Focus Strategies in African Languages

≥
Trends in Linguistics
Studies and Monographs 191

Editors
Walter Bisang
(main editor for this volume)
Hans Henrich Hock
Werner Winter

Mouton de Gruyter
Berlin · New York
Focus Strategies
in African Languages
The Interaction of Focus and Grammar
in Niger-Congo and Afro-Asiatic

edited by
Enoch Oladé Aboh
Katharina Hartmann
Malte Zimmermann

Mouton de Gruyter
Berlin · New York
Mouton de Gruyter (formerly Mouton, The Hague)
is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

앪
앝 Printed on acid-free paper which falls within the guidelines
of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

Focus strategies in African languages : the interaction of focus and

grammar in Niger-Congo and Afro-Asiatic / edited by Enoch Oladé
Aboh, Katharina Hartmann, Malte Zimmermann.
p. cm. ⫺ (Trends in linguistics studies and monographs ; 191)
Includes bibliographical references and index.
ISBN 978-3-11-019593-4 (alk. paper)
1. Niger-Congo languages ⫺ Grammar. 2. Afroasiatic languages ⫺
Grammar. 3. Focus (Linguistics) I. Aboh, Enoch Oladé. II. Hart-
mann, Katharina. III. Zimmermann, Malte, 1970⫺
PL8026.N44F63 2007
4961.36⫺dc22
2007042927

ISBN 978-3-11-019593-4
ISSN 1861-4302

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

” Copyright 2007 by Walter de Gruyter GmbH & Co. KG, 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.
Cover design: Christopher Schneider, Berlin.
Printed in Germany.
Contents

Focus and grammar: The contribution of African languages

Enoch Oladé Aboh, Katharina Hartmann and Malte Zimmermann ........... 1

Part I. Focus and prosody

Nuclear stress in Eastern Benue-Kwa (Niger-Congo)

Victor Manfredi ......................................................................................... 15

Investigating prosodic focus marking in Northern Sotho

Sabine Zerbian .......................................................................................... 55

Part II. Information structure and word order

Preverbal objects and information structure in Benue-Congo

Tom Güldemann ........................................................................................ 83

Focus strategies and the incremental development of semantic

representations: Evidence from Bantu
Lutz Marten .............................................................................................. 113

Part III. Ex-situ and in-situ strategies of focus marking

Ex-situ focus in Kikuyu

Florian Schwarz ....................................................................................... 139

Focus in the Force-Fin system: Information structure in

Cushitic languages
Mara Frascarelli and Annarita Puglielli ................................................. 161

Coptic relative tenses: The Profile of a morpho-syntactic

flagging device
Chris H. Reintges ..................................................................................... 185
vi Contents

Part IV. The inventory of focus marking devices

Identificational operation as a focus strategy in Byali

Brigitte Reineke ........................................................................................ 223

Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee

Katharina Hartmann and Malte Zimmermann ........................................ 241

Part V. Focus and related constructions

Narrative focus strategies in Gur and Kwa

Anne Schwarz and Ines Fiedler ................................................................ 267

Focused versus non-focused wh-phrases

Enoch Oladé Aboh ................................................................................... 287

List of contributors ................................................................................... 315

Subject index ............................................................................................ 317
Language index ........................................................................................ 323
Focus and grammar: The contribution of African
languages

Enoch Oladé Aboh, Katharina Hartmann, and Malte

Zimmermann
Starting with Chomsky (1971) and Jackendoff (1972), the formal study of
information structure and grammar has received an ever increasing interest,
and led to the formulation of various proposals centering on the expression
of focus in natural languages and its function in the architecture of gram-
mar (e.g. Dik 1981, Vallduvi 1990, Cinque 1993, Lambrecht 1994,
Reinhart 1997, Drubig 2007). This book aims at contributing to the ongoing
discussion of focus by investigating a range of African languages hitherto
under-represented in the literature. The volume consists of a selection of
articles that developed from presentations at a Workshop on “Topic and
Focus: Information Structure and Grammar in African Languages” held at
the University of Amsterdam in December 2004.
Notwithstanding the wide range of approaches to focus found in the lit-
erature, the contributions to the present volume all agree on the following
very general notion of focus as a pragmatic category that interacts with
grammar (Jackendoff 1972, Dik 1981, Lambrecht 1994):

(1) Focus refers to that part of the clause that provides the most rele-
vant or most salient information in a given discourse situation.

Typically, an expression will be most relevant or most salient if it is either

new or contrasted with another element in the preceding or subsequent
discourse. The non-focused part of a clause is often referred to as the back-
ground:

(2) Background refers to that part of the clause that contains the pre-
supposed and/ or given information, where givenness implies
having been mentioned in the preceding discourse.

In this volume, the background is also referred to as the out-of-focus part

by some authors (e.g. Reineke, A. Schwarz & Fiedler).
2 Enoch Aboh, Katharina Hartmann and Malte Zimmermann

The information-structural category of focus must be kept apart from the

notion of focus marking, which refers to the overt realization of focus by
special grammatical means, which is subject to cross-linguistic variation:
Languages can mark focus syntactically, or prosodically, or morphologi-
cally, or they can use combinations of these grammatical means. At the
same time, non-contrastive focus is not always marked, as shown by some
of the languages discussed in the present volume.
The articles in this volume look at focus strategies in a variety of Afri-
can languages (Niger-Congo, Afro-Asiatic) from several theoretical and
methodological perspectives. Their common aim is to deepen our under-
standing of how the information-structural category of focus is represented
and marked in the languages of the world. This inquiry into the focus sys-
tems of African languages will have repercussions on existing theories of
focus because it reveals new focus strategies as well as fine-tuned focus
distinctions that are not discussed in the literature, which is almost exclu-
sively based on intonation languages. Departing from the purely descriptive
studies that we are familiar with when it comes to African languages, this
book is – to our knowledge – unique in its effort to combine careful empiri-
cal study and theoretical analysis of the prosody, morphosyntax, and se-
mantics of focus. Though every single paper in this volume addresses more
than one issue in connection with the interaction of focus and grammar, the
whole collection revolves about the following theoretical and empirical
topics: (i) Focus and prosody; (ii) information structure and word order;
(iii) ex-situ and in-situ strategies of focus marking; (iv) the inventory of
focus marking devices; (v) focus and related constructions.

1. Focus and prosody

Studies on intonation languages (e.g. Germanic, Romance) show that these

use stress assignment (sometimes combined with syntactic transformations)
for marking focus (e.g. Selkirk 1984, 1995). The underlying stress is typi-
cally expressed in form of a nuclear pitch accent, as shown for the follow-
ing English examples (the focused constituent is underlined, the accented
syllable is printed in capitals):

(3) a. Where did Peter buy the cassava?

He bought them at the MARket.
 Focus and grammar: The contribution of African languages 3

b. What did Peter buy at the market?

He bought casSAva at the market.
c. Who bought the cassava at the market?
PEter bought them.

In contrast, some African tone languages (e.g. from the Kwa, Gur, Bantu,
and Chadic families) make use of syntactic transformations (e.g. fronting
rules) for the same purpose. A plausible assumption is that the presence of
tone in these languages somehow diminishes the prominence of intonation
(or stress assignment). The following examples from Gungbe would sup-
port this view a priori:

(4) a. Ét( w(a à n2a sà ?

what FOC 2sg HAB sell
‘What do you habitually sell?’
b. Hwèví w(a ùn n2a sà.
fish FOC 1sg HAB sell
‘I sell FISH.’

Things are not so clear-cut however. Kanerva (1990), for example, shows
that focus marking in Nkhotakota Chichewa (Bantu) has an effect on the
prosodic phrasing. In Chichewa, the right-edge of a prosodic phrase is indi-
cated by penultimate lengthening and tone lowering on the phrase-final
vowel, e.g. nyumbá > nyúumba in (5bc). The examples in (5a–c) show that
the expression of focus affects the prosodic phrasing of the Chichewa
clause in that a prosodic phrase boundary must be inserted after the focused
constituent, i.e. after the VP in (5a), after the object NP nyúumba in (5b),
and after the verb anaméenya in (5c), respectively:

(5) a. What did he do? (VP focus)

(anaményá nyumbá ndí mwáála)
he.hit house with rock
‘He hit the house with a rock.’
b. What did he hit with the rock? (object NP focus)
(anaményá nyuúmba) (ndí mwáála)
c. What did he do to the house with the rock? (V focus)
(anaméenya) (nyuúmba) (ndí mwáála)
4 Enoch Aboh, Katharina Hartmann and Malte Zimmermann

The examples in (5) show that certain tone languages can resort to prosodic
devices for marking focus (see also Yip 2002), giving rise to the following
questions: How do prosodic focus marking and lexical tone interact? To
what extent is prosody inherent to the expression of focus? And do speak-
ers always employ prosodic cues that are associated with the focused ex-
pression?
The articles by Manfredi and Zerbian in this volume address these is-
sues for a number of African tone languages and provide a closer inspec-
tion of the role of prosody in information packaging in these languages.
The two articles take different starting points and come to radically differ-
ent conclusions. Based on a careful experimental study of the prosodic
properties of Northern Sotho, Zerbian shows that prosody plays no role in
the marking of postverbal focus constituents in this language. Manfredi, in
contrast, takes a universalist perspective as his starting point. According to
his analysis, African tone languages do not differ significantly from intona-
tion languages when it comes to focus marking. Universally, focus is
marked by means of abstract stress (prominence), which is spelt out as a
pitch accent in intonation languages and as a high tone in Kimatuumbi and
Luhaya. It remains to be seen whether Manfredi’s analysis can be extended
to Northern Sotho, and if so, how abstract stress is prosodically realized in
this language.

2. Information structure and word order

In many languages, information structure affects word order. Such interac-

tion has been described for so-called discourse-configurational languages
(É. Kiss 1995). In German, for instance, constituents expressing old infor-
mation tend to precede the focus constituent in the middle field, sometimes
leading to reordering of the basic word order (scrambling). For instance, the
unmarked order of indirect object (IO) and direct object (DO) in the neutral
clause in (6a) is IO > DO. However, when the indirect object is focused, it
must follow the backgrounded direct object, as shown in (6b). Notice that
this word order variation correlates with the definiteness/ indefiniteness
distinction:

(6) a. Peter hat einem Mann ein Buch gegeben.

Peter has a man a book given
‘Peter gave a book to a man.’
 Focus and grammar: The contribution of African languages 5

b. Q: To whom did Peter give the book?

A1: Peter hat das Buch einem Mann gegeben.
Peter has the book a man given
A2:# Peter hat einem Mann das Buch gegeben
Peter has a man the book given

Similar effects of information structure on word order are found, for in-
stance, in Hungarian and in the Slavic languages. Even though some Bantu
languages have been observed to exhibit discourse-configurationality as
well (e.g. Bresnan & Mchombo 1987), the correlation between information
structure and word order variation in African languages has not been sub-
ject to in-depth study.
The articles by Marten and Güldemann in this volume are the first sys-
tematic studies in this domain: Marten gives an account of the various word
orders in Bantu as the result of an interaction between syntax and pragmat-
ics. Güldemann shows that the shift from SVO to SOV in a number of
Bantu and Kwa languages is triggered by the information-structural status
of the preposed object, comparable to scrambling processes in Germanic.

3. Ex-situ and in-situ strategies

Certain African language families, such as Kwa, Bantu, and Chadic, mark
focus syntactically by means of focus fronting (cf. e.g. Ameka 1992, New-
man 2000). This ex-situ strategy was illustrated in (4) for Gungbe, but it
also exists in Hausa, as illustrated by the question-answer pair in (7):

(7) a. Mè sukà kaamàà ?

what 3pl.rel.perf catch
‘What did they catch?’
b. Dawaakii nèè sukà kaamàà.
horses FOC 3pl.rel.perf catch
‘They caught HORses.’

Depending on the context, though, the same question may also be answered
by using an in-situ strategy in many of these languages: In the Hausa ex-
ample (8), the focus constituent appears in-situ in its base position.
6 Enoch Aboh, Katharina Hartmann and Malte Zimmermann

(8) Sùn kaamà dawaakii.

3pl.perf catch horses
‘They caught HORses.’

In other words, these languages allow for two focus strategies (at least), one
of which involves overt displacement and optional morphological marking.
The existence of two structural focus strategies gives rise to two additional
questions: Firstly, are there semantic/pragmatic differences between the
two focus strategies? And if so, what are they? Secondly, do the ex-situ
versus in-situ strategies comply with the new information versus contras-
tive focus dichotomy as proposed for certain intonation languages (e.g. É.
Kiss 1998, Drubig 2001)?
While the literature on certain African languages often deals with the
focus strategy involving displacement and/or morphological marking, it is
still unclear whether or how the in-situ strategy is marked, and what seman-
tic effects arise with in-situ foci. Several articles in the volume address
these questions to some extent: Concerning the formal marking of in-situ
focus, Reintges shows that in-situ focus in Coptic is marked by means of
special relative morphology, whereas Zerbian’s production and perception
studies suggest that some languages (e.g. Northern Sotho) do not mark
focus at all, neither in-situ nor ex-situ. With respect to interpretation, Aboh
proposes that there is no semantic difference between the two focus posi-
tions. In the same vein, Hartmann & Zimmermann argue that the semantic
notion of exhaustivity should not be linked to a particular syntactic position
in Hausa, but results from the presence of an exhaustivity marker.

4. Inventory of focus marking devices

The African languages under discussion exhibit a rich variety of focus

strategies, sometimes even within one and the same language, and conse-
quently the present book provides an extensive overview of possible focus
marking devices. Apart from syntactic dislocation, which is attested in most
languages investigated here, focus can also be marked morphologically by
means of special focus markers or by a change in the aspectual system: The
morphological strategy is found, for instance, in Byali (Reineke), Egyptian
Coptic (Reintges), Somali (Frascarelli & Puglielli), Ewe, Akan, Lelemi,
Buli, Dagbani (A. Schwarz & Fiedler), and Kikuyu (F. Schwarz).
 Focus and grammar: The contribution of African languages 7

The observed variety of focus marking devices should not obscure the
fact that many of the African languages discussed here exhibit a curious
absence of grammatical focus marking, especially when compared to into-
nation languages, where sentences must have a pitch accent. In most of
these languages, what has been referred to as new information focus (e.g. in
answers to wh-questions) tends to be unmarked. The absence of such
grammatical focus marking may constitute a serious challenge to the view
that focus must be marked somehow (see e.g. Gundel 1999). At the same
time, it might also turn out that tone and intonation interact in an intricate
way to mask the marking of the in-situ focused constituent. To our mind,
the articles in this volume contribute to a better understanding of such focus
related phenomena and will pave the way to some refinements of existing
focus theories.

5. Focus and related constructions

Another issue addressed in this book is the way in which clauses containing
an overtly marked focus expression, often informally referred to as focus
constructions in the volume, are related to other discourse-related clause
types, such as wh-questions, narratives, relative clauses, and clefts. A phe-
nomenon that is often mentioned, but still not fully understood, is the for-
mal identity between focus constructions and these other clause types. A
case in point is the identical shape of the aspectual marker in focus con-
structions, wh-questions, and relative clauses (e.g., Hartmann &
Zimmermann for Hausa, Reintges for Coptic). In addition, A. Schwarz &
Fiedler observe a much less studied structural correlation between ex-situ
focus and narrative structures in Kwa and Gur. Aboh discusses question-
answer pairs in Gungbe, showing that focused wh-questions and non-
focused wh-questions require different formal licensing as well as different
information structuring of the answer. Finally, Frascarelli & Puglielli (So-
mali), and Reineke (Byali) point to well-known but still controversial struc-
tural identities between focus constructions and clefts, wh-questions, and
equative sentences, respectively.
As mentioned above, many of the articles in this volume address more
than one of the theoretical and empirical issues outlined above. Nonethe-
less, it is possible to subsume each article under one of the five issues dis-
cussed above according to its main focus of interest. The volume thus con-
sists of five independent but interrelated parts. Part I, Focus and prosody,
8 Enoch Aboh, Katharina Hartmann and Malte Zimmermann

takes up the issue of whether and how focus is marked prosodically in Afri-
can languages and contains the articles by Manfredi and Zerbian. Part II,
Information structure and word order, deals with the influence of informa-
tion structure on word order and contains the articles by Güldemann and
Marten. Part III, Ex-situ and in-situ strategies of focus marking, examines
syntactic focus marking strategies in more detail and contains the articles
by F. Schwarz, Frascarelli & Puglielli, and Reintges. Part IV, The inventory
of focus marking devices, deals with the question of which functional ele-
ments should be considered bona fide focus markers, and which ones
should not. It contains the articles by Reineke and Hartmann &
Zimmermann. Part V, Focus and related constructions, investigates the
relationship of sentences containing a syntactically marked focus with other
discourse-related clause types, such as wh-questions, relative clauses and
clefts. It contains the articles by A. Schwarz & Fiedler and Aboh. We hope
that this ordering will allow prospective readers an easier access to those
articles reflecting their particular research interests. For ease of exposition,
the maps at the end of the introduction provide an overview of the lan-
guages discussed in this volume, and where in Africa they are spoken. Map
1 gives a general overview. Map 2 gives a more detailed overview of the
languages spoken in Nigeria and neighboring countries.

Acknowledgments

Work on the present volume would not have been possible without the
generous financial support from the following institutions: The Amsterdam
Center for Language and Communication (ACLC) and the Nederlandse
Organisatie voor Wetenschappelijk Onderzeok (NWO), Vidi-grant 276–75–
003 (Enoch Oladé Aboh); and the Deutsche Forschungsgemeinschaft
(DFG), and in particular the DFG-sponsored collaborative research centre
SFB 632 “Information Structure”, located at Potsdam University and
Humboldt University of Berlin (Katharina Hartmann, Malte Zimmermann).
We would further like to thank Felix Ameka, Jeff Good, Nancy Chongo
Kula, Roland Pfau, Harold Torrence, and an anonymous reviewer for help-
ing us in the reviewing process. Special thanks are due to Lars Hermann
and Anne Schwarz for their editing work in preparing the final manuscript,
to Maria Höger for helping with the indexes, and to Lars Marstaller for his
proof-reading.
 Focus and grammar: The contribution of African languages 9

Map 1. Africa complete

10 Enoch Aboh, Katharina Hartmann and Malte Zimmermann

Map 2. Nigeria and neighboring countries

 Focus and grammar: The contribution of African languages 11

References

Ameka, Felix
1992 Focus Constructions in Ewe and Akan: A Comparative Perspective.
MIT Working Papers in Linguistics 17, 1–25.
Bresnan, Joan, and Sam Mchombo
1987 Topic, Pronoun, and Agreement in Chichewa. Language 4, 741–782.
Chomsky, Noam
1971 Deep structure, surface structure and semantic interpretation. In
Semantics: an interdisciplinary reader in linguistics, philosophy and
psychology, Danny D. Steinberg and Leon A. Jacobovits (eds.), 183–
216. Cambridge: Cambridge University Press.
Cinque, Guglielmo
1993 A null theory of phrase and compound stress. Linguistic Inquiry 24:
239–298.
Dik, Simon et al.
1981 On the typology of focus phenomena. In Perspectives on functional
grammar, Teun Hoekstra (ed.), 41–74. Dordrecht: Foris.
Drubig, Bernhard
2001 Focus Constructions. In Language Typology and Language Univer-
sals. An International Handbook, Martin Haspelmath, Ekkehard
König, Wulf Oesterreicher, and Wolfgang Raible (eds.), 1079–1104.
Berlin: de Gruyter.
2007 Phases and the typology of focus constructions. In On Information
Structure, Meaning and Form, Kerstin Schwabe and Susanne
Winkler (eds.), 33–66. Amsterdam: John Benjamins.
É. Kiss, Katalin
1995 Discourse Configurational Languages. Oxford: Oxford University
Press.
1998 Identificational Focus Versus Information Focus. Language 74: 245–
273.
Gundel, Jeanette K.
1999 On Different Kinds of Focus. In Focus: Linguistic, Cognitive, and
Computational Perspectives, Peter Bosch and Rob van der Sandt
(eds.), 293–305. Cambridge: Cambridge University Press.
Jackendoff, Ray
1972 Semantic Interpretation in Generative Grammar. Cambridge, MA:
MIT Press.
Kanerva, Jonni
1990 Focusing on Phonological Phrases in Chichewa. In The Phonology-
Syntax Connection, Sharon Inkelas and Draga Zec (eds.), 145–161.
Chicago: CSLI.
12 Enoch Aboh, Katharina Hartmann and Malte Zimmermann

Lambrecht, Knud
1994 Information Structure and Sentence form. Cambridge: Cambridge
University Press.
Newman, Paul
2000 The Hausa Language. An Encyclopedic Reference Grammar. New
Haven & London: Yale University Press.
Reinhart, Tanya
1997 Interface Economy: Focus and Markedness. Studia Grammatica 40:
146–169.
Selkirk, Elisabeth
1984 Phonology and Syntax. The Relation between Sound and Structure.
Cambridge, MA: MIT Press.
1995 Sentence Prosody: Intonation, Stress, and Phrasing. In Handbook of
Phonological Theory, John Goldsmith (ed.), 550–569. Oxford:
Blackwell.
Vallduví, Enric
1990 The Informational Component. Ph.D. dissertation, University of
Pennsylvania, Philadelphia.
Yip, Moira
2002 Tone. Cambridge: Cambridge University Press.
Part I

Focus and prosody

Nuclear stress in eastern Benue-Kwa
(Niger-Congo)

Victor Manfredi

Abstract

In ‘Bantu’ (eastern Benue-Kwa, [BK]), standard analyses of information structure

assume autosegmental tonemes and templatic verb morphology (Odden 1984,
Hyman 1999). Dispensing with both devices, inflectional tones reduce to phrasal
accents (Bamba 1991, Idsardi & Purnell 1997), a VP constituent emerges from aux
and clitic shells (Keach 1986, Myers 1998, Kinyalolo 2003a,b), and focus prosody
appears Romance-like, with topics dislocated away from nuclear stress (Vallduví
1990, Zubizarreta 1998). The revised view enhances learnability, narrowing the set
of possible languages while capturing language-internal generalizations.

