69-70, 113-114, 330
268, 269, 316-323, 327-328, 336 Great Scandinavian Belt 100
- linguistic 1-2, 17, 19, 31-32, 52-
67, 69-70, 336 have 5-6, 17-82
- interaction, see logistic regression heterogeneity 9, 203, 225, 268,
analysis 276-277, 288-289, 301, 318,
factor analysis 4-5, 186, 199, 201, 319, 337, 339
202-231, 245, 247, 251, 252, homogeneity 276, 277, 289, 294,
254, 255, 258, 273, 290, 326, 295, 300, 309, 310, 311, 314,
338, 339 316, 318, 319, 320, 321, 324,
focus 121, 130, 199, 249 327, 328
- egocentric 194, 199, 212, 247, homomorphy 272
251, 257, see also egocentricity
- interactive 186, 230, 257, see also indeterminate meaning 27, 36, 70
interaction informal language, see formality
formality 2, 8-11, 31, 51, 322, information processing, see dimen-
330-332, 339 sion, textual dimension
368 Index of subjects

informational density 270, 277, 139, 149, 171

313, 326, 329 - Old Norse 99
institutionalized 188, 190, 230 - Portuguese 17
interaction 193, 194, 197, 199, - Romance 17, 64
210, 213, 232, 246, 247, 251, - Scottish English 11, 268, 270,
252, 271, 279, 289, 300, 301, 314, 336, 338
311, 314, 318, 321, 322, 324- - Spanish 17
325, 328, 330-332, see also - Swedish 17
focus: interactive linear model analysis 68-70
inteijection 186, 204, 213, 214, Linguistic Atlas of Late Middle
216, 232, 233, 234-239, 253, English 6, 95, 100, 178
256, 257 linguistic feature 2, 4, 8, 10-11
interpersonal 196, 199, 244, see and passim, see also paralin-
also interaction guistic feature
intertextuality 9 literacy 235
involvement 192-195, 198, 199, loan-word 31, 64-65, 99-100, 171
202-204, 212, 230, 246, 256- logistic regression analysis 19, 32,
258, 269-270, 273, 301, see 67, 68-70, 80
also attitude, point of view,
stance, subjectivity manuscript 173, 192, 245, 336
irony 234, 258 matrices 186, 190, 198, 201, 230,
iterative, see verb 256
meaning, see indeterminate mean-
language, see also dialect, non- ing, meaning potential
standard language, spoken lan- meaning potential 1
guage, standard modality 186, 201, see also adjec-
- American English 31, 38-39 tive: modal; subjunctive; verb:
- Danish 17 conditional, modal auxiliary,
- Early Modern English 2-7, 17- mood, optative
81, 101-102, 121-174, 185-261, - boulomaic 202, 254, 256, 283,
267-339 284
- French 17, 148, 154, 156, 157, - deontic 21, 273, 283, 284, 295,
159, 171 298-300, 301, 313, 315, 322,
- German 17 325, 328, 334-336, 338
- Germanic 17 - epistemic 273, 283, 295, 296-
- Italian 17 300, 300, 312-315, 320, 322,
- Latin 17 324-325, 326, 328, 333-336
- Middle English 2-7, 17-80, 92- - perception 273, 283, 290, 295,
100, 121-174, 204, 234, 235, 296-300, 301, 311, 313, 315,
239, 256 317, 322, 324-327, 333-336
- Modem English 2, 5 and passim - space-building 186, 257
- Old English 3-6, 17, 18, 89-91, narration 196, 210, 215, 219, 220,
116-117, 121, 122, 128, 138, 221, 226, 234-237, 241, 248,
 Index of subjects 369

253, 254, 257, 267, 277, 282 printing 336

- imaginative 143, 197, 230-232, process adjunct, see adverb
238, 251, 318-321, 328 progressive, see verb
- indirect 196, 216, 251 pronoun 218, 257
- non-imaginative 143, 197, 230- - demonstrative 199, 252
232, 238, 251, 257, 258, 268, - indefinite 135
288, 289, 305, 316, 318-320, 328 -personal 135, 187, 195, 196, 198,
- narrator 267, 280, 281, 282, 324 199, 203, 206, 210, 212, 214-
non-standard language 197, see also 217, 219, 221, 222, 230, 236,
swearing 239-251, 252, 257, 338
negation 31, 58-59 - reflexive 123
- relative 136
object 125, 127, 132 proximity 206, 210, 252-254, 256,
object(-like) complement 17, 27, 257
29, 31, 133, 272, see also com-
plement qualitative analysis 8-11, 186, 201,
orality 31, 49-50 202, 226-230, 258
orderly heterogeneity 1 quantitative analysis 4-5, 8-11,
186, 226-230, 258, 313, 315
paralinguistic feature 233 question 199, 206, 210, 213, 214,
paralinguistic gesture 196 215, 223, 231
parody 241, 245, 256, 258
participant 194, 213, 242, 245, 258 rank, see social class
- relations 193, 194, 198, 268 readership 192, 258
- role 8, 11, 267, 279, 324, 329 reductive feature 199-200, 251
past perfect, see tense regularization 6
perception modality, see modality: re-phrasing 121—174
perception
personal affect, see affect script text 31, 49
philology 1-3, 6, 8-9, 11 "self'/"other" 194, 212, 244-252
point of view 167, 193, 194-195, semantic role 269, 271, 272, 277-
201, 212, 252, 254, 267-270, 310, 317, 325, 329, 330
273, 277, 278, 279, 282, 284, sentence, see adverbial, apposition,
290, 310, 311, 316-321, 323, complement, conjunct, exclama-
324, 326-329, see also attitude, tion, existential sentence, generic
involvement, stance, subjectivity sentence, negation, object, ob-
politeness 233 ject(-like) complement, ques-
- formula 188 tion, subject complement, sub-
- strategies 188, 197, 201 junctive, syntactic complexity,
pragmatic particle 168, 200, 206, THAT-deletion
210, 215, 216, 221, 232, 233, simplification 6
238, 253 social class 4, 240, 242, 243, 257,
present perfect, see tense 268, 271, 323
370 Index of subjects