1. Caveat stressor

In generative studies of Germanic and Romance, the term nuclear stress

refers to default phrasal accentuation which codes information focus
(Chomsky 1971, 199f., Vallduví 1990, Zubizarreta 1998). The concept is
not universally applied, thanks to a stack of gratuitous assumptions. (i)
Prosodic phonology (Pierrehumbert 1980, Selkirk 1995) views accent as
“non-phonemic prominence structure having to do with various pragmatic
things people do” and thus as an unlikely source for “symbolic representa-
tions of phonological units that determine pitch differences” (D. Odden,
p.c.). In English, however, nuclear stress is phonologically conditioned
(Chomsky & Halle 1968, 17, Arregi 2002) and also shows “genuine ambi-
guity” between broad and narrow focus interpretations (Ladd 1996, 201, cf.
Gussenhoven 1983b), therefore, accent can’t be excluded from derivations
and a null theory of accent is the null hypothesis (Cinque 1993). (ii) Carto-
graphic syntax (Brody 1990, Rizzi 1997) treats ex-situ focus as attraction to
a formal/“criterial” feature with no intrinsic PF correlate. In Hungarian,
however, subordinating main stress (the “Eradication Rule”) tracks focus
movement (Kornai & Kálmán 1988, Szendröi 2003). (iii) Relativistic de-
16 Victor Manfredi

scriptions of prosody (§1.1) and morphology (§1.2) underanalyze focus

phenomena in Kimatuumbi (§2) and other ‘Bantu’ languages (§3). 1

1.1. Autosegmental luggage rack

Tonemes – alphabetic representations of paradigmatic pitch contrasts (Pike

1948) – were ported into generative grammar as unscreened conceptual
baggage three decades ago (Williams 1971, Leben 1973, Goldsmith 1976).
Less than robustly detected in instrumental and perceptual experiments
(Rouget 1965, LaVelle 1974, Hombert 1976, Abramson 1979, Gandour
1983, Connell 2000, Dilley 2005), tonemes have proven inseparable from
metrical effects like interpolation, skewed distribution, and juncture
(McCawley 1978, Abramson 1978, Akinlabip 1985, Sietsema 1989, Haragu-
chi 1991, Bamba 1991, Lapnipran 1992, Liberman et al. 1993, Liberman
1995, Idsardi & Purnell 1997). Despite such clues, tonemes are immune to
disproof if accents are considered as just another tier in an “autosegmental-
metrical” array (Leben 1976, Halle & Vergnaud 1982, Goldsmith 1982,
Beckman & Pierrehumbert 1986, Gussenhoven 2004).
The main justification for tonemic autonomy is that phonological con-
trasts obey phonetic biuniqueness, as in Hyman’s slogan “tone = pitch fea-
tures present lexically” (1989, 116) or in Odden’s sorting PF properties into
“stress” versus “tone” while rejecting “shared formal machinery” despite
“certain apparent similarities” (1999, 212). But biuniqueness loses gener-
alizations in taxonomic discovery procedures (Halle 1959, 23), translation
wordlists don’t reliably find minimal pairs, and language-internal puzzles
result. (i) Emic tone demands ad hoc junctures. In Kimatuumbi, for exam-
ple, Odden observes “tonal upsweep” – a gradual pitch rise from “L” to
“H” across multi-syllables and a phrase boundary – and writes it as a string
of H tonemes separated by [¡], the mark of “phonetic upstepping” (Pike &
Wistrand 1974, Meir et al. 1975). But only “the final H in such a sequence
is at the pitch level typical for an H-tone” (Odden 1996, 6), so to specify
each middle syllable as H entails an absurd rule of leftward downdrift. 2 (ii)
Many BK listemes divide into ‘tone classes’ – lexically latent pitch con-
trasts surfacing only in phrasal contexts – e.g. Gikuyu and Umbundu
“nouns” (Benson 1964, xxi–xxviii, Clements & Ford 1979, Clements 1984,
Schadeberg 1986) and Izgbo “verbs” (Swift et al. 1962, EpmepnanjoB 1981,
Nwapchukwu 1983). Such effects neatly diagnose prelinked phrasal accent
(Pulleyblank 1982, Déchaine 1993, Purnell 1998) but are more often han-
dled as postlexical “tonal morphemes” (Sharman & Meeussen 1955,
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 17

Welmers 1959) or “floating/boundary/register” tones (Hyman 1974, 1985a,

Clements 1981, Snider 1990). (iii) Tonemes obscure evidence of a phrase
boundary after the BK “noun prefix” (Welmers 1973b) – an obligatory
closed-class item displaying restricted tone and licensing a definite inter-
pretation even in the absence of an overt demonstrative (Stahlke 1971,
Manfredi 1993, 2003, 2004a, Déchaine & Manfredi 1998, Ajipbopyez 2005).
Tonemes also overstate the prosodic diversity of natural language: their
frequency is broadly complementary to suffixation and/or to the option of
branching rimes (Manfredi 2003). Therefore if “a major goal for linguistic
theory is to define the notion ‘possible language’” (Fromkin 1978, 1), it’s
reasonable “to reinterpret autosegmental phonology as a special case of
metrical phonology” (Leben 1982, 189). Accordingly, Duanmu eliminates
underlying monomoraic tone contours in Chinese (1a), wiping out the ini-
tial plausibility argument for autosegmental notation (Goldsmith 1976, 21–
30), and arriving at the universal in (1b). In the same vein, consider West
Germanic (2) and ‘Bantu’ (3).

Duanmu (1994, 567; 2005)

(1) a. “… stressless syllables either do not carry underlying tones… or
will lose their underlying tones”
b. “… while languages can differ in word stress, all languages have
the same rule for compound and phrasal stress”

Wagner (2005, 34, 273, 275) 3

(2) a. “If A and B are sisters and A is the functor and B its argument,
… B is [accentually] more prominent than A unless A already
contains an argument …” (cf. Cinque 1993, 244).
b. “Only associative domains are built in a single cycle. …
Each cycle consists of a right-branching structure.”
c. [An informationally given constituent] “becomes a functor”
(cf. Steedman 2000).

Hyman (1999, 153)

(3) a. “[–focus] o tonal integration = reduction”
b. “[+focus] o tonal finality = end demarcation”

Despite the morphosyntactic distance between these languages, patterns (1)

– (3) are empirically nondistinct. This convergence could be due to extrin-
18 Victor Manfredi

sic factors (4a), or it could show that the faculté de langage handles stress
and tone uniformly (4b).

(4) a. prosodic relativity: focus cues are incommensurable

grammaticalized “strategies” (Vallduví & Engdahl 1996),
interpreted beyond the semantic interface (Reinhart 1997) under
a generic functional label of “prominence” (Truckenbrodt 1995).
b. prosodic unity: all natural languages compute semantically-
relevant prosody as phrasal accent (‘nuclear stress’).

(4a) is unfalsifiable (cf. Kaye 1988), both because it’s evaluated in a global
output procedure (economy/OT) and because it admits any imaginable fo-
cus cue (pitch, duration, word order, morphological Gestalt…). 4 To limit
indeterminacy, it’s sometimes suggested that a focus-sensitive feature can’t
also support lexical contrasts – e.g. “… the phonetic cue to sentence accent
is duration in Bantu languages” since “Bantu languages cannot use the
prevalent component of pitch changes because of their tonal nature” (Zer-
bian 2005, 15) – but such an inference is false in general. In East Asian
languages, lexical tone doesn’t block “parallel encoding” of focus informa-
tion as F0 pitch (Xu & Xu 2005, cf. Potisuk et al 1996, Xu 2004), and in
‘nontonal’ Chimwiini-Kiswahili vowel duration is lexically contrastive as
well as governed by focus-sensitive phrasing (Kisseberth 2002). Appeals to
lexical tone are circular anyway: it’s “difficult to draw a dividing line be-
tween languages with contrastive tone on (almost) all syllables and lan-
guages with tone contrasts in more restricted locations in the word. Stan-
dard Chinese and Swedish are … both tone languages by this definition”
(Gussenhoven 2004, 47). Eastern BK is equivocal in these terms, because
lexical pitch is underspecified in all conceivable ways: paradigmatic
(Meeussen 1963, Stevick 1969), syntagmatic (Voorhoeve 1973) and rela-
tive to syntactic category (Odden 1988, Kimenyi 2002). The literature does
report one plausible instance where duration necessarily substitutes for
pitch as a focus cue: in second-occurrence focus – however, this effect de-
pends on configurational anaphora not lexical contrast sets and implicates
metrical structure (Rooth 1996, 219) consistent with (2c).
(4b), on the other hand, could be wrong but finds support beyond the
studies cited under (1) – (3). In an elegant analysis of Spanish and Catalan,
for example, verbs and nouns follow different stress rules reflecting the
difference between high and low pronunciation of the root in its extended
phrasal shell (Arregi & Oltra-Massuet 2005). To be sure, it’s not obvious
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 19

from inspection of BK citation forms that every (OCP) H tone recovers an

accent, thus implying the presence of a local weak (non-H) position in
keeping with metrics’ “relational” property (Liberman 1975, 51). It was not
even obvious in English, to English-speaking linguists, just 60 years ago:
Anticipating ToBI (Pierrehumbert 1980), Pike misses the phrasal distribu-
tion of stress (Bolinger 1951, Chomsky et al. 1956, 72) and sets up four
“relative but significant levels (pitch phonemes)” (1945, 25), stringing them
into intonational lines with a redundant word-level stress [°] for each “2”
tone. 5

Pike (1945, 27–30, cf. Trager & Smith 1949)

(5) a. He wanted to do it. [°2–4]
b. I want to go home. [3–°24]
c. The boy in the house is eating peanuts rapidly.
[3–°23–3–°23–3–°23–°23–°24]

In sum, the tonemic valise packs a load of warrantless assumptions plus a

bundle of pitch features in search of analysis. In eastern BK, this search
meets a second type of impediment.

1.2. The diacritical imperative

Some Bantuists identify a focus slot right-adjacent to the finite verb

(Watters 1979, Ndayiragije 1998, v.d. Wal 2005), but this can’t be the
verb’s syntactic complement – as in Hyman’s (1985b) analysis of Aghem –
if one recognizes a polysynthetic “verbal word” (Nurse 2003, 90, cf.
Meeussen 1967, Goldsmith 1985, Hyman 2003), a templatic string of ar-
gument-type clitics, an aux, the lexical root plus derivational “extensions”
(minimally, a default vowel), cf. (6a). Canonized in conjunctive orthogra-
phy (Guthrie 1948), the template is reborn in generative theory (Baker
1988, 1996, Odden 1996, 71, 228f., Carstens 2001) and entails computa-
tional explosion: Odden (1981, 17) reportedly estimates “that some
16,000,000,000,000 distinct forms can be built around a single verb radical,
not counting the differences induced by distinctive tonal features of various
morphemes” (Sietsema 1989, 90). 6 Luckily, a range of evidence shows that
the left edge of the verb root is separated by ordinary phrase boundaries
from subject “agreement” (CLs ), which is “merged as a DP, an independ-
ent syntactic object” (Kinyalolo 2003b, cf. Keach 1986, Bresnan &
Mchombo 1987, Myers 1998, Russell 1999, Kinyalolo 2003a), cf. (6b). 7
20 Victor Manfredi

(6) a. [X° CLs–(aux)–(CLo)–verb root–extensions ] [DP+foc __ ]

b. [TP [DP CLs] [T' (aux) [KP [DP (CLo) ] [VP [V° verb root–
extensions] [DP+foc __ ] ] ] ] ]

The choice between (6a) and (6b) has a clear consequence for focus: (6b)
allows a surface VP constituent, whereas under (6a) a VP must be simu-
lated by stratally-ordered lexical phonology (Pulleyblank 1983, Odden
1996, 228ff.) plus a focus diacritic. This diacritic is actually embraced:
“Bantu languages in particular are known for their ‘focus prominence’ … as
when a tense is marked differently according to whether the verb is included
in the focus or not. … Whether one is a syntactician or semanticist wishing
to study focus or one is a phonologist wishing to study tone, one must con-
sider all aspects of the grammatical system of a Bantu language. … [T]o not
do so would be to risk drawing the tempting – but wrong – conclusion that
there is a direct link between semantic focus and pitch in these languages.”
(Hyman 1999, 151, 174, his italics)
Similarly, Odden rescues (6a) with a semantic parameter, making Ki-
matuumbi a language where “focal-sensitivity is a general property of an
entire grammar” banning “two items focused in a clause” and deploying
“morphological processes whose sole purpose is marking focus”; the exis-
tence of examples like “Who likes only meat? or Tom likes Sally” are said
to show that such a ban “is not found in languages like English” (Odden
1984, 277, his italics, cf. 1996, 71).
On second thought, however, these arguments for (6a) fail. Hyman’s
rhetoric above is misdirected and overblown. Misdirected: It’s not the “syn-
tactician or semanticist” but rather the prosodic phonologist (Pierrehumbert
1980, Gussenhoven 1983a) or interface economist (Reinhart 1997) who’s
“tempted” to draw “a direct link between semantic focus and pitch”. Over-
blown: No reason is given why the road of indirectness must implicate “all
aspects of the grammatical system of a Bantu language” (my italics) since
just one grammatical feature (syntactic phrasing) may suffice. Odden’s
English examples involve operator absorption not multiple independent
foci (Krifka 1991), and an argument type expression modified by the Ki-
matuumbi item translating only, “whose sole purpose is marking focus” (cf.
Rooth 1985), requires a “noun focal tense” (Odden 1984, 292). In any case,
semantic parameters are doubtfully learnable (Gavarró et al 2005), so they
are slim support for exotic morphosyntax. 8
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 21

The remaining task is to reconsider focus in eastern BK without tonemic

and templatic distractions. Although I lack firsthand experience with these
languages, the standard literature contains numerous helpful hints.

2. Kimatuumbi

Odden holds that “there are no data in Kimatuumbi which help choose
among competing definitions of focus” (1984, 278), but empirical differ-
ences do distinguish the diacritic/tonal (§2.1) and syntactic/accentual (§2.2)
analyses.

2.1. H deletion, diminishing returns

Odden generates Kimatuumbi verbal prosody by lexical H-tone assignment,

followed by morphosyntactic tone mapping:
“Every verb is assigned a floating H-tone, which is mapped … to the third
vowel after the subject prefix in the subjunctive and participial; it is mapped
to the second mora in dependent clauses; otherwise it is mapped to the root-
initial mora. … [T]o the extent that H is assigned to different moras in the
stem, we have support for the independence of these processes” (Odden
1996, 191f.)
By this premise, every Kimatuumbi verb should end up with one H, but
actual occurrences range from zero to two, and that’s where focus comes
in. Sticking to main clause indicatives, verb forms are subclassified along a
“three-way contrast in focal properties”: (7a/b) are “neutral”, (7c/d) “noun-
focal”, and (8a–d) “verb-focal” (Odden 1984, 289, 295 fn), although all the
forms in (8) are systematically ambiguous between focus on VP as a whole
and narrow V. In accordance with the Bantuist template (6a), even the aux-
iliated, double-H forms in (8) count as “simple verbs” (Odden 1996, 71). 9

(7) a. CLs kalang-iBtee ñapma. ‘… recently fried meat.’

b. CLs a kaplang-iBtee ñapma. ‘…fried meat.’
c. CLs kalang-aa ñapma. ‘…is frying meat.’
d. CLs kalang-a-ee ñapma. ‘…was frying meat.’
(8) a. CLs tiBp kalaang-apa ñapma. ‘…recently fried meat/fried meat.’
b. CLs a-tipB kalaang-apa ñapma. ‘…fried meat/fried meat.’
c. CLs endap kalaang-apa ñapma. ‘…is frying meat/frying meat.’
d. CLs ende-ep kalaang-apa ñapma. ‘…was frying meat/frying meat.’
22 Victor Manfredi

In retrospect, these data are less friendly to morphological focus than ad-
vertised. The “focus-neutral” status of (7a/b) – illustrated in (9) – (11) be-
low – belies the notion that “focal-sensitivity is a general property of an
entire grammar” (Odden 1984, 277). A three-valued focus feature overgen-
erates, providing no clue as to why Kimatuumbi lacks a “noun focal” past
tense; why a “neutral” progressive specially requires periphrastic ‘be’; or
why the “nine forms of the future [are] all focally neutral” (Odden 1996,
62f., my italics). 10 Barring accident, these gaps show that focus is not a
morphological feature; rather, focus interpretations emerge as a composi-
tional product of freestanding elements. Even taking the morphology at
face value, if natural language has at least one generative engine – a syntax
– and conceding that morphology can emulate at least some syntactic ef-
fects (Keenan & Stabler 2003), then for any given interpretation – includ-
ing focus – a Kimatuumbi child needs to decide whether syntax or mor-
phology is driving it. Odden admits that the motor is non-morphological in
certain “neutral” forms which yield narrowly verb-focal readings even
though unassisted by dedicated “verb focal” inflection (1996, 62; 1984,
281): 11

(9) a. A tel-iBke tuBp.

3S cook-PERF only
‘S/he only cooked.’
b. ?Mamboondop a a kaplaang-iBte, a a yapn-iBte liBpiB-liBp.
M.H 3S AUX H.fry-PERF 3S AUX H.forge-PERF NEG-NEG.H
‘M. <fried> [something], he didn’t <forge> [something].’

From (9b), Odden reasonably concludes that “the focal properties of the
verb are determined independently and are not a direct result of the selec-
tion of the verbal morphology” (1984, 281f., fn. 4). Logically, the next
steps would be (i) to name these independent factors and (ii) to check
whether the same factors also explain “noun-focal” and “verb-focal” ef-
fects. Attaining step (ii) would mean that morphological focus diacritics
perform no indispensable work in Kimatuumbi.
As to step (i), notice that both examples in (9) display object pro-drop.
Assuming that this ellipsis is necessary in order to obtain narrow verb focus
in “neutral” forms, it follows that focus is only accidentally narrow in (9) –
just as expected for nuclear stress in a phonetically nonbranching VP. 12 In
the same vein, consider a wider sample of focal interpretations which can
contextually elicit a “neutral” form (Odden 1984, 280f.):
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 23

(10) a. A a kaplang-iBtee ñapma.

‘S/he <fried meat>.’
(How did s/he feed the children?)
b. A a tepl-iBke kindooplo.
‘S/he cooked <sweet potato(es)>.’
(What did s/he cook?)
c. KiBwiBiByop a a wiBpiB-le.
‘<K. died.>’
(Why are you crying?)
d. A a wipB-le KiBwiBpiByo.
‘<K.> died.’
(Who died?)
e. A a kaplang-iBte Mambopondo.
‘<M.> fried [something].’
(Who fried [something]?)
f. A a twep-tiB kiBndolop chaapngu.
‘S/he took <my> sweet potato.’
(Whose sweet potato did s/he take?)

With direct object in-situ, the neutral form returns either broad VP focus
(10a) or narrow focus on the object (10b) – a situation familiar from Eng-
lish. Narrow subject focus (10d/e) requires VS order, as in Italian (Anti-
nucci & Cinque 1977, 124). 13 Focus on the possessor of the complement
(10f) is also consistent with nuclear stress, assuming that the postnominal
possessor of Kimatuumbi is c-commanded by its possessum.
Not only does the “neutral” form have a semantics consistent with nu-
clear stress, it also has the phonetic earmarks. Apparently without excep-
tion, a main verb bears H in two circumstances: after an aux (7b, 8, 9b, 10)
or phrase finally – even with an aux absent (11e). But with no aux, any
clausemate phonetic material to the right of the verb, as in (11a–d), is suffi-
cient to block H on the verb itself (Odden 1996, 62, 233, 287) – conforming
to the first clause of (2a) above. 14
(11) a. A kalang-iBte yopopapta eepla.
‘S/he recently fried [something] to get money.’
b. A kat-iBte kaapmba. ‘S/he recently cut rope.’
c. A tel-iBke ñama tupB. ‘S/he recently cooked only meat.’
d. A tel-iBke Mambopondo. ‘M. recently cooked [something].’
e. Mamboondop a tepl-iBiBke. ‘M. recently cooked [something].’
24 Victor Manfredi

As mentioned, Odden pre-syntactically assigns H to all verb forms, then

maps this H to the left edge of the root in finite main clauses and lastly
deletes the same H in case the verb is not clause-final (7a, 9a, 11a–d) by
“Perfective Tone Loss” (1996, 233). Curiously, though, a deletion sensitive
to the righthand context happens to fail just in case the root is introduced by
an overt aux – a lefthand context exception which accommodates the tonal
contrast between the “recent” (7a) and “remote” (7b) perfectives. The (7b)
form shows that the absence of root H depends on a larger domain than
[VO]. (7b) aside, the more general problem with the sequence of lexical H
insertion followed by phrasal H deletion is that it is theory-internal: There
can be no independent evidence for the deletion, because Kimatuumbi hap-
pens to be a language “lacking lexical tones in verbs, … similar to Safwa,
Kinga, Makua, and Kikuria” (Odden 1996, 165, cf. Cheng & Kisseberth
1979, 31). The opposite conclusion – that the H in (7b, 11e) is epenthetic
and phrasal, and therefore accentual in the normal meaning of the term – is
compelled by an additional fact: H also shows up in unauxiliated progres-
sives (forms which are not “perfectives”) when these are preceded by an
object clitic (Odden 1996, 192), thus, the H-less forms in (7c/d) contrast
minimally with (12a–b): 15

(12) a. CLs n teplek-y-a kiBndooplo.

‘… is cooking sweet potato(es) for her/him.’ [‘her/him’ = n]
b. CLs niB teplek-y-a-e kiBndooplo.
‘… was cooking sweet potato(es) for me.’ [‘me’ = niB]

To handle (7c/d) as well as (12), Odden suggests to “postulate a special rule

deleting H from a mora that is both stem initial and superstem initial … in
noun focal tenses” (1996, 195). 16 This deletion won’t occur in (12), be-
cause the H in (12) is not superstem-initial – Odden’s “superstem” being a
sub-word constituent which includes the object clitic. Of course, this “spe-
cial rule” is a mere restatement of the problem, but worse, it is inadequate:
Due to the presence of the object clitic between the subject clitic and the
root in (13), H deletion should fail to apply, but no verb form H is reported
either in applicative (13a), in which direct object scrambling puts default
focus on the applied object, or in causative (13b), where the “superstem”-
initial clitic receives a causee interpretation (Odden 1984, 74, 296): 17

(13) a. Ñamap niBiB-n kalang-y-a Mambopondo.

‘I’m frying meat for M.’ [‘M.’ = n]
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 25

b. TuBuB-n kalang-iBy-a LiBbuBlupBlee ñapma.

‘We’re making L. fry meat.’ [‘L.’ = n]

The dim prospect of fiddling with “special” H-deletion so as to affect (13)

without wrongly also applying in (12) shows that tonemic investment
yields diminishing returns: Morphological, focus-sensitive tone rules lack
generality because tone deletion lacks a coherent structural description,
needing a new patch-up rule for every new context.

2.2. Clitics all the way up

If tonemes are the lexical hypostasis of accent (§1.1 above), a natural alter-
native to H-deletion as an account of focus prosody is to turn the toneme
‘on its head’ and express the complement set of H-deletion contexts as
accent-driven H-epenthesis. Doing this for Kimatuumbi, given the absence
of lexical pitch contrasts on verb roots, leaves syntactic phrasing as the sole
basis for verb accentuation in the language. As already discussed (§1.2),
this is blocked by the morphological template (6a), as assumed by Odden
(1996, 71, 228f), but it’s possible with the phrasal syntax of (6b) which
admits a surface VP and so brings focus effects in reach of nuclear stress,
as in (2). Tentative results can be glimpsed in diverse constructions.
Unauxiliated progressives, being “noun-focal tenses”, necessarily “have
no prepausal form in main clauses” (Odden 1996, 195), thus, the nuclear
stress rule (2a) can’t explain H epenthesis in these forms – nor should it if
extrinsic factors are responsible. 18 One such factor, seen in (12) above, is
where stacking an object clitic between the subject clitic and the root
evokes a lexically spurious H. Assume with Seidl (2001) that the object
clitic is licensed in a KP shell whose null head doubles thematic material in
VP: the applicative formative [ú ] ~ [y]. 19 Then the cross-root dependency
between KP shell and applicative extension, flagged in (12' ) by coindexing,
diagnoses a spellout domain, as in (2b): The lexical predicate can’t be con-
strued with two structurally external arguments symmetrically, therefore,
KP is accented. This accent is however realized neither on the object clitic
nor on the null head of KP but on the nearest accentable syllable, which
happens to be on the left edge of VP. Auxiliated, nonapplicative (7b) is
parallel: The verb root is linearized between the overt aux [a] and the as-
pectual suffix [úte]; regardless of whether this suffix is an argument-type
expression (Manfredi 2005b) or a predicate operator, its functional compo-
sition within TP is nonassociative, therefore by (2b), Asp inhabits a differ-
26 Victor Manfredi

ent cycle from the subject/tense, so AspP is accented. Completing the paral-
lel to the KP shell in (12' ), the AspP shell in (7' b) is phonetically sub-
minimal – in this case null – and again, accent is realized on the nearest
accentable material, namely the left edge of VP. 20

(12' ) TP
3
SPEC3KP
g T 3
CL g SPEC 3VP
‡i g K 6
CL g
‡j teplek-yj-a(-ei) kindooplo

(7' b ) TP
3
SPEC 3AspP
g T 3 VP
CL g Asp 6
a g
‡i kaplang-iBteei napma

The presence of VP-initial accent in (12) contrasts with its absence in (13),
though the object clitic occurs throughout. This difference is beyond the
expressive power of tonemic-templatic description, as pointed out above:
No revision of the H-deletion rule can cover both (12) and (13) at once.
Considered in phrasal terms, however, a differentia specifica does appear:
For construction-particular reasons affecting the position of the direct ob-
ject, both (13a) and (13b) display a discontinuous VP constituent, and the
relevance of this fact to accentuation is consistent with (2a). 21
Another lexically spurious, phrasally-assigned H appears at the right
edge of any XP before a conjoint/adjoint phrase or clause. The rule’s name,
“Phrasal Tone Insertion” (Odden 1996, 234–38), tells the whole story: 22

(14) a. mpuBuBngap [ConjP na kiBndooplo].

‘rice and sweet potato(es)’
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 27

b. mpuBuBnga ntepepenga uBp [ConjP naa ñapma ]

‘wet rice and meat’
c. Ñamap [TP a a liBpiB-le].
‘S/he ate meat.’ (cf. ñama ‘meat’)
d. MuBnduBuB ntokopBma uBp [TP a wipBiB-le].
‘A sluggish person died.’ (cf. muBnduBuB ntokopBmauB ‘sluggish person’)

A third type is the “verb-focal” paradigm, cf. (8) above. The phonetic gen-
eralization is that the inflected domain (aux plus verb) displays two pitch
peaks (call them H1 and H2) whose distributions are patently phrasal. H1
occurs at the right edge of the (minimally CV) aux or aux stack. 23 H2 is
found at the right edge of the lexical V° (extension included). The tonemic-
templatic approach treats H2 as derivationally identical to the H that shows
up immediately after the aux in “focally neutral” (7b); by Odden’s rule of
Focal Flop, the first H gets “delinked” and “set afloat – later … [d]ocking
… to the final syllable” of the verb (1996, 193). Viewed accentually, how-
ever, the root-initial H in (7b) is only superficially linearized on the verb
but is generated on a later cycle. As to H2 in (8), its own appearance de-
pends on the appearance of H1, moreover its position apparently at the right
edge of VP draws semantic support from the fact in (8) that the object can-
not bear narrow scope: What is in focus in (8) is either the verb by itself or
the entire V+O sequence if any. It’s incorrect to describe the interpretation
of (8) as “translated into English with contrastive stress on the verb” (Od-
den 1984, 279). That may be true in (15a), necessarily read with narrow
scope on the verb, and it may indeed be “preferable to topicalize an object
noun phrase when a verb-focal tense is selected” (Odden 1984, 296), but
(15b/c) with the object not topicalized are both cited as felicitous answers
to questions with broad VP scope (1984, 280, 290). A narrowing of <VP>
focus to <V> in examples like (15a) could be imposed contextually, easier
than a widening from <V> to <VP>. 24

(15) a. Ne endap ly-apa ñapma,

‘I’m <eating> meat.’ (… not <frying> it)
b. A a tiBp telek-ap kindooplo.
‘S/he <cooked sweet potato(es).>’ (What did s/he <do>?)
c. E endap kalaang-apa ñapma.
‘S/he’s <frying meat>.’ (What is s/he <doing>?)
28 Victor Manfredi

The hard question for accent is how the position of H at the right edge of
V° tracks VP focus. It must be the case that the object is evacuated from
VP without removing it from the focus domain, but this surprising infer-
ence is supported by several considerations. Phonetically, Odden wonders
why “the final H-tone which derives via Focal Flop cannot undergo Retrac-
tion” onto a preceding double vowel (1996, 199), as expected in the tone-
mic framework, yielding something like the ungrammatical [*…kalaapng-aa
…] in (8) or (15b). This failure of retraction is banal under the accentual
assumption that the verb-final position of the H in (8) and (15) is already
retracted from the complement, as expected for nuclear stress. Indeed, Ki-
matuumbi presents independent evidence for retraction in (8) and (15): the
concomitant failure to shorten the vowel of the verb root, which normally
“applies when some word follows within the VP (Odden 1996, 226). For
example, the root ‘fry’ appears as [kalang] throughout (7) but as [kalaang]
in (8) and (15) where the complement is overt, as well as in (16) and (17)
where the complement is null. Note that (16) is not morphologically “verb
focal” but rather “neutral”. 25

(16) a. CLs kaplaang-iBte.

‘… recently fried [something].’
b. CLs a kaplaang-iBte.
‘… fried [something].’

(17) a. CLs tiBp kalaang-ap.

‘… recently fried [something].’
b. CLs a tipB kalaang-ap.
‘… fried [something].’

“It is not clear how this exceptionality is to be handled” says Odden (1996,
227) when shortening fails before an overt object in (8) and (15). But the
syllabic and tonal facts follow if the object is phrased external to VP in
those cases, as required by the accentual hypothesis. As to interpretation,
more Kimatuumbi data are needed, but consolation comes from a remarka-
bly similar phenomenon in Italian. According to Cardinaletti, the logical
object in (18) is deaccented in-situ, no less than in (19) where it is patently
dislocated when doubled by a clitic. In both contexts, there is a “low pitch
intonation contour … separated from the clause by an intonation break
(signalled by a comma)” (2002, 30 fn 1). Despite the similar contour, Car-
dinaletti argues for a structural difference between (18) and (19). In (19b),
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 29

for example, the participle is XP-extractable by itself, but this extraction

fails in (18b). Conversely, (18c) but not (19c) allows a quantified direct
object (Cardinaletti 2003, 36ff.). 26

(18) a. Ho già comprato, il giornale.

AUX.1S already bought the newspaper
‘I’ve already bought, the newspaper.’
b. *Comprato, non ho, il giornale.
bought NEG AUX.1S the newspaper
c. Non ha invitato Gianni nessuno.
NEG AUX.3S invited G. nobody
‘G. hasn’t invited anybody.’

(19) a. L’ho già comprato, il giornale.

3S-AUX.1S already bought the newspaper
‘I’ve already bought it, the newspaper.’
b. Comprato, non l’ho, il giornale.
bought NEG 3S-AUX.1S the newspaper
‘I haven’t bought the newspaper.’
c. *Non l’ha invitato Gianni, nessuno.
NEG 3S-AUX.3S invited G. nobody

In sum, both Kimatuumbi “verb focus” (8) and Italian emarginazione (18)
display a right-peripheral direct object which is outside of nuclear stress but
still within sentential nuclear scope – though it may be excluded from that
scope by further operations including dislocation. §3 concludes this paper
with some parallels elsewhere in eastern BK. 27

3. “Disjoint” forms

Unlike Kimatuumbi (§2), most of eastern BK shows accentual prelinking

(‘lexical tone’) in verb roots to some degree, but this difference doesn’t
undermine the accentual treatment of focus as presented above.
As illustrated in (20) from Byarushengo et al. (1976, 196, 199) and in
(21) from Hyman (1999, 153, 155), word order in Luhaya correlates with
the prosody of penultimate syllables. [HL] is found in the penults of certain
argument type phrases just in case they are sentence-final (20, 21a), nar-
rowly focused (20b), or clitic-doubled after the verb (20c/d, 21c), otherwise
30 Victor Manfredi

the penult is simple [H] (20a, 21a/b). As for certain finite verbs, the penult
is L (20a/b, 21a) unless an object clitic precedes, in which case the penult is
[H] if the verb is in narrow focus (20c/d) or is followed by a nondoubled
argument (21a), otherwise the penult is HL (21b/c). To highlight these al-
ternations, I use a simplified transcription exploiting tone/length redun-
dancy, writing [xx] for HL, [x] for H and no brackets for no H. 28

(20) a. Abak[a]zi ba bon’ omw[aa]na.

women CL see child
‘The women see the child.’ [broad focus]
b. Ba bon’ omw[aa]n’ abak[aa]zi.
CL see child women
‘They see the chíld, the women.’ [narrow focus on direct object]
c. Ba mu b[o]n’ abak[aa]zy’ omw[aa]na.
CL CL see women child
‘They sée him, the women, the child.’ [narrow focus on the verb]
d. Ba mu b[o]n’ omw[aa]n’ abak[aa]zi.
CL CL see child women
‘They sée him, the child, the women.’ [narrow focus on the verb]

(21) a. Abak[a]zi ni ba bal[i]la omw[a]na emb[uu]zi.

women ASP CLs counted.for child goats
‘The women are counting the goats for the child.’ [broad focus]
b. Abak[a]zi omw[a]na emb[u]zi ni ba zi mu bal[ii]la.
women child goats ASP CLs CLo CLo counted.for
‘The women, the child, the goats, they are counting them for him.’
c. Ni ba zi mu bal[ii]la abak[aa]zi omw[aa]na emb[uu]zi.
ASP CLs CLo CLo counted.for women child goats
‘They are counting them for him, the women, the child, the goats.’

Hyman maps H tonemes to surface distributions by autosegmental spread-

ing and deletion, filling in L tonemes. A rule changing penult [H] to [HL] is
triggered before %, a diacritic that “marks assertive focus, i.e. the end of an
assertion” (1999, 154). This entails that examples (21a/b), each with a sin-
gle [HL] contour, count as simplex assertions, but that (21c) with a total of
four such contours contains four % diacritics and therefore “four completed
assertions” in one sentence (1999, 156). While saving the diacritic ap-
proach to focus (§1.2) and the tonemic hypothesis (§1.1), Hyman also tac-
itly rejects a treatment of Luhaya % as a clause boundary, i.e. a formal ob-
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 31

ject with only indirect, syntactically-mediated impact on focus interpreta-

tion. This nondiacritic view assimilates a Luhaya ‘noun’ after a % to clitic
“right dislocation” in Romance languages, i.e. as a strategy that “allows
speakers to remove an argument, or other bit of information from within the
scope of assertion” (Byarushengo et al. 1976, 197, crediting Francesco
Antinucci p.c.). 29 The same view, according to which “the effect of the %
boundary is to ‘defocus’ non-asserted information” (Byarushengo et al.
1976, 198), is close to the syntactic treatment of Kimatuumbi in §2.2
above, modulo the extra option of prelinked H in certain verb roots as well
as the richer, triple contrast of syllable prosodies (HL, H, ‡). Both consid-
erations favor the accentual treatment of H: A phrasal regime of nuclear
stress expects the nonparsing of H in an unperturbed, broad focus VP lack-
ing object clitics (20a/b), as well as the positional “enhancement” (Halle &
Vergnaud 1987, 37) of H to [HL]. Neither phenomenon fits in a phonemic
framework: The former requires an arbitrary deletion rule, and the latter
entails mysterious action-at-a-distance from the % trigger. 30
Based on available published data, it’s hard to argue directly against
Hyman’s (1999) theory-internal claim of “four completed assertions” in
(21c). A sharper problem for the claim is posed by narrow focus data like
(20b–d): The most natural interpretation of the English glosses allows one
assertion per example despite the presence of multiple [HL] contours, so
either the glosses or the multiple focus analysis must be wrong. 31 Con-
versely, the freedom of post-verb word order in (20c/d) can be taken as
evidence for the nondiacritic 1976 treatment of % – as a sentence boundary
lacking intrinsic semantic content – in view of the parallel word order free-
dom in Spanish examples like (23), which has been argued to diagnose
right dislocation (Zubizarreta 1998, 156). 32

(22) a. Le envió un regalo, María, a Mamá.

CL sent a gift M. to mom
‘She sent her a present, Maria, to Mom.’
b. Le envió un regalo, a Mamá, María.
CL sent a gift to mom M.
‘She sent her a present, to Mom, Maria.’

Another reason not to build semantic assertion into the % boundary is that
real semantic assertion is transparently implicated in a different Luhaya
contrast, discussed by Hyman (1999, 160–62) under the rubric of “con-
joint/disjoint” phrasing of post-verb material (cf. Schadeberg 2004). A
32 Victor Manfredi

relevant pattern in both languages is that lexically prelinked H is “sup-

pressed” in environments of what Hyman & Watters (1984) call “auxiliary
focus”. Because this distribution differs from the suppression of verb root
H in (20) and (21), conditioned by % and tied by hypothesis to a focus dia-
critic, Hyman is obliged to introduce “a secondary focus that has become
morphologized and which, therefore, only imperfectly corresponds to the
semantics that motivates it” (1999, 162). Assuming with Hyman that both
phenomena are indeed focus-related, his dilemma can be blamed on use of
a focus diacritic which has at most two values, minus (3a) and plus (3b).
An indirect analysis which reads focus from syntax naturally admits more
than one kind of focus related phrasing.
In Kirundi, as first observed by Meeussen (1959, 119–28), the “disjoint”
aux ra is necessary for the appearance of H in the lexically accented root
‘pick’ in (23a). Absence of ra suppresses the H (23b). The prosodic con-
trast is missing with a lexically unaccented root, as in (24) from Nday-
iragije (1998), leaving auxiliation as the only cue. The correlation of aux-
iliation and lexical H in (23) follows from nuclear stress if the aux and VP
form separate cycles, as is guaranteed by the pleonastic character of ra,
which would otherwise be blocked by economy. As to interpretation, the
presence of ra allows either the verb root alone or the whole VP to consti-
tute new information (24a), whereas a narrow information focus on the
object, as in a content question (24b), entails ra’s absence. The correlation
is 100%, because the ra auxiliary is pleonastic apart from focus considera-
tions, much like affirmative (i.e stressed, declarative) English do, as help-
fully hinted by Ndayiragije’s translation of (24a).

(23) a. N-ra áam-uur-a intore.

1S-ra pick-EXT-V plum
‘I’m picking plums.’ [disjoint]
b. N aam-uur-a intore.
1S pick-EXT-V plum
‘I’m picking the plum.’ [conjoint]

(24) a. Yuvinari a-á-ra somye ibitabo.

Y. a-a-ra read books
‘Y. read/did read books.’ [disjoint]
b. Yuvinari a-á (*-ra) somye iki?
Y. a-a-ra read what
‘What did Y. read?’ [conjoint]
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 33

According to Goldsmith, Kirundi’s “-ra- Focus marker … is itself a recent

innovation shared with Kinyarwanda but no other languages (a reanalysis
of what was formerly the present tense marker, presumably)” (1985, 127).
Even so, the focus effect of auxiliation is not limited to this pleonastic
form, as shown by Setswana. Literal pause is not necessary after a disjoint
verb (Chebanne et al. 1997, 56), but disjoint phonetic traits remain prosodic
in the broad sense: a language-particular mix of pitch/timing and affixa-
tion/auxiliation, both consistent with an analysis by phrasing. Creissels
(1996, 110f.) reports systematic tonal minimal pairs in the Setswana present
perfect (25) and negative nonpast (26). In the superimposed paradigms in
(27), the tonal difference appears in tandem with auxilation, whether the
aux in question is pleonastic (affirmative nonpast) or a contentful sentence
operator (future). If the aux is substantive, the audible conjoint/disjoint
distinction reduces to tone alone. In contexts with a nonroot character, in-
cluding those labeled pluperfect or consecutive, the distinction is phoneti-
cally neutralized altogether, yielding radical ambiguity, (28).

(25) a. Bap jep-lez lep boznep.

3P eat-PERF with 3PL
‘Theyi have eaten, even theyi.’ [disjoint]
b. Bap jez-lep lep boznep.
3P eat-PERF with 3PL
‘Theyi have eaten with themj.’ [conjoint]

(26) a. Gaz bap bipn-ez lep boznep.

NEG 3P dance-V with 3PL
‘Theyi don’t dance/aren’t dancing, even theyi.’ [disjoint]
b. Gaz bap bipn-ep lep boznep.
NEG 3P dance-V with 3PL
‘Theyi don’t dance/aren’t dancing with themj.’ [conjoint]

(27) a. Bap {az/tlapaz} bipn-ap lep boznep.

3P AUX dance-V with 3PL
‘Theyi {‡/will} dance, even theyi.’ [disjoint]
b. Bap {‡/tlapaz} bipn-az lep boznep.
3P AUX dance-V with 3PL
‘Theyi {‡/will} dance with themj.’ [conjoint]
34 Victor Manfredi

(28) Bap nez bap bipn-nep lep boznep.

3P PERF 3P dance-PERF with 3PL
a. ‘Theyi had danced, even theyi.’ [disjoint]
b. ‘Theyi had danced with themj.’ [conjoint]

Setswana differs from Kirundi semantically in that the Setswana disjoint

form puts narrow focus on the verb excluding the direct object (Creissels
1996, 113f.). A similar difference even divides Kirundi from Kinyarwanda,
where the ra form excludes an adverbial (29a) or a direct object (30a) from
the domain of new information (Givón 1975b, 194; tone outside the aux not
marked in the source); and where the ra form is also impossible in a nega-
tive or relative predicate (reported but not illustrated in the source). The
non-ra form can’t be followed by a discourse-old (pronominalized or
scrambled) object (31b), which the ra form allows (31a).

(29) a. *Yohani y-à-rá koze vuuba/ mumusozi.

Y. y-a-ra work fast/ in the village [disjoint]
b. Yohani y-à koze vuuba/ mumusozi.
Y. y-a work fast/ in the village
‘Y. worked fast/in the village.’ [conjoint]

(30) a. *Yohani y-à-rá riiye iffi.

Y. y-a-ra eat fish [disjoint]
b. Yohani y-à riiye iffi.
Y. y-a eat fish
‘Y. ate (a) fish.’ [conjoint]

(31) a. Yohani y-à-rá yi-riiye (iffi).