social function 267-268, 269 216, 221, 231, 232-233, 251,

socio-historical linguistics 5, 19 253, 256
sociolinguistics 4-5, 8, 267 syntactic complexity 268, 269
speech-based text 2, 21, 31, 49,
143, 155, 167, 168, 322 tense 187, 199, 200, 219, 252, 253
spoken language 167, 196, 197, - past perfect 17-80
199, 200, 226, 232, 239, 252, - perfect infinitive 31, 52, 53-55,
271, 321, 322 57, 70
stance 11, 232, 270, 281, 325, 328, - perfective 17-80
see also attitude, involvement, - present perfect 17-80
point of view, subjectivity - present perfect versus past perfect
standard English 2, 336 52-53, 70, 78
standard deviation 224, 225, see - tense auxiliary 104, 106, 108, 112
also heterogeneity, homogeneity text category 144-147, 163-166,
standardization 87, 101, 171, 268 224-225, 268, 271, 311, 318-
statal, see verb: stative 321, 327-328, 330, 332, 339
statistical method, see chi-square text taxonomy 187
test, factor analysis, f-test, Glim text type 2, 3-5, 8-11 and passim,
program, linear model analysis, see also genre
logistic regression analysis, quan- text-type marker 2, 8-11, 185, 186,
titative analysis, standard devia- 190, 230, 231, 233, 255, 257-258
tion, t-test, variance analysis textual affinities 224-226, 258
style 9, 128, 148, 155, 159, 167, textual dimension, see also dimen-
171, 187, 188, 190, 191, 197, sion, involvement
199, 202, 203, 204, 251, 253, - detachment 269
256, 258, 288-289, 323, see -fragmentation 269, 329, 313-315,
also genre styles 329
subject complement 105, 133, 272 - integration 269, 290, 305, 313-
subjectivity 185, 186, 193, 258, see 315, 328-329
also affect, involvement, point of THAT-deletion 196, 199, 206, 210,
view, stance 219, 221, 231, 251, 338
subjunct 281, see also adverb: time reference 279, 289-290, 311,
focusing, intensifying 316-317, 323, 324, 326, 327,
- courtesy 329 333-336
- emphasizing 283 transitivity 273, 315-316, 325, 339,
- item subjunct 271, 280, 281, 337 see also verb: intransitive, transi-
subjunctive 90, 91, 93-95, 97, 99, tive
101 translation 137, 140, 150, 151, 152,
surge 186, 194, 232, 240, 241, 243, 153, 154, 156, 157, 159, 160,
252, 253, 256, 257, see also af- 171, 192, see also Boethius
fect, interjection t-test 186, 230, 231, 232, 245, 247,
swearing 197, 200, 206, 210, 214, 252, 254, 258, 321, 322
 Index of subjects 371

variance analysis 186, 231, 252, see - passive auxiliary 103, 106, 108,
also f-test 112
variation 1, 5-6, 7-8, 8-11 and - private 201, 210, 217, 218, 222,
passim 254, 269, 338
verb 127, 130, 132, see also modal- - process 17-80
ity, subjunctive, tense -progressive 104, 106, 111
- action(al) 17-80, 89, 272, 315 - pseudo-passive 79
- action versus process 31-32, 36- - relational 272
38, 41, 70, 76-78 - semantic category 272, 273
- be/have 17-80 - sensory perception 201, 206, 210
- come 29, 31-32, 45-47 - speech-art 190, 272, 329
- conditional 31-32, 56-58, 70 - stative 27-29, 32, 45, 53, 89, 282
- copula 102, 104-105, 107, 108, - stative versus mutative 31-32,
111 34-36
- durative 31, 56-58, 70, 91 - tense auxiliary 5-6, 17-80, 87-
- emotion(ality) 191, 193, 194, 118
197, 200, 201, 206, 210, 213, - transitive 17, 18, 27, 28, see also
216, 222, 234-237, 239-244, transitivity
253, 256, 257 - v o i c e 272, 315-316, 339
- 'expanded form' 104, 105 - volition 272
- go, see come
- imagination 201, 206, 210, 222 word-class, see adjective, adverb,
- mg-form 31, 54, 55-56, 109-111 conjunct, disjunct, inteijection,
- inert perception and cognition pronoun, subjunct, verb
272, 273, 283, 301, 339
- intransitive 17-80, see also transi-
tivity
- iterative 31, 56-58, 70, 91
- main veib 18, 27, 31-32, 53, 64-
67, 70
- mental activity 201, 210, 217,
218, 254
- mental state 201, 210, 232
- metalinguistic comments 272
- modal auxiliary 200, 201-202,
206, 210, 217, 218, 231, 254-
256, 257
- mood 193
- morphology 89-102
- mutative 17-80, 112
- observation 201, 206, 210, 217
- optative 32, 57-58