Y. y-a-ra CL-eat fish
‘Y. ate it (the fish).’ [disjoint]
b. *Yohani y-à yi-riiye.
Y. y-a CL-eat [conjoint]

In Setswana, auxiliation is neither necessary nor sufficient for a disjoint

form which is narrow: (25a) is disjoint with no aux, whereas the nonfuture
variant of (27a) is indeed auxiliated but could also be conjoint depending
on the tone (27b). In Kirundi by contrast, all disjoint forms described by
Meeussen are auxiliated, with the further prosodic correlate noted above,
and they are all broad. Crucially, in the Kirundi conjoint, the verb root is
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 35

deaccented – a robust correlate of old information (Williams 1997) – so the

fact that it’s narrow is no surprise. Now, why is the Setswana conjoint
broad? Prosody distinguishes two subcases. Assuming that accent appears
as a [HL] pitch contour, an accented verb is broad conjoint after a ‘middle
field’ aux – one which follows the subject clitic, as in (27b) – but narrow
disjoint otherwise – if nothing separates the subject clitic and the verb root,
as in (25), (26), and (28). In other words, focus projects from an accented
verb iff an aux directly precedes. This generalization also holds in Kirundi,
in fact trivially so because verb root accent (in this language, a simple H) is
limited to auxiliated forms. Turning to Setswana sentences which lack a
middle field aux, an accented verb is narrow/does not project, but this is not
a comparative problem because Kirundi ra has no Setswana counterpart,
for the independent reason already stated. The remaining case is an un-
stressed (and therefore nonauxiliated) verb in Kirundi, which is narrow
conjoint; the corresponding form in Setswana is broad. Both languages
share this principle of stress-to focus mapping: focus projects from a
stressed verb to VP (is broad) only after a middle field aux. In the absence
of either VP-internal stress or an aux, focus includes the object (is conjoint)
in both languages. Thus far unexplained is the difference that neutral phras-
ing is broad in Setswana but narrow in Kirundi; tentatively, this could be
related to the different accentuation of argument type phrases (‘nouns’),
presumably an independent fact.

4. Perspective

The foregoing shows that the choice to forego tonemes and Bantuist tem-
plates brings many BK prosodic phenomena under a general theory of
phrase structure, a hypothesis of prosodic unity (4b), as well as a modular
architecture of information structure (2c). Numerous language-particular
consequences across BK and more widely remain to be drawn, e.g. the
appearance of left-peripheral focus and “compulas”, i.e. copular focus
markers (Manfredi 1987, 110, cf. Bergvall 1987).

Notes

This paper revises Manfredi (2004b, 2005a). Thanks to E. O. Aboh, ’SB.

AdepsBoBlap, A. Akinlabip, A. Bachrach, M. Bamba, L. Brunetti, A. Cardi-
36 Victor Manfredi

naletti, G. Cinque, R.-M. Déchaine, L. Domínguez, S. Duanmu, I. Fiedler,

J. Frampton, G. Giusti, F. del Gobbo, M. Halle, K. Hartmann, A. Ki-
menyi, K. Kinyalolo, G. Longobardi, J. Lowenstamm, C. Mayr, A.
Nevins, D. Odden, J. Rooryck, V. Samek-Lodovici, B. Reineke, T.
Schadeberg, R. Schuh, A. Schwarz, A. Seidl, K. Szendröi, C.
Upchezchupkwu, M. Wagner, J. v.d. Wal, M. Zimmermann and M.-L. Zubi-
zarreta.
1. The term Bantu is sharp ideology but fuzzy linguistics. In apartheid Zuid
Afrika it was the state synonym for “black” – as in this parliamentary
speech by Minister P.W. Botha in 1964 (quoted by McGreel 2006):
“I am one of those who believe that there is no permanent home for even a
section of the Bantu in the white area of South Africa and the destiny of
South Africa depends on this essential point. If the principle of permanent
residence for the black man in the area of the white is accepted then it is
the beginning of the end of civilisation as we know it in this country.”
Turning the tables, Sowetan students replied to Botha’s Bantustans with
black consciousness (Biko 1972). The Bantu mystique persists in U.S.
Afrocentrism (Holloway & Vass 1993) and also in linguistics, even
though it is “impossible to draw a clear line between Bantu, however de-
fined, and non-Bantu Niger-Congo” (Nurse & Philippson 2003, 5). In
Neogrammarian terms, the nearest affiliation of any two ‘Bantu’ lan-
guages above the local cluster is the Benue-Kwa “dialect continuum”
(Williamson & Blench 2000, 17f., 27, cf. Greenberg 1963, 39, Givón
1975a), also called Volta-Congo (Stewart 1976, 1994).
2. Odden (p.c.) notes that “[l]exicality is not the real issue, though that is of-
ten a fact that makes people comfortable with the idea of something being
phonological.” Indeed, and once ‘tone’ is phrasal, a reanalysis in terms of
nuclear stress comes within reach.
3. The “unless” clause of (2a), accentually weighting the predicate over the
subject, can be dropped if the default focus operator is the middle field as-
sertion head, Sigma (Piñón 1992, Surányi 2004, Manfredi 2004b, cf.
Gleitman 1969, Laka 1990). A “non-associative domain”, which by (2b)
must be multi-cyclic, is defined as one in which compositional order af-
fects truth value.
4. Bamba & Liberman state that in Manding (Niger-Congo) “(some of) the
functions of English intonational focus are performed by explicit and or-
dinary morphological marking” (1999, 1) but then go on to note that this
marking is homophonous with the non-negative copula – anything but
“ordinary”!
5. The tonemic format of (5) may have a sociological explanation: Pike “util-
ized the study of English intonation as one step in the teaching of [S.I.L.]
students how to reduce tone languages to writing” (1947, 131), but sus-
ceptibility to commonsense Africanist analysis persisted half a century
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 37

later when, for example, Ladd defended “tonal targets” in English intona-
tion based on “the existence of languages like Yoruzbap in which it is un-
controversial that the system of lexical tones is based on distinctive levels
… Once we have such a theory, it is plausible to assume that it will apply
to English or Dutch as well” (1996, 61).
6. Guthrie exempts from the template “some of the languages of the extreme
north-western part of the Bantu area” (1948, 24 fn. 2). Templatics and
tonemics have a close affinity, e.g. in Izgbo (BK), Clark’s (1989) templatic,
level ordering analysis – developing Goldsmith (1976) in tune with
Clements & Goldsmith (1980) – is obliged to posit a phonemic “down-
step” (an unpredictable tonemic juncture) whereas a syntactic, accentual
analysis can derive this juncture from phrase boundaries, e.g. between aux
and V or within the Genitive DP (Déchaine 1992, 1993, 497–520, Man-
fredi 1993, cf. Clark 1980).
7. Odden (1996, 228f.) considers the possibility that narrow focus on V re-
sults from verb raising, i.e. if the X0 in (6a) = “Infl”, or alternatively from
extraposition of [-focus] arguments. Neither kind of rephrasing is inde-
pendently motivated, however. In (6b), the label V° is intended to cover
the categorial conflations of “lexical syntax” à la Hale (1995).
8. I can’t find a Kimatuumbi example with a ‘perfective’/focally ‘neutral’
verb and a noun phrase or PP modified by ‘only’. One may exist, given
the situation with wh-expressions: Odden initially says that these “cannot
appear with the verbfocal tense, but may appear only with the noun focal
tense” (1984, 292) but later observes that “Intrinsically focused elements
such as wh-words may appear in clauses containing a focally neutral verb”
(1996, 62). The crucial case for focus-sensitivity would combine a neutral
verb, a wh expression, and a distinct nominal or prepositional phrase
modified by ‘only’. See also discussion of the data in (9) below. Both
Hyman’s “focus prominence” and Odden’s “focus sensitivity” recall the
functional parameter of subject prominence vs. topic prominence (Li &
Thompson 1976).
9. Throughout (7) and (8), the verb root is kala(a)ng ‘fry’ and the logical ob-
ject ñapma ‘meat’; the remaining morphemes, glossable only with diffi-
culty, are discussed in the text. Italics in the translations mark obligatory
focus as described in the source. I’ve cosmetically enhanced Odden’s
transcription by changing hyphen to wordspace between the aux and the
lexical verb root and by bolding the inflectional H tones. The consistently
lengthened final vowel of the verb throughout (7) and (8), spelled as a
double letter, has no morphological import (Odden 1996, 253ff.), just re-
flecting a phonetic constraint on the left edge of bisyllabic nouns (e.g.
ñama ‘meat’).
38 Victor Manfredi

10. The nine Kimatuumbi futures include both “periphrastic” forms and not.
In the closely related language Makua, an effable focus distinction in the
future is reported (Stucky 1979a, 363), but without illustration.
11. The question mark on (9b) denotes “acceptable, if not optimal” grammati-
cality (Odden 1984, 281). In the translation of (9b) and henceforth, <angle
brackets> mark actual scope elicited in context versus italics which indi-
cate obligatory focus as described without explicit tests. In the published
translation of (9b), I correct a typo : “…forget”. (9b)’s interlinear gloss
notes phrasally assigned H tones on the subject, verb, and negation –
without prejudging their status as tonemes versus accents. In the glosses, I
mark person/number (e.g. 3S) rather than noun class. In examples (9)–
(11), the subject clitic is [a], denoting a third person singular human sub-
ject which is traditionally called class 1 (Odden 1996, 34). Examples (7b,
8b, 9b, 10, 11) include an aux [a], labeled “remote” (Odden 1996, 57) and
linearized immediately after the subject clitic. Thus, (9b, 10, 11) show
homophony between the default/epenthetic aux and third singular human
pro. Welmers (1973b) expresses scepticism about standard Bantuist tense
labels denoting multiple degrees of remoteness.
12. A phrasal origin of focus in (9) is further supported by the appearance of
the doubled, specifically phrase-final form of negation (lùú-lú ) in the sec-
ond clause. A desideratum is defining the general freedom of object pro-
drop in this language.
13. In contrast to Italian as well as to Kimatuumbi, VS order denotes broad
(“all new”) focus in Catalan (Vallduví 1990), and VOS is generally possi-
ble for narrow subject focus in Spanish (Zubizarreta 1998). Plausible
sources of this variation – case assignment and clitic doubling – remain
controversial in these languages.
14. The locution complement is used here rather than direct object in order to
denote an argument phrase in immediate post-verb position, thus includ-
ing the focused, inverted subjects in (10d/e). This accords with Cinque’s
view that, for purposes of nuclear stress, a non-topical post-verb argument
is more deeply embedded than the verb itself “on the recursive side”.
Nevertheless, the configurational definition is distinguished from an edge-
based theory by cases of right detachment; faced with these, the edge-
based theory needs to proliferate prosodic boundaries, while the syntactic
view can refer to phrasal discontinuity.
15. For me, epenthetic describes minimal overt material not drawn from the
numeration. For example, prosodic unity (4b) claims that children know
how to construct metrical grids ex nihilo although of course they can learn
lexicalized accents.
16. On the same page, Odden floats a second, more technical solution without
enthusiasm.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 39

17. Both exceptional forms have two overt phrasal objects; I assume that
scrambling of the direct object is obligatory in (13a). Focus on the applied
object is implied in the description of (13a); nothing is said about focus in
the causative (13b). Although relevant data are lacking, it is implied that
the rule designed for (12) cannot be collapsed with “Perfective Tone
Loss”, i.e. that even with an object clitic an unauxiliated recent past still
lacks stem H, as in hypothetical (i), based on Odden (1996, 74, ex 112):
(i) [CLs n telek-iB Mamboondo kiBndooplo.]
[‘…recently cooked sweet potato(es) for M. (= n)’]
18. Apart from progressives, the source mentions no other “noun focal
tenses”.
19. Cf. Kisseberth & Mmusi (1990) for similar facts in Setswana. The poly-
synthetic view of applicatives (Baker 1988b, c), saving templatic (6a), de-
rives the “verb extension” as a lexical head raised from an abstract PP in a
downward lexical cascade (Baker 1988a). Hyman (2003) reviews prob-
lems for applicative incorporation – problems eluded by the inverse archi-
tecture in (6b) and (12' ) where the applicative licensor is above, not be-
low, its spellout position, consistent with upwardly exploded treatments of
clausal superstructure (Rizzi 1997, Manzini & Savoia 1998, Cinque
1999).
20. In rightward H displacement forms like (7' b), Goldsmith marks the on-
setless aux [a] as “a ‘post-High’ morpheme” (1985, 123).
21. Note the scopal difference between (12) and (13a). In lieu of a phrasing
for (13), I offer the slogan flat syntax, flat prosody.
22. The phrase structure of conjunction is notoriously murky, but the two en-
vironments potentially form a natural class (Williams 1989). The similar-
ity is underlined by the accentuation of the scrambled object in (14c).
23. “In general, for all verb-focused tenses there is an H-tone on the last mora
of the tense-aspect prefix” (Odden 1996, 194).
24. For example, if the object in (15a) is anaphoric; Kenesei (2005) argues
that narrowing is the only way to get narrow V focus.
25. Odden (1996, 225) gives examples in which descriptive shortening applies
twice in one form: once to the root vowel and once to the complex of ap-
plicative extension plus tense suffix. However, the likely operation of vo-
calic epenthesis in the latter context, though little explored in Bantuist lit-
erature, places the theoretical relevance of shortening in doubt.
Instrumental data may clarify the correlation between length and pitch dy-
namism along the lines of the Chinese literature referred to above. Odden
speaks of shortening “providing a second argument for syntactic constitu-
ency in potentially unclear cases” (p. 233), hinting at acoustic unclarity.
26. Examples tweaked. The presence of subject focus in (18c) is not indicated
in Cardinaletti’s data but is implied in descriptive remarks (2002, 32).
Emarginazione is the closest Italian counterpart to Germanic “metrical in-
40 Victor Manfredi

visibility” (Zubizarreta 1998, 49). Although the typical example in both

languages is anaphoric/topical, the requirement is less stringent as shown
by (18c) and by (i) from Cinque (1993, 255 fn. 19 citing G. Grewendorf
p.c., cf. also Wagner 2005, 211 and references cited there):
(i) dass der Árzt bereitwillig [FP jeden Patienten [untersúchte]].
that the doctor willingly every patient investigated
‘…that the doctor willingly examined every patient.’
In (18c), nontopical material in XP is deaccented, and in (i), the transitive
verb fails to deaccent despite occupying final position.
27. My treatment of Kimatuumbi differs notationally from Pulleyblank’s
(1982) use of a Halle & Vergnaud accentual grid plus tonemes. The simi-
larity is that reference to accent lets Pulleyblank avoid rules which first
assign tones lexically and then delete them in phrasal contexts. A substan-
tive difference – ignored here – stems from Pulleyblank’s emphasis on
nominal prosody.
28. Hyman (1999) abstracts away from automatic vowel elision and glide
formation, both of which are marked as apostrophes by Byarushengo et al
(1976), as in (20). Hyman’s glosses for (21c/d) contain the apparent typo
“have counted” for ‘are counting’.
29. Right dislocated items were accordingly described as either “after-
thoughts” or “recapitulation” (1976, 201f.). Similarly, Hyman & Bya-
rushengo state that “the focus (or assertion) of a Haya utterance is placed
last in a clause” (1984, 70, their italics).
30. Unitary formal treatment of phonetic [HL] has a precedent in eastern BK:
Luganda (Hyman & Katamba 1993, 49).
31. Corresponding forms with narrow focus on a sentence final verb, like (i)
from Byarushengo et al. (1976, 201), or a broad-focus sentence-final verb,
like (ii) from Hyman & Byarushengo (1984, 94), contain no instances of
[HL], but it is not clear how the 1999 analysis can avoid generating [HL]
there, or on the finite verbs in (20c/d) for that matter, while still correctly
producing [HL] contours on the verbs in broad focus sentences like (21
b/c). (Throughout, the verb root belongs to the prelinked H class.)
(i) a. Kak[u]lw’ abak[a]zi ba mu b[o]na. (ii) Ba mu k[o]ma.
K. women CL CL see CL CL tie.up
‘Kakulu, the women (they) sée him.’ ‘They tie him/her up.’
b. Abak[a]zi Kak[u]lu ba mu b[o]na.
women K. CL CL see
‘The women, Kakulu, they sée him.’
32. In (22), nuclear stress is transcribed as underlining.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 41

References

Abramson, Arthur S.
1978 Static and dynamic acoustic cues in distinctive tones. Language &
Speech 21, 319–25.
1979 The noncategorical perception of tone categories in Thai. In Fron-
tiers of speech communication research, Björn Lindblom and Sven
Ohman (eds.), 127–134. London: Academic Press.
Ajipbopyez, OB.
2005 Topics on Yoruzbap nominal expressions. Dissertation, University of
British Columbia, Vancouver.
Akinlabip, Akinbiyi M.
1985 Tonal underspecification and Yoruzbap tone. Dissertation, University
of Izbazdazn.
Antinucci, Francesco, and Guglielmo Cinque
1977 Sull’ordine delle parole in Italiano; l’emarginazione. Studi di gram-
matica italiana 6, 121–46.
Arregi, Karlos
2002 Focus on Basque movements. Dissertation, M.I.T., Cambridge, MA.
Arregi, Karlos, and Isabel Oltra-Massuet
2005 Stress-by-structure in Spanish. Linguistic Inquiry 36, 43–84.
Baker, Mark
1988a Incorporation; a theory of grammatical-function changing. Univer-
sity of Chicago Press.
1988b The mirror principle and morphosyntactic explanation. Linguistic
Inquiry 16, 373–415.
1989 Elements of a typology of applicatives in Bantu. In Current Ap-
proaches to African Linguistics 7, John P. Hutchison and Victor
Manfredi (eds.), 111–24. Dordrecht: Foris.
1996 The Polysynthesis Parameter. Oxford University Press.
Bamba, Moussa
1991 De l’interaction entre tons et accent. Dissertation, Université du
Québec à Montréal.
Bamba, Moussa, and Mark Liberman
1999 Focus in Manding.
http://www.ldc.upenn.edu/myl/mawu_LSA_handout.pdf
Beckman, Mary E., and Janet Pierrehumbert
1986 Intonational structure in Japanese and English. Phonology Yearbook
3, 255–309.
Benincà, Paola
1988 L’ordine degli elementi della frase e le construzioni marcate. In
Grande grammatica italiana della consultazione 1, Lorenzo Renzi
(ed.), 115–94. Bologna: Il Mulino. Benson, Thomas G. (ed.)
42 Victor Manfredi

1964 Kikuyu-English Dictionary. Oxford University Press.

Bergvall, Victoria L.
1987 Focus in Kikuyu and Universal Grammar. Dissertation, Harvard
University.
Biko, Steve
1972/78 Let’s talk about Bantustans. I Write What I Like; a selection of his
writings edited with a personal memoir by Aelred Stubbs. 81–86.
London: Bowerdean.
Bolinger, Dwight
1951 Intonation; levels vs. configurations. Word 7, 199–210. Reprinted in
Forms of English; accent, morpheme, order, 3–16. Cambridge,
Mass.: Harvard University Press.
Bresnan, Joan, and Sam A. Mchombo
1987 Topic, pronoun and agreement in Chichewa. Language 63, 741–82.
Brody, Michael
1990 Remarks on the order of elements in the Hungarian focus field. In
Approaches to Hungarian 3; Structures & Arguments, István Kene-
sei (ed.), 95–122. Szeged: JATE.
Byarushengo, Ernest Rugwa, Larry M. Hyman, and Sarah Tenenbaum
1976 Tone, accent and assertion in Haya. In Southern California Occa-
tional Papers in Linguistics 5: Studies in Bantu Tonology, Larry M.
Hyman (ed.), 183–205. Los Angeles: University of Southern Cali-
fornia.
Cardinaletti, Anna
2002 Against optional and null clitics; right dislocation vs. marginaliza-
tion. Studia Linguistica 56, 29–57.
Carstens, Vicki
2001 Multiple agreement and Case deletion; against M-incompleteness.
Syntax 4, 147–63.
Chebanne, A., Creissels, Denis, and Nkhwa, H.
1997 Tonal Morphology of the Setswana Verb. München: Lincom Europa.
Cheng, Chin-Chuan, and Charles W. Kisseberth
1979 Ikorovere Makua tonology (part 1). Studies in the Linguistic Sciences
9, 31–63.
1981 Ikorovere Makua tonology (part 3). Studies in the Linguistic Sciences
11, 181–202.
Chomsky, Noam
1971 Deep structure, surface structure and semantic interpretation. In
Semantics; an interdisciplinty reader in linguistics, philosophy &
psychology, Danny D. Steinberg and Leon A. Jakobovits (eds.), 183–
216. Cambridge University Press.
Chomsky, Noam, and Morris Halle
1968 The Sound Pattern of English. New York: Harper.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 43

Chomsky, Noam, Morris Halle, and Fred Lukoff

1956 On accent and juncture in English. In For Roman Jakobson, Morris
Halle, Horace G. Lunt, Hugh McLean, Cornelis H. van Schooneveld
(eds.), 64–80. The Hague: Mouton.
Cinque, Guglielmo
1991 Types of A’-Dependencies. Cambridge, Mass.: MIT Press.
1993 A null theory of phrase and compound stress. Linguistic Inquiry 24,
239–98.
1999 Adverbs & Functional Heads, a cross-linguistic perspective. Oxford
University Press.
Clark, Mary M.
1980 On the treatment of syntactically-distributed downstep. Studies in
African Linguistics 11, 101–37.
1989 The Tonal System of Izgbo. Dordrecht: Foris.
Clements, George N.
1981/83 The hierarchical representation of tone features. In Harvard Studies
in Phonology 2, George N. Clements (ed.), 50–108/ In Current Ap-
proaches to African Linguistics 1, Ivan R. Dihoff (ed.), 145–76.
Dordrecht: Foris.
1984 Principles of tone assignment in Kikuyu. In Autosegmental Studies in
Bantu Tone, George N. Clements and John Goldsmith, 281–339.
Dordrecht: Foris.
Clements, George N., and Kevin Ford
1979 Kikuyu tone shift and its synchronic consequences. Linguistic In-
quiry 10, 179–210.
Clements, George N., and John Goldsmith
1980 What is downstep? A reply to Clark [1980]. Studies in African Lin-
guistics 11, 239–54.
Connell, Bruce A.
2000 The perception of lexical tone in Mambila. Language & Speech 43,
163–82.
Creissels, Denis
1994 La tonalité des finales verbales et la distinction entre formes verbales
conjointes et formes disjointes en tswana. Africana linguistica 11 (=
Annales du Musée Royale de l’Afrique Centrale, Sciences humaines
142), 27–45.
1996 Conjunctive and disjunctive verb forms in Setswana. South African
Journal of African Languages 16, 109–15.
Déchaine, Rose-Marie
1992 Inflection in Izgbo and Yoruzbap. MIT Working Papers in Linguistics
17, 95–120.
1993 Predicates across categories; towards a category-neutral syntax.
Dissertation, UMass, Amherst.
44 Victor Manfredi

2001 On the left edge of Yoruzbap complements. Lingua 111, 81–130.

Déchaine, Rose-Marie, and Victor Manfredi
1998 SVO ergativity and abstract ergativity. Recherches Linguistiques de
Vincennes 27, 71–94.
Dilley, Laura
2005 The phonetics & phonology of tonal systems. Dissertation, M.I.T.,
Cambridge, Mass.
Duanmu, San
1994 Against contour tone units. Linguistic Inquiry 25, 555–608.
2004 Tone and non-tone languages; an alternative to language typology
and parameters. Language & Linguistics 5, 891–923. <www-
personal.umich.edu/~duanmu/ToneTyp04.pdf>.
2005 The tone-syntax interface in Chinese; some recent controversies. In
Cross-Linguistic Studies of Tonal Phenomena, Historical Develop-
ment, Tone-Syntax Interface & Descriptive Studies, Shigeki Kaji
(ed.), Institute for the Study of Languages & Cultures of Asia & Af-
rica, Tokyo University of Foreign Studies. <www-
personal.umich.edu/~duanmu/ToneSyntax05.pdf>.
EpmepnanjoB, E. ’N.
1981/85 Auxiliaries in Izgbo Syntax. Dissertation, University of Izbazdazn/ Indi-
ana University Linguistics Club, Bloomington.
Fromkin, Victoria
1978 Introduction. In Tone; a linguistic survey, Victoria Fromkin (ed.), 1–
4. New York: Academic Press.
Gandour, Jack
1983 Tone perception in far eastern languages. Journal of Phonetics 11,
149–75.
Gavarró, Anna, Ana Teresa Pérez-Leroux, and Thomas Roeper
2005/ to appear Definite and bare noun contrasts in child Catalan. In The
Romance Turn, Vicenç Torrens and Linda Escobar (eds.),
Amsterdam: Benjamins.
Givón, Talmy
1975a Serial verbs and syntactic change: Niger-Congo. In Word order and
word order change, Charles N. Li (ed.), 49–112. Austin/ London:
University of Texas Press.
1975b Focus and the scope of assertion. Studies in African Linguistics 6,
185–205.
Gleitman, Lila R.
1969 Coordinating conjunctions in English. In Modern Studies in English:
Readings in Transformational Grammar, David A. Reibel and San-
ford A. Schane (eds.), 80–112. Englewood Cliffs, New Jersey: Pren-
tice-Hall.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 45

Goldsmith, John
1975 Tone melodies and the autosegment. OSU Working Papers in Lin-
guistics 20, 135–47.
1976 Autosegmental phonology. Dissertation, M.I.T., Cambridge, Mass.
1978/81 English as a tone language. Communication & Cognition 11, 453–
76. /In Phonology in the 1980’s, Didier L. Goyvaerts (ed.), 287–
308. Ghent: Story-Scientia.
1982 Accent systems. In The Structure of Phonological Representations
Part 1, Harry van der Hulst and Norval Smith (eds.), 47–63.
Dordrecht: Foris.
1984 Tone and accent in Tonga. In Autosegmental Studies in Bantu Tone,
George N. Clements and John Goldsmith (eds.), 19–51. Dordrecht:
Foris.
1985 Bantu –a–; the far past in the far past. Studies in African Linguistics
Supplement 9, 123–34.
1987 The rise of rhythmic structure in Bantu. In Phonologica 1984. Pro-
ceedings of the Fifth International Phonology Meeting, Wolfgang U.
Dressler, Hans C. Luschützky, Oscar E. Pfeiffer, and John R. Renni-
son (eds.), 65–78. Cambridge University Press.
Greenberg, Joseph H.
1963 The Languages of Africa. The Hague: Mouton. [= International
Journal of American Linguistics 29.1].
Gussenhoven, Carlos
1983a Focus, mode and the nucleus. Journal of Linguistics 19, 377–417. [=
Gussenhoven 1984, 11–62].
1983b Testing the reality of focus domains. Language & Speech 26, 61–80.
[= Gussenhoven 1984, 151–75].
1984 On the Grammar & Semantics of Sentence Accents. Dordrecht:
Foris.
2004 The Phonology of Tone & Intonation. Cambridge University Press.
Guthrie, Malcolm
1948 Bantu Word Division; a new study of an old problem. [= Interna-
tional African Institute Memorandum 22.] Oxford University Press.
Hale, Ken
1995 Universal grammar and the necessity of linguistic diversity. Presi-
dential address, Linguistic Society of America, New Orleans. Pub-
lished as: Universal grammar and the roots of linguistic diversity.
MITWPL 28 (1996),137–61.
Halle, Morris
1959 The Sound Pattern of Russian. The Hague: Mouton.
46 Victor Manfredi

Halle, Morris, and Jean-Roger Vergnaud

1982 On the framework of autosegmental phonology. In The Structure of
Phonological Representations Part 1, Harry van der Hulst and Nor-
val Smith (eds.), 65–82. Dordrecht: Foris.
1987 An Essay on Stress. Cambridge, Mass.: MIT Press.
Holloway, Joseph E., and Winifred K. Vass
1993 The African Heritage of American English. Bloomington: Indiana
University Press.
Hombert, Jean-Marie
1976 Perception of tones in bisyllabic nouns in Yoruzbap. Studies in African
Linguistics Supplement 6, 109–21.
Hyman, Larry M.
1974 The great Izgbo tone shift. In Third Annual Conference on African
Linguistics, Erhard Voeltz (ed.), 111–25. Bloomington: Indiana Uni-
versity Press.
1985a Word domains and downstep in Bamiléké-Dschang. Phonology
Yearbook 2, 47–83.
1985b Dependency relations in syntax; the mysterious case of the empty
determiner in Aghem. Studies in African Linguistics Supplement 9,
151–56.
1989 Accent in Bantu; an appraisal. Studies in the Linguistic Sciences 19,
115–34.
1999 The interaction between focus and tone in Bantu. In The Grammar of
Focus, Georges Rebuschi and Laurice Tuller (eds.), 151–77. Am-
sterdam/Philadelphia: Benjamins.
2003 Suffix ordering in Bantu; a morphocentric approach. In Yearbook of
Morphology 2002, Geert Booij and Jaap van Marle (eds.), 245–81.
Kluwer Academic Publishers.
2004 How to become a ‘Kwa’ verb. Journal of West African Languages
30, 69–88.
Hyman, Larry M., and Ernest Rugwa Byarushengo
1984 A model of Haya tonology. In Autosegmental Studies in Bantu Tone,
George N. Clements and John Goldsmith (eds.), 52–103, Dordrecht:
Foris.
Hyman, Larry M., and Francis X. Katamba
1993 A new approach to tone in Luganda. Language 69, 34–67.
Hyman, Larry M., and John R. Watters
1984 Auxiliary focus. Studies in African Linguistics 15, 233–73.
Idsardi, William J., and Thomas C. Purnell
1997 Metrical tone and the Elsewhere Condition. Rivista di Linguistica 9,
129–56.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 47

Kaye, Jonathan
1988 The phonologist’s dilemma: a game-theoretic approach to phono-
logical debate. GLOW Newsletter 21.
<http://www.meertens.knaw.nl/glow2002/kaye.pdf>
[Barrett-]Keach, Camillia
1986 Word-internal evidence from Swahili for Aux/Infl. Linguistic Inquiry
17, 559–64.
Keenan, Edward, and Edward Stabler
2003 Bare Grammar. Stanford, Calif.: CSLI.
Kenesei, István
2005 Focus as identification. In The Architecture of Focus, Valéria
Molnaar and Susanne Winkler (eds.), Berlin: Mouton.
<http://www.nytud.hu/oszt/igazg/kenesei/publ/keneseifocusmay.pdf
Kimenyi, Alexandre
2002 A Tonal Grammar of Kinyarwanda. Lewiston, New York: Edwin
Mellen Press.
Kinyalolo, Kasangati
2003a Limiting the scope of V-movement in KiLega (D25). Presented at
Niger-Congo workshop, Rutgers University, 14 March.
2003b The case for null subject-verb agreement morphology in Bantu.
NELS 34, SUNY Stony Brook.
<http://www.linguistics.stonybrook.edu/events/nels/abstracts/kinyalo
lo.pdf>.
Kisseberth, Charles W.
2002 Prosody and the phonology-syntax interface: Chimwiini revisited.
M.I.T. Linguistics Colloquium, 15 February.
Kisseberth, Charles W., and Sheila Mmusi
1990 The tonology of the object prefix in Setswana. Studies in the Linguis-
tic Sciences 20, 151–61.
Kornai, András, and László Kálmán
1988 Hungarian sentence intonation. In Autosegmental Studies on Pitch
Accent, Harry van der Hulst and Norval Smith (eds.), 183–95.
Dordrecht: Foris.
Krifka, Manfred
1991 A compositional semantics for multiple focus constructions. SALT 1,
127–58.
Ladd, D. Robert
1996 Intonational Phonology. Cambridge University Press.
Laka, Itziar
1990 Negation in syntax; on the nature of functional categories & projec-
tions. Dissertation, M.I.T., Cambridge, Mass.
48 Victor Manfredi

Lapnipran, Y.
1992 Intonation in tone languages; the phonetic implementation of tones
in Yoruzbap. Dissertation, Cornell University, Ithaca, N.Y.
LaVelle, Carl
1974 An experimental study of Yoruzbap tone. Studies in African Linguistics
Supplement 5, 185–204.
Leben, William R.
1973 Suprasegmental phonology. Dissertation, M.I.T., Cambridge, Mass.
1976 The tones in English intonation. Linguistic Analysis 2, 69–107.
1982 Metrical or autosegmental. In The Structure of Phonological Repre-
sentations Part 1, Harry van der Hulst and Norval Smith (eds.), 177–
90. Dordrecht: Foris.
Li, Charles N., and Sandra A. Thompson
1976 Subject and topic; a new typology of language. In Subject and Topic,
Charles N. Li. (ed.), 457–85. New York: Academic Press.
Liberman, Mark Y.
1975 The intonational system of English. Dissertation, M.I.T., Cambridge,
Mass.
1995 The sound structure of Mawu words. In Invitation to Cognitive Sci-
ence, Lila R. Gleitman and Mark Y. Liberman (eds.), 55–86. Cam-
bridge, Mass.: MIT Press.
Liberman, Mark Y., J. Michael Schultz, Soonhyun Hong, and Vincent Okeke
1993 The phonetic interpretation of tone in Izgbo. Phonetica 50, 147–60.
Manfredi, Victor
1987 Anti-logophoricity as domain extension in Izgbo and Yoruzbap. Niger-
Congo Syntax & Semantics 1, 97–113. African Studies Center, Bos-
ton University.
1993 Spreading and downstep; prosodic government in tone languages. In
The Phonology of Tone; the Representation of Tonal Register, Harry
van der Hulst and Keith Snider (eds.), 133–84. Berlin: De Gruyter.
1995 Tonally branching s in Yoruzbap is [LH]. Niger-Congo Syntax & Se-
mantics 6, 171–82. African Studies Center, Boston University.
2003 A fonosyntactic parameter within Benue-Kwa and its consequences
for EBzdop. In Typologie des langues d'Afrique et universaux de la
grammaire, vol. 2: Benue-Kwa, Soninke, Wolof, Patrick Sauzet and
Anne Zribi-Hertz (eds.), 127–62. Paris: Presses Universitaires de
Vincennes/ Éditions de l’Harmattan.
2004a A toneless theory of 2-and-a-half tonemes in Gbez. UParis-7, 1 De-
cember.
2004b The prosodic infrastructure of focus in Benue-Kwa. Universiteit van
Amsterdam, 4 December.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 49

2005a Conjoint/disjoint in western Benue-Kwa. Humboldt-Universität

Berlin, 7 October; Leiden University Centre for Linguistics, 18 Oc-
tober.
2005b Aspect or the serialization parameter. Institute for African Studies,
Universität Leipzig. 12 October.
2006 Nuclear stress in western Benue-Kwa. Dipartimento Scienze del
Linguaggio, Università di Venezia Ca’ Foscari, 13 June; Workshop
on Minority Languages, Università di Roma-3, 17 June; submitted
to ZAS Working Papers in Linguistics, Berlin.
Manzini, Maria Rita, and Leonardo M. Savoia
1998 Clitics and auxiliary choice in Italian dialects; their relevance for the
person-ergativity split. Recherches Linguistiques de Vincennes 27,
115–37.
McCawley, James
1970 Some tonal systems that come close to being pitch accent systems
but don’t quite make it. Chicago Linguistic Society 6, 526–32.
1978 What is a tone language? In Tone; a linguistic survey, Victoria
Fromkin (ed.), 113–31. New York: Academic Press.
McGreal, Chris
2006 Brothers in arms; Israel’s secret pact with Pretoria. The Guardian
[U.K.], 7 February. <http://www.guardian.co.uk/israel/Story/
0,,1704037,00.html>.
Meeussen, Achille E.
1959 Éssai de grammaire rundi. Musée Royal du Congo Belge, Tervuren.
1963 Morphotonology of the Tonga verb. Journal of African Languages 2,
72–92.
1967 Bantu grammatical reconstructions. Africana linguistica 3, 80–122.
Meier, Paul E., Ingeborg Meier, and John T. Bendor-Samuel
1975 A Grammar of Izzipiz, an Izgbo Language. Summer Institute of Linguis-
tics/ Wycliffe Bible Translators, Normal, Oklahoma.
Myers, Scott
1998 AUX in Bantu morphology and phonology. In Theoretical Aspects of
Bantu Tone, Larry M. Hyman and Charles W. Kisseberth (eds.),
231–64. Stanford, Calif.: CSLI.
Ndayiragije, Juvénal
1998 TP-internal focus in Kirundi and “Attract-F”. In Proceedings of
Workshop on Focus, Elena Benedicto, Maribel Romero, and Satoshi
Tomioka (eds.), 175–90. [= University of Massachusetts Occasional
Papers in Linguistics 21.] Amherst, Mass.: G.L.S.A.
50 Victor Manfredi

Nurse, Derek
2003 Aspect and tense in Bantu languages. In The Bantu languages, Derek
Nurse and Gerard Philippson (eds.), 90–102. London: Routledge.
Nurse, Derek, and Gerard Philippson
2003 Introduction. In The Bantu languages, Derek Nurse and Gerard
Philippson (eds.), 1–12. London: Routledge.
Odden, David
1981 Problems in tone assignment in Shona. Dissertation, University of
Illinois, Urbana-Champaign. [Not personally consulted; cited by
Sietsema 1989.]
1984 Formal correlates of focusing in Kimatuumbi. Studies in African
Linguistics 15, 275–99.
1988 Predictable tone systems in Bantu. In Autosegmental Studies on
Pitch Accent Systems, Harry van der Hulst and Norval Smith (eds.),
225–51. Dordrecht: Foris.
1996 The Phonology & Morphology of Kimatuumbi. Oxford University
Press.
1999 Typological issues in tone and stress in Bantu. In Cross-linguistic
Studies of Tonal Phenomena, Tonogenesis, Typology & Related Top-
ics, Shigeki Kaji (ed.), 187–215. Institute for the Study of Languages
& Cultures of Asia & Africa, Tokyo University of Foreign Studies.
2003 Rufiji-Ruvuma (N10, P10–20). In The Bantu languages, Derek
Nurse and Gerard Philippson (eds.), 529–45. London: Routledge.
Pierrehumbert, Janet
1980 The phonetics & phonology of English intonation. Dissertation,
M.I.T., Cambridge, Mass.
Pike, Eunice V., and Kent Wistrand
1974 Step-up terrace tone in Acatlan Mixtec. In Advances in Tagmemics,
Ruth Brend (ed.), 83–104. Amsterdam: North-Holland.
Pike, Kenneth L.
1945/47 The intonation of American English. Washingon, D.C.: Foreign
Service Institute/ Ann Arbor: University of Michigan Press.
1948 Tone languages; a technique for determining the number & type of
pitch contrasts in a language, with studies in tonemic substitution &
fusion. Ann Arbor: University of Michigan Press.
1965 On the grammar of intonation. In Proceedings of the 5th Interna-
tional Congress of Phonetic Sciences (Münster), Eberhard Zwirner
and Wolfgang Bethge (eds.), 105–19. Basel: Karger.
Piñón, Christopher
1992 Heads in the focus field. In Approaches to Hungarian 4; the Struc-
ture of Hungarian, István Kenesei and Csaba Pléh (eds.), 99–121.
Szeged: JATE.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 51

Potisuk, Sirpong, Jack Gandour, and Mary P. Harper

1996 Acoustic correlates of stress in Thai. Phonetica 53, 200–20.
Pulleyblank, Douglas
1982 Accent in Kimatuumbi. In Current Approaches to African Linguis-
tics 2, Jonathan Kaye, Hilda Koopman, Dominique Sportiche, and
André Dugas (eds.), 195–215. Dordrecht: Foris.
1983 Tone in lexical phonology. Dissertation, M.I.T., Cambridge, Mass.
Purnell, Thomas C. Purnell
1998 Principles & parameters of phonological rules; evidence from tone
languages. Dissertation, University of Delaware.
Reinhart, Tanya
1997 Interface economy; focus and markedness. Studia Grammatica 40,
146–69.
Rizzi, Luigi
1997 The fine structure of the left periphery. In Elements of Grammar:
Handbook in generative syntax, Liliane Haegeman (ed.), 281–337.
Dordrecht: Kluwer.
Rizzi, Luigi, and Leonardo M. Savoia
1992 Conditions on /u/ propagation in southern Italian dialects; a locality
parameter for fonosyntactic processes. In Syntactic Theory & the
Dialects of Italy, Adriana Belletti (ed.), 252–318. Torino: Rosenberg
& Sellier.
Rooth, Mats
1985 Association with focus. Dissertation, UMass, Amherst.
1996 On the interface principles for intonation. SALT 6, 202–226.
Rouget, Gilbert
1965 Analyse des tons du guzn par le «détecteur de mélodie» de l’institut
de Phonétique de Grenoble; rapport d’expériences. Langage &
comportement 1, 31–48/ 50, 3–29.
Russell, Kevin
1999 What’s with all these long words anyway? MIT Occasional Papers
in Linguistics 17, 119–130.
Schadeberg, Thilo C.
1986 Tone cases in Umbundu. Africana linguistica 10, 423–47.
2004 Conjoint & disjoint; grammaticalized information structure in Bantu.
Universiteit van Amsterdam, 3 December.
Seidl, Amanda Hallie
2001 An Optimality-Theoretic account of applicative asymmetries and
their phonological domains. ApplFest, Dept. of Linguistics, M.I.T.,
Cambridge, Mass, 28 January.
52 Victor Manfredi

Selkirk, Elisabeth
1995 Sentence prosody; intonation, stress and phrasing. In The Handbook
of Phonological Theory, John Goldsmith (ed.), 550–69. Oxford:
Blackwell.
Sharman, John C., and Achille E. Meeussen
1955 The representation of structural tones, with special reference to the
tonal behavior of the verb, in Bemba, Northern Rhodesia. Africa 25,
393–404.
Sietsema, Brian M.
1989 Metrical dependencies in tone assignment. Dissertation, M.I.T.,
Cambridge, Mass.
Stahlke, Herbert
1971 The noun prefix in Ez8ez. Studies in African Linguistics Supplement 2,
141–59.
Steedman, Mark
2000 Information structure and the syntax-phonology interface. Linguistic
Inquiry 34, 649–89.
Stevick, Earl W.
1969 Tone in Bantu. International Journal of American Linguistics 35,
330–41.
Stewart, John M.
1976 Towards Volta-Congo reconstruction. Leiden: University Press.
1994 The comparative phonology of Gbå and its significance for that of
Kwa and Volta-Congo. Journal of African Languages & Linguistics
15, 175–93.
Stucky, Susan
1979a Focus of contrast aspects in Makua; syntactic and semantic evidence.
Berkeley Linguistic Society 5, 362–72.
1979b The interaction of tone and focus in Makua. Journal of African Lan-
guages & Linguistics 1, 189–98.
Surányi, Balazs
2004 The left periphery and cyclic spellout; the case of Hungarian. In
Peripheries; syntactic edges & their effects, David Adger, Cecile de
Cat, and George Tsoulash (eds.), 49–73, refs. 407–33. Dordrecht:
Kluwer.
Swift, Lloyd B., A. AzhapghiBpotuz, and E. UzgopBjiBp
1962 Izgbo Basic Course. Washington, D.C.: Foreign Service Institute.
Szendröi, Kriszta
2003 A stress-based approach to the syntax of Hungarian focus. Linguistic
Review 20, 37–78.
Trager, George, and Henry Lee Smith
1949 An Outline of English Structure. Washington, D.C.: Foreign Service
Institute.
 Nuclear stress in eastern Benue-Kwa (Niger-Congo) 53

Truckenbrodt, Hubert
1995 Phonological phrases; their relation to syntax, focus & prominence.
Dissertation, M.I.T., Cambridge, Mass.
Vallduví, Enric
1990 The informational component. Dissertation, University of Pennsyl-
vania, Philadelphia.
Vallduví, Enric, and Elisabet Engdahl
1996 The linguistic realization of information packaging. Linguistics 34,
459–519.
Voorhoeve, Jan
1973 Safwa as a restricted tone system. Studies in African Linguistics 4,
1–22.
Voorhoeve, Jan, Achille E. Meeussen, and K. de Blois.
1969 New proposals for the description of the Izgbo completive phrase.
Journal of West African Languages 6, 79–84.
Wagner, Michael
2005 Prosody and recursion. Dissertation, M.I.T., Cambridge, Mass.
Wal, Jenneke van der
2005 Prosody and focus in Setswana. Ms., Leiden University.
Watters, John R.
1979 Focus in Aghem: a study of its formal correlates and typology. In
Aghem grammatical structure, Larry M. Hyman (ed.), 137–97. [=
Southern California Occasional Publications in Linguistics 7]. Los
Angeles: Department of Linguistics, University of Southern Califor-
nia.
Welmers, William E.
1959 Tonemics, morphotonemics and tonal morphemes. General Linguis-
tics 4, 1–9.
1973a Functional and vestigial noun class systems. African Language
Structures, 184–210. Berkeley: University of California Press.
1973b Verbal constructions in Niger–Congo. African Language Structures,
343–83. Berkeley: University of California Press.
Williams, Edwin
1971/76 Underlying tone in Margi and Izgbo. Linguistic Inquiry 7, 462–84.
1989 The clause as a coordinate structure. Ms., University of Massachu-
setts, Amherst.
1997 Blocking and anaphora. Linguistic Inquiry 28, 577–628.
Williamson, Kay
1989 Niger-Congo overview/ Benue-Congo overview. In The Niger-
Congo Languages, John T. Bendor-Samuel and Rhonda L. Hartell
(eds.), 3–45, 247–74. Lanham, Md.: American Universities Press.
54 Victor Manfredi

Williamson, Kay, and Roger Blench

2000 Niger-Congo. In African languages: an introduction, Bernd Heine
and Derek Nurse (eds.), 11–42. Cambridge: Cambridge University
Press.
Xu, Liejiong
2004 Manifestation of informational focus. Lingua 114, 277–99.
Xu, Yi, and Ching X. Xu
2005 Phonetic realization of focus in English Declarative intonation.
Journal of Phonetics 33, 159–97.
Zerbian, Sabine
2005 Why a prosodic approach to focus in Northern Sotho does not work.
Bantu Network Meeting, S.O.A.S., London, 13 May.
Zubizarreta, Maria Luisa
1998 Prosody, Focus and Word Order. Cambridge, Mass.: MIT Press.
Investigating prosodic focus marking in Northern
Sotho

Sabine Zerbian

Abstract

The paper contributes to the ongoing debate on the expression of information struc-
ture in the languages of the world by investigating prosodic focus marking in
Northern Sotho, a Bantu language spoken in South Africa. It presents the results of
a production and perception study that shows the absence of prosodic focus mark-
ing in this tone language. It thereby challenges both the universality of the promi-
nent status of focused constituents as well as the understanding of re-phrasing as a
universal cue to focus. However, Northern Sotho is not unique in this respect. The
finding therefore supports further research on a context-dependent theory of focus
interpretation.

1. Introduction

Over the last two decades, the expression of information structure in

grammar has become a prominent topic in linguistic research. This paper
contributes to the ongoing debate by investigating prosodic focus marking
in Northern Sotho, a Bantu language spoken in the northern part of the
Republic of South Africa, thereby deepening the understanding of whether
and how the category of focus is represented and marked in the languages
of the world.
Although the Bantu languages form the largest language family in Af-
rica, consisting of approximately 300 languages (Nurse & Philippson 2003:
2f), comparatively little research has been carried out on the prosodic ex-
pression of the information-structural categories topic and focus. Excep-
tions are Kanerva (1990), Jokweni (1995), Hyman (1990), and Downing et
al. (2004). This paper joins these studies by providing data on the prosodic
expression of in-situ focus in one further Bantu language. It applies ex-
perimental methodology that has been tested on Indo-European languages
and thereby makes the data comparable cross-linguistically.
56 Sabine Zerbian

The production and perception studies carried out show the absence of
prosodic marking in the verbal and postverbal domain in Northern Sotho.
The lack of grammatical marking of new information focus therefore chal-
lenges existing theories of focus that predict focus to be marked by pro-
sodic prominence. More specifically with respect to focus prosody, the
results reported here show that neither culminative stress prominence (Sel-
kirk 2004) nor (re-) phrasing (Ladd 1996, Hayes & Lahiri 1991) are univer-
sal cues to focus. At the same time, the results underline the importance of
typological research, in this case of African languages, for the development
and improvement of linguistic theories.
By investigating a tone language, this paper also contributes to the re-
search of the interaction of (lexical) tone and intonation. Whereas non-tonal
languages (e.g. Germanic languages like English and German) typically use
pitch accent for focus marking, it is reported that tone languages (for Bantu,
see Bearth 2003) often use syntactic transformations for the same purpose.
The role of prosody in African tone languages remains under-investigated
(Bearth 1999, Creissels 1996). Contrary to prior views and in accordance
with work on non-African tone languages, research on the Bantu languages
Chichewa, Kinande, and Xhosa has given evidence for the fact that tone
languages do make use of prosodic means to indicate focus. In showing the
absence of prosodic focus marking in Northern Sotho, this paper reveals
language-specific variation among Bantu languages and thereby encourages
further research into the microvariation of Bantu prosody.
The paper is organized as follows. Section 2 discusses phrasal phonol-
ogy in the Bantu language Chichewa in order to illustrate the kind of su-
prasegmental parameters that can be influenced by focus in a tone lan-
guage. It relates to Northern Sotho in presenting the kind of prosodic
changes to be expected under focus in this language. In section 3, the pro-
duction study is presented that has been carried out in order to investigate
the prosodic expression of in-situ focus in Northern Sotho. Section 4 deals
with the perception study. Finally, section 5 summarizes the results of the
studies and discusses their implications.

2. Prosody in Bantu tone languages

Contrary to prior views (Cruttenden 1986), research on a range of different

tone languages has shown that these languages use suprasegmental means
in order to indicate focus. Among the suprasegmental means used in tone
 Investigating prosodic focus marking in Northern Sotho 57

languages are the widening of the overall pitch range (Xu 1999 for Chi-
nese), the insertion of boundary tones (Hyman 1990 for Kinande), the in-
sertion of phonological phrase boundaries with resulting tonal effects (Kan-
erva 1990 for Chichewa; Jokweni 1995 for Xhosa), and the manipulation of
downdrift (Downing et al. 2004 for Chichewa).
Section 2.1 presents the prosodic expression of focus in Chichewa
(Kanerva 1990, Downing et al. 2004), thereby illustrating how focus can be
indicated in a Bantu tone language. Section 2.2 illustrates further use of
suprasegmental means for conveying discourse-pragmatic meaning. Section
2.3 relates to the prosody of Northern Sotho. Section 2.4 summarizes and
formulates the research question that will be investigated in section 3.

2.1. Influence of focus on phrasal phonology in Bantu

Bantu languages have two-tone systems in which high tones (H) are speci-
fied underlyingly and low tones (L) are inserted by default (Kisseberth &
Odden 2003). Following common practice, in the examples cited here, high
tones are marked by acute and underlying high tones are underlined. Sur-
face realizations of tone patterns are determined by principles governing
the realization of underlying tones and the inventory and application of
tonal processes as well as by the syntactic position of a word. The setting of
these parameters is language-specific.
Research on Chichewa, a Bantu language spoken in Malawi, has shown
that, besides syntactic structure, it is also focus that influences the surface
tone pattern of a word (Kanerva 1990, Downing et al. 2004). Focus is
shown to insert phonological phrase boundaries that determine the applica-
tion of tonal processes and thereby result in surface tone sandhi. Among the
cues to phonological phrase boundaries in Chichewa are the lack of high
tones on phrase-final vowels and lengthened vowels in the penultimate
syllable of phrase-final words (See Kanerva 1990 for a detailed discussion
of phonological processes that motivate phonological phrases in
Chichewa). Phonological phrase boundaries are indicated by parentheses,
focus structure is indicated by square brackets.

Chichewa (Kanerva 1990: 98)

(1) a. Q: ‘What did he do?’
A: ([anaményá nyumbá ndí mwáála]F)PP
he hit house with stone
‘He hit the house with a stone.’
58 Sabine Zerbian

b. Q: ‘What did he hit the house with?’

A: (anaményá nyumbá [ndí mwáála]F)PP

c. Q: ‘What did he hit with the rock?’

A: (anaményá [nyuúmba]F)PP (ndí mwáála)PP

d. Q: ‘What did he do to the house with the rock?’

A: ([anaméenya]F)PP (nyuúmba)PP (ndí mwáála)PP

Under wide focus, (1a), the entire VP forms a single phonological phrase.
The same phrasing is observed for narrow focus on the sentence-final
prepositional phrase, as in (1b). Narrowing the focus within the sentence, as
to the object in (1c) and to the verb in (1d), results in additional phonologi-
cal phrase boundaries. The focused constituent is followed by a phonologi-
cal phrase boundary.
For further details of the analysis concerning the phrasing of post-focal
constituents, see Kanerva (1990) and Truckenbrodt (1995, 1999). For com-
parable expression of focus by suprasegmental means in Xhosa, see Jok-
weni (1995) and a reanalysis in Zerbian (2004).

2.2. Intonation in Bantu

Apart from focus, suprasegmental means also convey further discourse-

pragmatic meaning in Bantu languages, such as sentence types. Again,
there is microvariation concerning the phonetic implementation (see, e.g.,
Downing 1996 for the Bantu language Jita).
In Northern Sotho and Southern Bantu languages more generally, raised
overall pitch and lack of length on the penultimate syllable distinguish de-
claratives from yes/no-questions (Poulos & Louwrens 1994: 374, Jones et
al. 2001 in a detailed phonetic study on Xhosa). Independent of the lexical
tone patterns of the words used in an utterance, the yes/no-question dis-
plays a raised overall pitch, which is indicated by the arrow in (2). Fur-
thermore, the penultimate syllable of the sentence-final word is not length-
ened in yes/no-questions.

(2) a.ÇO dumediUhC dikgarebe ?

2nd greet ladies
‘Are you greeting the ladies?’
 Investigating prosodic focus marking in Northern Sotho 59

b. Ee, ke dumedKUhC dikgare:be.

yes 1st greet ladies
‘Yes, I am greeting the ladies.’

Also, a non-lexical high tone on the ultimate syllable of a clause indicates

continuation, as shown in (3).

(3) Kgarebe é búla lefasetere H% mo-kgalabje ó tswálela lematí.

CL9.lady SM9 open window CL1-old.man SM1 close door
‘The lady opens the window, the old man closes the door.’

The examples in (2) and (3) show that Northern Sotho and Bantu languages
in general make use of intonation for discourse-pragmatic meaning in ac-
cordance with cross-linguistic tendencies.

2.3. Phrasal phonology in Northern Sotho

Just as in Chichewa, Northern Sotho shows suprasegmental and tonal proc-

esses that are sensitive to prosodic domains. The two most prevalent
phrasal processes are penultimate lengthening and the application of the
finality restriction at prosodic boundaries.
Vowel length in Northern Sotho is not contrastive. It can therefore be
exploited to convey grammatical meaning. Like in Chichewa, in Northern
Sotho, the penultimate syllable of a word in phrase-final position is regu-
larly lengthened (Doke 1954: 125). Length is indicated by a colon in the
examples to follow.
The rule of high tone spread (HTS), which is common in Bantu lan-
guages, accounts for the spreading of an underlying high tone to the imme-
diately right-adjacent syllable (see also Zerbian 2006b). This rule applies
within a word, as shown in the verbs in (4a). If, however, the word appears
in phrase-final position, spreading to a phrase-final syllable is blocked.
Such a restriction that exempts a final syllable from prosodic processes is
called finality restriction and is common across Bantu (Kisseberth & Odden
2003: 64). The tonal effect of the finality restriction is exemplified in (4b).

Northern Sotho
(4) a. Ke thúshá mosá:di. b. Ke a thú:UhC
1st help woman 1st A help
60 Sabine Zerbian

Ke Uhómá polase:-ng. Ke a Uhó:ma.

1st work farm-at 1st A work
Ke bóná gagó:lo. Ke a bó:na.
1st see well 1st A see

The blocking of HTS is therefore an indicator of prosodic domain edges.

The prosodic domain for lengthening and the application of the finality
restriction in Northern Sotho is smaller than the entire clause. In (5), length
and tone is manipulated on the verb which is in non-final position.

(5) Ke a mo thú:UhC mo-sá:di.

st
1 A CL1 help CL1-woman
‘I am helping her, the woman.’

Prosodic structure therefore depends on syntactic structure in Northern

Sotho. Zerbian (2006a) argues that prosodic boundaries occur at the right
edges of syntactic IPs in this language. In the study reported in the next
section, it is tested if the prosodic domain which penultimate lengthening
and the finality restriction give evidence for is sensitive to focus.

2.4. Summary

Focus-induced variation in phrasing in some Bantu languages and the use

of intonational changes for conveying discourse-pragmatic meaning in
Northern Sotho motivate the hypothesis that suprasegmental means also
play a role in postverbal in-situ focus in Northern Sotho. At the same time,
language-specific variation among Bantu languages with respect to the
parameters that determine and influence surface tone realization justifies
research on the focus prosody of further Bantu languages.
The next section presents a production and perception study on the pro-
sodic marking of in-situ focus in Northern Sotho.

3. Production study

This section presents the production study that was carried out in order to
investigate the means of prosodic marking of in-situ focus in Northern So-
tho. 1 The expectation based on comparative studies is that prosodic mark-
ing of focus in Northern Sotho results in perceptible suprasegmental
 Investigating prosodic focus marking in Northern Sotho 61

changes, either concerning the realization of underlying tone patterns, reali-

zation of length on the penultimate syllable (as discussed in section 2.1 and
2.3), or overall register height (see section 2.2). For more details on the
studies, the reader is referred to Zerbian (2006a: 149ff).

3.1. Production set up

In order to establish if there are prosodic differences between sentences

displaying narrow focus on the verb phrase, the verb, the object, or the
adverb, a production study was carried out with native speakers of Northern
Sotho. Sentence pairs with different focus structures were recorded. The
individual pairs consist of a trigger sentence and a target sentence. The
trigger sentence is a question that determines the focus structure of the cor-
responding answer, the target sentence. Thereby, this study follows the
underlying assumption, first used in Paul (1880) and applied in experimen-
tal set ups by Gussenhoven (1983), Oppenrieder (1989), Birch & Clifton
(1995), and Xu (1999) among others, that the focus of an utterance is con-
trolled by a preceding question. An example for the question/answer-pairs
is given in (6).

(6) a. What do you do?

Ke [mémá mo-hú:mi]F.
1st invite CL1-rich
‘I [invite the rich man]F.’
b. Who do you invite?
Ke mémá [mohú:mi]F.
‘I invite [the rich man]F.’
c. What are you doing with the rich man?
Ke [mémá]F mohú:mi.
‘I [invite]F the rich man.’

The current study tried to prevent reading intonation (Gussenhoven 1983,

referring back to Brazil et al. 1980: 83) by having two native speakers in-
teract in posing and answering questions which have been provided on
cards. The subjects were instructed to answer in a full sentence repeating all
constituents. Although this is not the most natural answer (discourse-old
constituents are often reduced or deleted), this methodology is justified by
the need for a set-up in which focus is controlled for.
62 Sabine Zerbian

The production study consisted of 69 sentences altogether, having five

different syntactic target structures: intransitive verbs followed by an ad-
verb, and transitive and ditransitive structures both with and without ad-
verb. Examples are given in (7).

(7) a. Intransitive sentences with adverb

Ke bíná monyányé-:ng.
1st dance party-LOC
‘I am dancing at the party.’

b. Transitive sentences
Ke mémá mo-hú:mi.
1st invite CL1-rich
‘I invite the rich man.’

c. Transitive sentences with adverb

Ke rémá morúlá moséga:ré.
1st chop marula tree midday
‘I chop the marula tree at midday.’

d. Ditransitive sentences
Ke fá mmá hém:pe.
1st give mother shirt
‘I give mother a shirt.’

e. Ditransitive sentences with adverb

Ke néá malómé lengwáló bo-dúló:-ng.
st
1 give uncle letter CL14-live-LOC
‘I give uncle a letter at the dwelling place.’

Narrow focus on each constituent of these sentences plus wide focus

yielded 69 target sentences, displaying different focus structures. Eight
native speakers of Sotho languages participated in the study. The five
speakers who provided the target sentences were all speakers of the Sepedi
dialect of Northern Sotho.
 Investigating prosodic focus marking in Northern Sotho 63

3.2. The target sentences

The target sentences were chosen in such a way that a maximal supraseg-
mental contrast, both in length and tone, could have been observed if focus
induced boundaries in Northern Sotho had indicated differences in informa-
tion structure (see section 2.3). This hypothesis is illustrated in (8).

(8) a. VP and object focus

Ke mémá mohú:mi)PP.
‘I invite [the rich man]F.’
b. Verb focus
Ke mé:ma)IP mohú:mi)PP.
‘I [invite]F the rich man.’

In (8a), the verb stem initial high tone spreads onto the immediately right-
adjacent vowel under wide focus and object focus. If focus induced a
phrase boundary, the application of HTS would be blocked because the
finality restriction would apply, as shown in (8b). Also, the syllables that
undergo penultimate lengthening are expected to differ under different fo-
cus structures, as is also shown in (8). 2

3.3. Quantitative analysis

A sample of the recorded data, viz. the basic SVO/SVAdv structure, was
chosen for the analysis. The reason for the decision in favor of basic
SVO/SVAdv structures, as in (7a,b), as compared to structures in which
two or more constituents follow the verb, as in (7c-e), is that if there is pro-
sodic expression of focus, it will most clearly emerge within short sen-
tences. In short sentences, the register size at disposal for pitch manipula-
tions towards the end of the sentence is wider than in long sentences, due to
downdrift. The sentences are given in (9).

(9) Sentences of prosodic analysis

a. i. Ke mémá mo-hú:mi.
1st invite CL1-rich
‘I invite the rich man.’
ii. Ke rwálá mo-rwá:lo.
1st carry CL3-load
‘I am carrying a load.’
64 Sabine Zerbian

b. i. Ke bíná mo-nyányé:-ng.
1st dance CL3-party-LOC
‘I am dancing at the party.’
ii. Ke bóá mo-lálé:-ng.
1st return CL3-field-LOC
‘I am returning from the field.’

Although the sets are different with respect to the grammatical function of
the postverbal constituent, they are identical in their phrasal and tonal
make-up. It has been shown in Zerbian (2006a: 122ff) that objects and ad-
verbs are phrased parallel in Northern Sotho. This insight is exploited in the
analysis. Although the sentences in (9a) differ from those in (9b) in the
absolute number of syllables (the locative suffix -ng in (9b) is syllabic), this
difference is not crucial as the syllables of interest are the two syllables of
the verb and the following nominal prefix.
Moreover, all sentences in (9) contain a high-toned disyllabic verb
which is followed by a constituent that has a low-toned nominal prefix but
bears a high tone on the noun stem initial syllable. This tonal environment
allows HTS to occur on the verb and has been chosen because it has been
shown that high tones are indeed manipulated at IP-boundaries in Northern
Sotho (see section 2.3).
The sounds of these utterances were segmented according to common
phonetic practice (see e.g. Ladefoged 2001, 2003). Two phonetic parame-
ters were measured: pitch height and duration. 3 These two parameters are
phonetic correlates of stress in intonation languages, and also the tonal and
durational changes expected under phrasing differences in Bantu tone lan-
guages are captured by these parameters. Changes in the tonal structure due
to the application of the finality restriction and the resulting blocking of
HTS will result in differences in F0. Changes in penultimate lengthening
due to the insertion of prosodic boundaries after focused constituents will
result in longer duration of vowels.

3.4. Results – Fundamental frequency

In the following diagrams, the F0-average values for each of the five
speakers are represented in a plot. The three horizontal lines correspond to
the three different focus conditions: VP focus, object focus, and verb focus.
The values indicated in the diagrams are average values of the respective
segment over the four utterances indicated in (9). 4
 Investigating prosodic focus marking in Northern Sotho 65

On the x-axis, the vowels of the different constituents in the utterance

are displayed. The second and third value correspond to the verb, the fourth
till final value correspond to the object or locative adverbial. The first value
refers to the vowel of the personal pronoun (ke). On the y-axis, the average
F0-value is given in Hertz. Four speakers are female. The fifth speaker is
male wherefore the lines occur in a lower pitch range. The dashed line
shows VP-focus, the continuous line represents focus on the verb, and the
dotted line indicates focus on the object or local adverbial.

Diagrams 1–5: Average F0

Diagram 1: Speaker DE

300
250
Obj/Adv-focus
F0 (Hz)

200
150 V-focus
100 VP-focus
50
0
SM

dv

dv

dv
V

v
Ad

/A
/A

/A
bj

bj

bj
O

Diagram 2: Speaker LY

250

200

150
F0 (Hz)

100

50

0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv
66 Sabine Zerbian

Diagram 3: Speaker MA

350
300
250
F0 (Hz)

200
150
100
50
0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv

Diagram 4: Speaker Mak

350
300
250
F0 (Hz)

200
150
100
50
0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv

Diagram 5: Speaker MO

180
160
140
120
F0 (Hz)

100
80
60
40
20
0

SM V V Obj/Adv Adv Obj/Adv Obj/Adv

 Investigating prosodic focus marking in Northern Sotho 67

The general tonal information, discussed in section 2.3, emerges from the
diagrams. The verb bears a high tone on the stem initial syllable underly-
ingly. The peak of F0 is only reached in the second syllable of the verb,
which is captured by the process of high tone spread (HTS). After the F0
peak in the second syllable of the verb, F0 drops again.
The noun stem initial syllable, i.e. the second syllable of the noun, is
transcribed in the examples as bearing a high tone. This is motivated by the
tonal realization the noun receives if it is pronounced in isolation. Within
the sentence, no F0 peak is visible. This is due to declination, which nar-
rows the pitch range available to speakers towards the end of the utterance
(see Gussenhoven 2004 for an overview of downtrends in intonation).
In the speech of LY, MA, Mak, and MO, we find a nearly perfect over-
lap in what concerns the F0 for all three focus conditions. F0 raises in the
verb due to HTS and falls smoothly towards the end of the sentence. Inter-
estingly, in the speech of DE a slightly higher F0 can be found for the ini-
tial syllable of the object when it is in focus. Also, a slightly raised pitch
and a lower initial syllable of the object can be found when the verb is in
focus, which results in a sharper fall in pitch. Although the differences are
only small, it is nevertheless important to follow up on these differences as
the second phenomenon, viz. the sharper fall after a focused verb, has been
reported as a focus strategy in Chichewa (Downing et al. 2004). By the
term ‘anti-accent’, Downing et al. (2004: 177) describe that in Chichewa
the focused element is “made prominent by raising the pitch enough to
make the following elements relatively much lower in pitch”.
However, what we do not find in Northern Sotho is a pitch accent on the
focused constituent. The diagram shows that F0 falls towards the end of the
sentence (due to declination), which is also the case if the object is the fo-
cused constituent. Judging from the pitch accent distribution in intonation-
only languages like English and German, if focus were expressed by pitch
accent in Northern Sotho as well, one would expect to find some reversal of
the trend in the condition of object focus.
In addition to the lack of evidence for the use of a pitch accent to indi-
cate focus, it is interesting to note that tone sandhi do not emerge as a con-
sequence of different focus conditions either. Against expectation, no
speaker shows a fall of F0 on the second syllable of the verb, as predicted if
focus inserted a phrase boundary that blocks the application of high tone
spread (HTS).
Also, a third option reported for tone languages, namely the raising of
the overall pitch register, is not mirrored in the diagrams 1–5.
68 Sabine Zerbian

The preliminary conclusion is that focus is indicated neither by pitch ac-

cent nor by tone sandhi in Northern Sotho. The perception study presented
in section 4 checks this preliminary conclusion, and also checks if the
slightly sharper fall after a focused verb in the speech of DE is interpreted
as a focus strategy, parallel to the ‘anti-accent’ described in Downing et al.
(2004) for Chichewa.

3.5. Results – duration

In diagrams 6–10, the duration of each vowel over the three different focus
conditions is given for all five subjects separately. The duration values
represent relative duration of the segments. Because of the uneven number
of syllables in (9a) and (9b), the absolute duration of the segment has been
divided by the absolute time of the whole utterance in order to account for
differences which are not only induced by utterance length but also by
speech tempo etc.
Again, on the x-axis the vowels of the different constituents of the utter-
ance are displayed. Each line represents the average values within one fo-
cus condition. On the y-axis, the relative duration of the segments is given
in seconds. The diagrams are discussed below.

Diagrams 6–10: Average duration

Diagram 6: Speaker DE
relative duration (sec)

0,25
0,2 Obj/ Adv-focus
0,15
V-focus
0,1
0,05 VP-focus
0
SM

v
v

v
V

v
d

d
Ad
/A

/A

/A
bj

bj

bj
O

O
 Investigating prosodic focus marking in Northern Sotho 69

Diagram 7: Speaker LY
relative duration (sec)

0,25
0,2
0,15
0,1
0,05
0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv

Diagram 8: Speaker MA

0,25
relative duration (sec)

0,2

0,15

0,1

0,05

0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv

Diagram 9: Speaker Mak

relative duration (sec)

0,25
0,2
0,15
0,1
0,05
0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv
70 Sabine Zerbian

Diagram 10: Speaker MO

0,25
relative duration (sec)

0,2
0,15
0,1

0,05
0
SM V V Obj/Adv Adv Obj/Adv Obj/Adv

A quite homogeneous picture emerges for all five speakers in all focus
conditions. The speech of DE, Mak, and MO again shows a nearly perfect
overlap with respect to duration in different focus conditions. The lengthen-
ing of the vowel in the penultimate syllable is clearly discernible in the
diagrams 6–10.
In the speech of LY and MA, we find a slightly lengthened verb final
syllable. However, the implication of this observation with respect to the
encoding of information structure is doubtful. First, the increase in duration
is not as expected in the penultimate syllable of the focused constituent and
also is not as high as the lengthened penultimate syllable of the sentence.
Second, the lengthening of the final syllable is not restricted to one focus
condition. It occurs with verb focus in the speech of MA but with object
focus in the speech of LY.
Again, the preliminary conclusion from a phonetic investigation of the
production data suggests no prosodic expression of focus through lengthen-
ing. However, the lengthening of the verb final syllable, although slight,
might be significant enough for native speakers to serve as an indicator for
the focused constituent in the utterance.

4. Perception study

The prosodic analysis revealed that across all speakers there are no system-
atic, significant differences between different focus structures if tested for
the common stress-related parameters pitch and duration. The objective of
the perception study is therefore to investigate if there are nevertheless
prosodic cues in the production data that enable native speakers of Northern
 Investigating prosodic focus marking in Northern Sotho 71

Sotho to identify instances of in-situ focus. These prosodic cues might be

very subtle so that they have escaped an acoustic analysis of the data, or
they might differ from the stress-related phonetic correlates that were in-
vestigated in section 3. Furthermore, it needs to be established if the exist-
ing suprasegemental differences seen in the speech of LY, MA, and DE are
indeed related to differences in focus structure of the utterances. Perception
tests for establishing distinct categories in focus structures have been car-
ried out by Gussenhoven (1983), Oppenrieder (1989), Jannedy (2002) a.o.
for focus in sentences and by Swerts et al. (2002) for focus in NPs.

4.1. Perception set up

The target sentences obtained by the recordings are used in the perception
study in order to test if prosodic differences in the non-identical pairs can
be related to information structure (study 1) and if there are perceivable
prosodic differences in the pairs that sound identical (study 2). The percep-
tion study asked for appropriateness judgments: Subjects were asked to
listen to an identical question followed by two differing answers in direct
comparison. The task is illustrated in (10).

(10) Question: ‘Which of these question-answer pairs matches best?’

a. O mémá má:ng?
‘Who do you invite?’
Ke mémá [mohú:mi]F.
‘I invite [the rich man]F.’
b. O mémá má:ng?
‘Who do you invite?’
Ke [mémá]F mohú:mi.
‘I [invite]F the rich man.’

As indicated in (10), one of the answers is the original answer to the ques-
tion. The competing answer is taken from a different focus-context. Sub-
jects had to decide between the first and the second question/answer-pair.
Study 1 contained utterances with perceptible suprasegmental differ-
ences. Four lexically different contexts were used, which are given in (11).
Study 2 contained utterances with no perceptible difference. Three lexically
different contexts were used, which are given in (12).
72 Sabine Zerbian

(11) Study 1
a. Ke néá náré lé:e. ‘I give the buffalo an egg.’
1st give buffalo egg
b. Ke néá molámó lemá:o. ‘I give the uncle a needle.’
1st give uncle needle
c. Ke nyálwá labo:ne. ‘I marry for the fourth time.’
1st marry four times
d. Ke rwálá morwá:lo. ‘I carry a load.’
1st carry load

(12) Study 2
a. Ke já namú:ne. ‘I eat an orange.’
1st eat orange
b. Ke rémá morúlá moséga:ré. ‘I chop the marula at midday.’
1st chop marula tree midday
c. Ke fá m´má hém:pe. ‘I give mother a shirt.’
1st give mother shirt

In the perception study, the focus structures of a target sentence were

checked against all possible constituent questions.
The expectation for this study is that the speaker chooses one of the two
alternative answers as more appropriate if there are subtle prosodic differ-
ences in the realization of various in-situ foci. The set up in terms of mini-
mal answer pairs was chosen in order to focus the listener’s attention on
possible subtle differences in the realization of different focus structures.
The appropriateness judgments made by the four subjects participating
in this study were recorded by the experiment software (Praat, Boersma &
Weenink 2005) and evaluated.

4.2. Results

The responses in experiment 1, in which the question/answer-pairs showed

a slight difference in prosody, were distributed according to chance both
within and across speakers. The same result was obtained for experiment 2,
in which the question/answer-pairs did not show prevalent perceptible dif-
ferences (for details see Zerbian 2006a). Therefore, one can conclude that
the existing prosodic differences are not related to the focus structure of an
utterance (study 1) and furthermore, that the utterances with no perceptible
 Investigating prosodic focus marking in Northern Sotho 73

differences do not contain subtle prosodic cues that could be related to fo-
cus structures (study 2).
A closer inspection of the answers to the perception study reveals that
suprasegmental differences influence the judgments independent of focus
contexts. To give an example: in one test item (DE_15), downstep occurs
after the focused verb. However, it is not considered a significant cue to
focus structure. Although all subjects perform exceptionally well when
choosing the downstepped answer as the inappropriate answer, they do not
choose the downstepped answer as the appropriate one in the verb focus
context. Also, ordering effects become visible in the responses (Ladd &
Morton 1997, Schiefer & Batliner 1991). Whereas one subject nearly al-
ways chooses the second answer, another nearly always chooses the first
answer.

5. Conclusion and discussion

The experimental part had as its objective to test if there is a prosodic ex-
pression of focus in Northern Sotho when the focused constituent is in-situ
and is followed by discourse-old material. Studies on Chichewa and Xhosa
report changes in the realization of the tonal pattern or lengthening of pe-
nultimate syllables. Northern Sotho tonology shows characteristics which
are comparable to Chichewa and Xhosa and which lead one to expect pro-
sodic changes under focus in Northern Sotho as well.
The production study tested if Northern Sotho native speakers produce
prosodic differences which are correlated to diverging focus structures.
Target sentences were prepared in such a way as to allow tonal as well as
durational changes. Question/answer-pairs that were controlled for different
focus conditions were recorded from native speakers. A phonetic analysis
of the data with respect to fundamental frequency and duration showed no
systematic prosodic expression of focus. A follow-up perception study
showed that the existing prosodic differences were not interpreted with
respect to focus. Neither did any other prosodic cue emerge as relevant for
the encoding of focus in Northern Sotho.
The conclusion drawn from this production/perception study is that in
syntactically identical sentences there is no suprasegmental marking of
focus in Northern Sotho. The study therefore showed that in Northern So-
tho, in contrast to Chichewa and Xhosa, focus does not affect phrasing.
Differences in production of diverging underlying information structures
74 Sabine Zerbian

can only be attributed to idiolects or differences in speech tempo, but they

cannot be regarded as language-inherent functional means of marking fo-
cus.
The production and perception study has shown that there are no pro-
sodic means available in the grammar of Northern Sotho for marking in-
situ focus in the verbal and postverbal domain. The absence of prosodic
marking has also been shown for other languages (e.g. Wolof in Rialland &
Robert 2001). However, Northern Sotho shows the simultaneous lack of
any other grammatical focus marking, such as morphological (as in Wolof)
or syntactic (see example (8) and also Zerbian (2006a)). Hartmann &
Zimmermann (2004, 2007) report similar findings for Tangale and Hausa,
two Chadic languages. In general, the findings in these languages question
the universal need for focus marking in grammar.
However, the conclusion does not imply that Northern Sotho lacks
grammatical means of indicating the focused constituent in the verbal and
postverbal domain. For example, Northern Sotho uses deletion, pronomi-
nalization, and dislocation of discourse-old constituents; processes which
conspire to the tendency that the focused constituent is often final in a
clause.
Neither does the lack of prosodic marking of postverbal in-situ focus
mean that Northern Sotho does not show any prosodic reflexes of informa-
tion structuring. For example, a right-dislocated discourse-old constituent is
set off from the sentence by an optional pause, preboundary lengthening,
and tone sandhi, see (5). However, the necessity of an object marker on the
verb when being dislocated shows that the prosodic changes are dependent
on the syntactic structure and not a means on their own.
As pointed out above, the absence of prosodic means with the simulta-
neous absence of syntactic marking in languages like Northern Sotho poses
a problem for standard theories of focus (from Jackendoff 1972 to
Schwarzschild 1999 and beyond) that generally assume that focus must be
marked somehow in the grammar, e.g. by pitch accent in languages like
English and German. Following Hartmann & Zimmermann (2007), who
discuss the theoretical implications of the lack of grammatical focus mark-
ing in Hausa, there are two ways to account for this lack.
First, one could argue that focus does not exist as a grammatical cate-
gory in the grammar of Northern Sotho, which is reflected by the lack of
formal F-features. However, syntactic processes of deletion, pronominali-
zation, and dislocation (also called indirect focus marking strategies by
Bearth 1999) make indirect reference to formal F-features as only non-F-
 Investigating prosodic focus marking in Northern Sotho 75

marked constituents are targeted by these processes. Furthermore, the sub-

ject cannot be F-marked in its canonical preverbal position (Zerbian 2006a,
to appear).
The alternative view is that Northern Sotho, just like Hausa, is among
the languages that do not require obligatory F-marking in grammar. This
characterization is reminiscent of the distinction between obligatory and
non-obligatory focus-marking systems in Heine & Reh (1983). This dis-
tinction was first proposed with respect to morphological F-marking but
seems to be applicable more widely also with respect to prosodic and syn-
tactic focus marking.
From the study of African languages such as Hausa (Chadic) and North-
ern Sotho (Bantu), evidence against the obligatoriness of marking focus in
grammar emerges. Therefore, existing theories of focus marking cannot lay
claim to universality, and future research has not only to bring further evi-
dence for such systems but also to accommodate the strategies of those
languages in appropriate focus theories (but see Manfredi, this volume, for
an alternative). A first proposal has been advanced in Hartmann &
Zimmermann (2007), drawing attention to topic marking and inter-
sentential marking as two alternative strategies of information structuring.

Notes

This work is part of the project ‘The expression of information structure in

languages of Southern Africa’ funded by the DFG. Further financial assis-
tance also came from a fellowship from the DAAD. I wish to express my
gratitude to Makopi Mogodi, Mogoke Mogodi, Simon Mashanya, Mam-
paka Lydia Mojapelo, Matseleng Mokhwesana, and Deborah Mampuru
who made this study possible in the first place. Furthermore, I wish to
thank Bernd Pompino-Marschall for practical advice, and Hubert Truck-
enbrodt, Laura J. Downing, and two anonymous reviewers for helpful
comments.
1. The investigation is restricted to in-situ focus in the verbal and postverbal
domain as focus on the subject cannot be expressed with having the sub-
ject in its canonical preverbal position in Northern Sotho (see Zerbian
2006a, to appear).
2. Furthermore, there are independent phonetic restrictions on the use of
segmental material in the target utterances to assure that changes in fun-
damental frequency are indeed related to extra-phonetic facts like dis-
course-pragmatics. On the one hand, obstruents interrupt and disturb the
F0-contour by their inherent phonetic properties. On the other hand, with
76 Sabine Zerbian

short vowels, the pitch reaches its peak only at the end of a following con-
sonant (Ladd et al. 2000 for Dutch). Because of the inherent phonetic
properties of obstruents as well as the lack of underlyingly long vowels in
Northern Sotho, the target sentences of the current study are constructed
in such a way that they contain sonorants only, which allows a theoreti-
cally undisturbed F0-contour.
3. Intensity is left out for two reasons: first, it is the parameter that indicates
stress in intonation-only languages least reliably. Second, it is the parame-
ter that is the most sensitive to changes in the recording surrounding, such
as distance to the microphone, background noises that might lead the sub-
ject to speak louder. As these parameters were not controlled for (flexible
distance to the microphone and no sound-proof box), it is expected that
deviation in intensity values might not alone be attributed to focus.
4. Fundamental frequency is influenced by the preceding sound. Sometimes,
no F0-value could be calculated for the vowel following the plosive of the
subject marker, as is the case in the speech of DE. Along the same lines, it
was sometimes impossible to calculate a F0-value for the final vowel of
an utterance. Because of the unequal number of syllables in (9a) and (9b),
the fifth value in the diagrams is only calculated for the examples in (9b)
and therefore represents the average of two utterances only. The penulti-
mate and last syllable of all examples have been calculated together. This
is especially important for the duration measurement, as the duration
measurement, as the penultima is the lengthened syllable.

References

Bearth, Thomas
1999 The contribution of African linguistics towards a general theory of
focus. Update and critical review. Journal of African Languages and
Linguistics 20: 121–156.
2003 Syntax. In The Bantu languages, Derek Nurse and Gerard Philippson
(eds.), 121–142. London: Routledge.
Birch, Stacy, and Charles Clifton
1995 Focus, Accent, and Argument Structure: Effects on Language Com-
prehension. Language and Speech 38: 365–391.
Boersma, Paul, and David Weenink
2005 Praat: doing phonetics by computer (Version 4.3.01) [Computer
program]. Retrieved Apr 23, 2004 from http://www.praat.org/.
Brazil, David, Malcolm Coulthard, and Catherine Johns
1980 Discourse intonation and language teaching. Essex: Longman.
 Investigating prosodic focus marking in Northern Sotho 77

Creissels, Denis
1996 Conjunctive and disjunctive verb forms in Setswana. South African
Journal of African Languages 16: 109–115.
Cruttenden, Alan
1986 Intonation. Cambridge: Cambridge University Press.
Doke, Clement Martyn
1954 The Southern Bantu Languages. London, New York, Cape Town:
Oxford University Press.
Downing, Laura J.
1996 The tonal phonology of Jita. München: Lincom Europa.
Downing, Laura J., Al Mtenje, and Bernd Pompino-Marschall
2004 Prosody and Information Structure in Chichewa. In Papers in Pho-
netics and Phonology, Susanne Fuchs and Silke Hamann (eds.), 167–
187. Berlin: ZAS.
Gussenhoven, Carlos
1983 Testing the Reality of Focus Domains. Language and Speech 26:
61–80.
2004 The Phonology of Tone and Intonation. Cambridge: Cambridge
University Press.
Hartmann, Katharina, and Malte Zimmermann
2004 Focus Strategies in Chadic - the Case of Tangale Revisited. In Inter-
disciplinary Studies on Information Structure 1. Working Papers of
the SFB 632, Shinichiro Ishihara, Michaela Schmitz, and Anne
Schwarz (eds), 207–243. Potsdam: Universitätsverlag Potsdam.
2007 In Place - Out of Place: Focus in Hausa. In On Information Struc-
ture, Meaning and Form, Kerstin Schwabe and Susanne Winkler
(eds.), 365–403. Amsterdam: John Benjamins.
Hayes, Bruce, and Aditi Lahiri
1991 Bengali Intonational Phonology. Natural Language and Linguistic
Theory 9: 47–96.
Heine, Bernd, and Mechthild Reh
1983 Grammaticalization and Reanalysis in African Languages. Ham-
burg: Helmut Buske Verlag.
Hyman, Larry M.
1990 Boundary Tonology and the Prosodic Hierarchy. In The Phonology-
Syntax Connection, Sharon Inkelas and Draga Zec (eds.), 109–126.
Chicago: The University of Chicago Press.
Jackendoff, Ray
1972 Semantic Interpretation in Generative Grammar. Cambridge
(Mass.): Massachusetts Institute of Technology.
78 Sabine Zerbian

Jannedy, Stefanie
2002 Hat Patterns and Double Peaks: The Phonetics and Psycholinguistics
of Broad versus late narrow versus double focus intonations. PhD-
thesis, Ohio State University.
Jokweni, Mbulelo Wilson
1995 Aspects of Isixhosa Phrasal Phonology. PhD-thesis, University of
Illinois.
Jones, Jackie, Jacobus Abraham Louw, and Justus Christian Roux
2001 Queclaratives in Xhosa: An acoustic analysis. SAJAL Supplement
36: 3–18.
Kanerva, Jonni M.
1990 Focus and Phrasing in Chichewa Phonology. New York, London:
Garland Publishing.
Kisseberth, Charles W., and David Odden
2003 Tone. In The Bantu languages, Derek Nurse and Gerard Philippson
(eds.), 59–70: London & New York: Routledge.
Ladd, D. Robert
1996 Intonational Phonology. Cambridge: Cambridge University Press.
Ladd, D. Robert, and Rachel Morton
1997 The perception of intonational emphasis: continous or categorical.
Journal of Phonetics 25: 313–342.
Ladefoged, Peter
2001 Vowels and Consonants - An Introduction to the Sounds of Lan-
guage. Malden, MA: Blackwell.
2003 Phonetic Data Analysis - An Introduction to Fieldwork and Instru-
mental Techniques. Malden: Blackwell Publishing.
Nurse, Derek, and Gerard Philippson
2003 Introduction. In The Bantu languages, Derek Nurse and Gerard
Philippson (eds.), 1–12. London & New York: Routledge.
Oppenrieder, Wilhelm
1989 Fokus, Fokusprojektion und ihre intonatorische Kennzeichnung. In
Zur Intonation von Modus und Fokus im Deutschen, Hans Altmann,
Anton Batliner, and Wilhelm Oppenrieder (eds.), 267–280. Tübin-
gen: Niemeyer.
Paul, Hermann
1880 Prinzipien der Sprachgeschichte. Leipzig: Niemeyer.
Poulos, George, and Louis J. Louwrens
1994 A Linguistic Analysis of Northern Sotho. Hatfield: Via Afrika.
Rialland, Annie, and Stephane Robert
2001 The intonational system of Wolof. Linguistics 39: 893–939.
 Investigating prosodic focus marking in Northern Sotho 79

Schiefer, Liselotte, and Anton Batliner

1991 A ramble round the order effect. Forschungsberichte des Instituts für
Phonetik und Sprachliche Kommunikation der Universität München
(FIPKM) 29: 125–180.
Schwarzschild, Roger
1999 GIVENness, AvoidF and other Constraints on the Placement of
Accent. Natural Language Semantics 7: 141–177.
Selkirk, Elisabeth
2004 Bengali intonation revisited. In Topic and Focus: A Cross-Linguistic
Perspective, Chungmin Lee, Matthew Gordon, and Daniel Büring
(eds.), 217–246: Kluwer Academic Publishers.
Swerts, Marc, Emiel Krahmer, and Cinzia Avesani
2002 Prosodic marking of information status in Dutch and Italian: a com-
parative analysis. Journal of Phonetics 30: 629–654.
Truckenbrodt, Hubert
1995 Phonological phrases: Their relation to syntax, focus, and promi-
nence. PhD-thesis, Massachussetts Institute of Technology.
1999 On the Relation between Syntactic Phrases and Phonological
Phrases. Linguistic Inquiry 30: 219–255.
Xu, Yi
1999 Effects of tone and focus on the formation and alignment of f0 con-
tours. Journal of Phonetics 27: 55–105.
Zerbian, Sabine
2004 Phonological Phrases in Xhosa. In Papers in Phonetics and Phonol-
ogy, Susanne Fuchs and Silke Hamann (eds.), 71–100. Berlin: ZAS.
2006a Expression of Information Structure in Northern Sotho. PhD thesis,
Humboldt-University Berlin.
2006b High tone spread in the Sotho verb. In Selected Proceedings of the
35th Annual Conference on African Linguistics, John Mugane, John
P. Hutchinson, and Dee A. Worman (eds.), 147–157. Somerville,
MA: Cascadilla Proceedings Project.
to appear The subject/object-asymmetry in Northern Sotho. In Information
Structure and the Architecture of Grammar: A Typological Perspec-
tive, Kerstin Schwabe and Susanne Winkler (eds.), Amsterdam: John
Benjamins.
Part II

Information structure and word order

Preverbal objects and information structure in
Benue-Congo

Tom Güldemann

Abstract

It is well known that Africa hosts a considerable number of languages with an

inconsistent word order profile, including a typologically rare clause word order S-
(Auxiliary)-O-V-Other. The most recent non-generative accounts of this pattern
(Claudi 1993, Heine and Claudi 2001) try to explain it exclusively in terms of
grammaticalization scenarios. I will address its possible motivation in Benue-
Congo where important preconditions for the proposed grammaticalizations do not
hold. By observing that preverbal objects are here regularly restricted to subclasses
of objects (e.g. extrafocal participants, pronouns) or to special clause types (e.g.
negation and certain TAM categories), I will try to show that this language group
gives evidence for another factor triggering O-V: this order is arguably associated
with the status of the object as less focal or even extrafocal, non-asserted informa-
tion.

1. Introduction

A number of African languages of different genealogical and areal affilia-

tion show a variable object position, namely before or after the verb, as
exemplified in (1) from Idoma (Idomoid).

(1) ó mç$cí vs. ó lç$cí má

3S see:tree 3S ?:tree see
‘She saw the tree.’ (Abraham 1951: 18)

As argued by Gensler and Güldemann (2003) and Güldemann (forth.),

this feature has a biased distribution on the continent: geographically it is
particularly frequent in a large East-to-West belt South of the Sahara; in
genealogical terms, it is a property of many sub-groups of Niger-Congo.
While early non-generative 1 assessments of such “inconsistent” word or-
84 Tom Güldemann

ders in the area had assumed an older O-V stage, in particular for Niger-
Congo (e.g. Givón 1975), the standard explanation today is in terms of
grammaticalization and reanalysis from a basic V-O clause to innovative
O-V patterns (e.g. Claudi 1993, Heine and Claudi 2001). The two basic
developments, starting (a) in an auxiliary periphrasis involving a nominal-
ized complement and (b) in a serial verb construction, can be schematized
as follows: 2

(a) Verbx - [Genitivey - Verbal nounz] - Other

> Auxiliaryx - Objecty - Verbz - Other
(b) [Verbx - Object] - Verb - Other
> Auxiliaryx - Object - Verb - Other

The grammaticalization hypothesis can account for a number of language-

specific cases, as argued by Heine and Claudi (2001) for languages of the
Kwa (Niger-Congo), Moru-Mangbetu (Central Sudanic), and Ju (a.k.a.
“Northern Khoisan”) families. However, looking at the phenomenon only
from this perspective runs into problems in many other cases; see Gensler
(1997) for a detailed assessment of Claudi (1993).
From a methodological point of view, it often requires one to invoke the
ad-hoc idea that processes of structural generalization and/or analogical
leveling that target the new O-V pattern have erased all or most traces of
earlier grammaticalization. 3 Moreover, the validity of a grammaticalization
scenario does not exclude other explanations, even within one and the same
language. That is, the word order alternation in Africa might well be the
result of “multiple causation”, and there is no a priori reason why it should
be an entirely unitary phenomenon with a single evolutionary explanation
for all its attested cases.
The factual problems of the two grammaticalization scenarios have
mostly to do with the assumed input features that make a change possible in
the first place. As these will be illustrated in section 2 by examples from
Benue-Congo languages, I will just list them at this point. First, the word
order alternation may not be coupled, in a given language, with genitive-
noun order and/or verb serialization. There are even cases where genitive-
noun order actually holds but the object shift still occurs, according to the
language-specific source, without the involvement of a genitive. Rather, the
marked O-V structure represents a true preverbal object complement con-
trolled by a nominalized verb (cf. in this respect German non-finite struc-
tures like Schuhe putzen ‘polishing, to polish shoes’, Fahrkarten kaufen
 Preverbal objects and information structure in Benue-Congo 85

‘buying, to buy tickets’, etc.). These languages suggest the existence of at

least one additional grammaticalization scenario, in which, however, the
preverbal object position would remain an unexplained phenomenon:

(c) Verbx - [Object - Verbal nouny] - Other

> Auxiliaryx - Object - Verby - Other

A further empirical problem with the grammaticalization hypothesis is that

the relevant word order alternation sometimes applies without the involve-
ment of an auxiliary or can fail to apply despite the presence of an auxil-
iary; here, auxiliation cannot, without additional assumptions, be the trig-
gering factor.
Finally, the available grammaticalization scenarios have not properly
addressed the fact that the word order alternations can concern constituents
other than direct objects, including adverbial adjuncts, and/or multiple par-
ticipants.
This paper is an attempt to remedy some of the shortcomings of the cur-
rently dominant explanation for the alternating word order in African lan-
guages. It will look more closely at the phenomenon in a single language
group, Benue-Congo 4 , where (a) the above problems associated with the
exclusive grammaticalization approach are particularly salient and (b) the
relevant word order changes are recurrently associated with a specific func-
tional effect, namely that the marked preverbal position for objects (and
other participants) is typical for information units that are less salient in the
pragmatic context.

2. Object order alternation in Benue-Congo

2.1. Family overview

Benue-Congo languages have a fairly consistent head-initial word order

(cf., e.g. Williamson and Blench (2000: 30–36) and the relevant chapters in
Bendor-Samuel (ed., 1989) for overviews of the basic word order in these
languages): (a) they show a basic V-O clause while O-V order is a marked
alternative in specific grammatical contexts, and (b) the noun phrase dis-
plays a head-modifier pattern, which, except for Nupoid, also includes
genitive constructions. This situation is in line with the currently prevailing
reconstruction of Benue-Congo and Niger-Congo as S-V-O (Heine 1980,
86 Tom Güldemann

Claudi 1993, Manfredi 1997). Nevertheless, as will be shown below, a pos-

sible alternation between post- and preverbal objects is widespread in the
group.
While a generally head-initial noun phrase would exclude the gram-
maticalization scenario under (a) in section 1 for almost all of Benue-
Congo, the development under (b) is also not viable for many languages,
especially in the eastern part of the group, because they lack verb serializa-
tion of the type found in Kwa and western Benue-Congo. Another salient
feature of several Benue-Congo languages is the partial irrelevance of aux-
iliaries for the occurrence of certain O-V structures. This property in par-
ticular has led me to pursue a slightly different approach to the general
problem. That is, instead of looking only for the pattern S-Auxiliary-O-V-
Other, I have undertaken a general survey of O-V and related structures,
irrespective of whether an auxiliary is involved.
The result of this still preliminary investigation is given in Table 1. This
list is not exhaustive because it omits several languages which are gene-
alogically close to those given in the table and show similar behavior.

Table 1: Attested cases of preverbal objects in Benue-Congo

__________________________________________________________________
Language Group Sub-branch(es) Source
______________________________________________________________________________________________________________________________________________________________________

Nupe Nupoid George 1971, Kandy-

bowicz and Baker 2003
Idoma Idomoid Abraham 1951
Yoruba Defoid Manfredi 1997
Igbo Igboid Manfredi 1997
Ibibio Cross River Lower Cross Urua 1997
Leggbo Cross River Upper Cross Good 2004
Kana Cross River Kegboid Ikoro 1996
Kaje Platoid Central Plateau Wolff and Gerhardt 1977
Mambila Bantoid Northern, Mambiloid Perrin 1994
Vute Bantoid Northern, Mambiloid Thwing and Watters 1987
Tikar Bantoid Southern Stanley 1991
Naki Bantoid Southern, Beboid Good 2005
Aghem Bantoid Southern, Grassfields Watters 1979
Bafut Bantoid Southern, Grassfields Chumbow and Tamanji
1994, Tamanji 2002
Nen Bantoid Southern, Mbam Mous 1997, 2005
Ewondo Bantoid Southern, Narrow Bantu Redden 1979
Proto-Bantu* Bantoid Southern, Narrow Bantu Meeussen 1967
______________________________________________________________________________________________________________________________________________________________________

Note: * except Northwest

 Preverbal objects and information structure in Benue-Congo 87

In the following, I will present data from individual languages grouped

according to certain recurrent factors that trigger the shift of an object, and
partly of other participants, to the marked preverbal clause position.

2.2. Pronominal objects

A first group of cases is characterized by the shift of PRONOMINAL ob-

jects. For example, this is reported in Ibibio (Cross River) for an apparently
isolated context, namely reflexive experiencer pronouns. Compare (2), with
a nominal object canonically following the verb, and (3), where the 1st-
person singular object m@ precedes the experiencer predicate biák ‘hurt, be
painful’; note that no auxiliary is involved in the construction.

(2) Òkôn á dêp ébót

PROP 3S buy goat
‘Okon is buying a goat.’ (Urua 1997: 201)

(3) úbç@k á m@ biák

hand 3S 1S be.painful
‘My hand hurts.’ [lit.: ‘The hand hurts me.’] (Urua 1997: 204)

In Vute (Bantoid), the marked preverbal position of pronoun objects is

more frequent in that it applies to human referents in all clauses with imper-
fective aspect. This is shown by the difference between (4) with a nominal
object and (5) with a pronominal object, both of which involve an auxiliary.

(4) yáyá taá sè-nà ni tá ƾgée júk

PROP PST:IPFV wash-IO ? father 3S.POSS clothes
‘Yaya was washing clothes for his father.’
(Thwing and Watters 1987: 100)

(5) ƾgáb á ƾgé jìm ni

3P IPFV 3S beat ?
‘They are beating him.’ (Thwing and Watters 1987: 100)

Preverbal pronoun objects are even more salient in Kana (Cross River)
since here the word order change is possible or obligatory for a wider range
of preverbal auxiliaries and predicate inflections (Ikoro 1996: 206-13). One
88 Tom Güldemann

marker which generally triggers the change encodes past tense; compare (6)
where the 1st-person singular pronoun mE# precedes a serial verb structure.

(6) we#è mE#-te#e#ra# p"#" #

3S:PST 1S-run meet
‘He ran to me.’ (Ikoro 1996: 212)

However, example (7), which is semantically identical to (6), illustrates an

important fact about the general phenomenon in Kana: as soon as the pro-
nominal object is pragmatically salient, even a past tense clause no longer
has the object in the marked preverbal position.

(7) we#è te#e#ra# píí n#da#

3S:PST run meet 1S.EMPH
‘He ran to ME.’ (Ikoro 1996: 212)

An even stronger correlation between the presence of an auxiliary and the

preverbal position of a pronoun object is found in certain Plateau languages
such as Kaje and Afusare (Wolff and Gerhardt 1977: 1527–8). Another
example is Ewondo (Bantoid, Bantu A70), as described by Redden (1979:
126, 166–7). Compare (8), with a postverbal object pronoun in a clause
without an auxiliary, and (9), with a preverbal object pronoun in a clause
with an auxiliary (habitual kad); the latter example also shows that Ewondo
is a language where preverbal position can apply to multiple objects.

(8) ma-v‹" Ջ Ջ"

1S-give 3S.H 3S
‘I am giving it to him.’ (Redden 1979: 166)

(9) a-kad m‹ dzn v‹"

3S-HAB 1S 3S give
‘He usually gives it to me.’ (Redden 1979: 167)

2.3. Negation

Another recurrent context for the preverbal occurrence of, often multiple,
objects is negation. In Vute (Bantoid), this applies to perfective clauses, as
shown in (10).
 Preverbal objects and information structure in Benue-Congo 89

(10) a. mvèìn yi ˜wáb-na tí ƾgé cene

chief REM.PST buy-IO ? 3S chicken
‘The chief bought him a chicken.’
b. mvèìn yi ní ƾgé cene ˜wáb-na wá
chief REM.PST NEG 3S chicken buy-IO PFV.NEG
‘The chief did not buy him a chicken.’
(Thwing and Watters 1987: 101)

In other languages, negative clauses cause the object to precede the verb in
a more generalized fashion. Compare in this respect (11a) with (11b) from
Leggbo (Cross River, Upper Cross), or (12a) with (12b) from Bafut (Ban-
toid, Grassfields). Similar cases are reported in other closely related lan-
guages, for example Lokaa (Iwara 1982) from Upper Cross, and Nweh
(Nkemnji 1995: 112–47) and Mbili (Ayuninjam 1998: 338–53) from Grass-
fields Bantu.

(11) a. wàdum s'" e-nii wà'" s'" ntààmi

man DEF 3S-give child DEF gift
‘The man gave the child a gift.’
b. wàdum s'" wà'" s'" ntààmi eè-nii
man DEF child DEF gift 3S:NEG-give
‘The man didn’t give the child a gift.’ (Good 2004)

(12) a. Sùù kì kó mbà

PROP TA catch animal
‘Suh killed an animal.’
b. ka#a# Sùù kì wa#'à mbà ko#
NEG PROP TA NEG animal catch
‘Suh did not kill an animal.’ (Chumbow and Tamanji 1994: 224)

2.4. Non-finite verb phrases

A third typical context for preverbal objects in Benue-Congo is non-finite

verb phrases which may or may not be controlled by an auxiliary. This
occurs recurrently inter alia in Western groups of Benue-Congo; it is re-
ported for Nupoid (cf. Hyman and Magaji (1971: 28–33) for Gwari, and
Kandybowicz and Baker (2003: 131–145) for Nupe), Idomoid (cf. Abra-
ham (1951: 71–7) for Idoma), Defoid, and Igboid. A representative exam-
ple is given here from Standard Igbo (Igboid).
90 Tom Güldemann

(13) óB mára-na bùbá a-kwá

3S know-PFV blouse NOM-sew
‘S/he knows how to sew blouses.’ (Manfredi 1997: 98)

Several facts should be noted in connection with the construction in (13).

First, this particular combination of verbal noun and object cannot be used
outside the present context after the controlling verb mára ‘know’ – for
example, it cannot appear as a sentence subject. Second, this is the only
pattern possible as long as the object is semantically specific. As soon as
the object is generic and – similar to a cognate object – notionally con-
tained in the verb, an interesting variability can be observed: one can either
use the structure given in (13), as shown in (14a), or have recourse to a
nominalization type with a different nominalizer, where the object
FOLLOWS the verb, as in (14b). 5

(14) a. óB mára-na akwà a-kwá

3S know-PFV cloth NOM-sew
b. óB mára-na iB-kwá akwà
3S know-PFV INF-sew cloth
‘S/he knows how to sew.’ (Manfredi 1997: 98)

The following example of a non-finite verb phrase with a preverbal object

is from Standard Yoruba (Defoid); according to the author, this particular
nominalization type need not be preceded by a complement-taking verb
like ‘learn’ in (15).

(15) ó kóB ìwé kí-kà

3S learn book NOM-read
‘S/he learned the art of reading.’ (Manfredi 1997: 96)

While the O-V order in this Yoruba construction is fixed with verbs imply-
ing a lower object that is generic, like ‘know’ and ‘learn’, the quite similar
structure after the verb ‘want’ allows for both marked O-V, as in (16a), and
normal V-O, as in (16b) (note again the different nominalizers).

(16) a. mo féB asBoB óB-hun

1S want cloth NOM-weave
 Preverbal objects and information structure in Benue-Congo 91

b. mo féB é-hun asBoB

1S want NOM-weave cloth
‘I want to weave (some) cloth.’ (Manfredi 1997: 96)

The Upper Cross languages Leggbo and Lokaa, already mentioned in Sec-
tion 2.3, are also cases where preverbal objects are typical for nominaliza-
tions in general. As the following examples from Leggbo show, the feature
is not only characteristic of verbal nouns controlled by another verb, like
phasal ‘begin’ in (17); it also occurs when the nominalized verb phrase
functions as the subject, as in (18). This last example also demonstrates that
the alternation in object position not only can be independent of an auxil-
iary but also that it can apply to more than one participant.

(17) b a-ttnngn lídzil gèdziE$

3P 3P-begin food eat:NOM
‘They began eating food.’ (Good 2004)
(18) n$tààmi bE$'" gèniE$ nEE$-dàa m
gift children give:NOM 3S:HAB-please 1S
‘I like giving gifts to children.’ (Good 2004)

2.5. Preverbal inflections and auxiliaries

As is well known from the literature on the topic, the shift between a post-
and a preverbal object is often tied to the absence or presence of an auxil-
iary or some other preverbal inflection. While in the cases discussed in the
previous sections certain additional conditions pertain (pronominal object,
negation, nominalization, etc.), the phenomenon can be more general as, for
example, in Tikar (Stanley 1991: 102–28, 132–58). This Bantoid language
has S-V-O order in perfective and modal clause types without a preverbal
auxiliary; this pattern is shown in (19).

(19) wù sh-ê mùn

2S say-IRR 1S
‘Si tu me l’avais dit, ...’ (Stanley 1991: 71)

With preverbal inflections and auxiliaries, however, the object occurs be-
fore the verb, as with the imperfective marker tă in (20) and the phasal
auxiliary Èunmi ‘begin’ in (21).
92 Tom Güldemann

(20) à tă nshe sh'

3S IPFV luggage carry
‘Il porte le bagage.’ (Stanley 1991: 103)

(21) à Èunmi nun ji f'

3S begin:PFV 3S food give
‘Il a commencé à lui donner de la nourriture.’ (Stanley 1991: 133)

However, such a situation does not seem to be typical for Benue-Congo as

a whole; more often O-V patterns are restricted to a smaller set of inflec-
tions/auxiliaries (see Manfredi’s (1997) discussion of “aspectual licensing
and object shift”). Important for the following discussion is the observation
that several of these cases seem to be sensitive to the pragmatic aspect of
information structure.
One such structure, fairly wide-spread in the Westernmost subgroups of
Benue-Congo, is the likely result of the grammaticalization scenario men-
tioned in section 1 under (b) involving an earlier verb ‘take, have, hold, get’
(cf. Lord 1982). Nupe (Nupoid) is a case in point. Past clauses display a
distinction between a plain V-O phrase, as in (22a), and an O-V structure,
as in (22b); the latter involves an element á which derives etymologically
from a verb ‘take’.

(22) a. Musa zu@) tsùku$)

PROP break stick
‘Musa broke the stick.’
b. Musa á tsùku$) zu@)
PROP PERF stick break
‘Musa got the stick broken/ has broken the stick.’
(George 1971: 90, 93)

While both sentences convey a perfective event, they differ in another re-
spect: the first pattern with unmarked V-O would be a response to a ques-
tion like “What did Musa do?” or “What did Musa break?” while the sec-
ond structure with marked O-V answers “What did Musa do with (or to) the
stick?”, where the object noun is given (also cf. Lord (1982) for other cases
where the preverbal object correlates with definiteness). Since the á-
construction in Nupe is characterized by a focus on the predicate, including
its aspectual operator of perfectivity, it has often been interpreted in terms
of aspect, namely as a resultative~perfect.
 Preverbal objects and information structure in Benue-Congo 93

The example pair in (23) shows the á-construction with a verb and a
cognate-like object. Here, the object can either precede the verb or follow
it. This suggests that the structure involves more than aspect and that the
object position itself is more directly related to information structure. That
is, it can be argued that the preverbal object itself characterizes sentences
where the focus is on the predicate as the carrier of the assertion.

(23) Musa á eci bi or á bi (e)ci

PROP PERF race run PERF run race
‘Musa (really) ran (well) / ran off.’ (Manfredi 1997: 100)

Verb serialization with ‘take’ is also at the base of the constructions with
preverbal objects reported by Adive (1989: 80–1, 130–1) for Ebira, by
Hyman and Magaji (1971: 56–8, 63) for Gwari (both Nupoid), by Manfredi
(1997: 97) for OBwoBroB Yoruba (Defoid), and by Abraham (1951: 18) for
Idoma (Idomoid, cf. (1)). While not all authors mention a functional differ-
ence vis-à-vis the unmarked clause with a postverbal object, it is likely that
the situation is – at least from a diachronic perspective – similar to that in
Nupe, where the marked O-V pattern can be characterized as a combination
of perfectivity and focus on the predication instead of the object.
Another case where a preverbal object seems to be associated with a
tense-aspect form that involves a specific information structure can be iden-
tified in southern Igbo varieties (Igboid). For example, in Avu` Igbo, the
future auxiliary gà can be followed by two phrase types involving different
verbal nouns; in one, the object follows the nominalization, as in (24a), and
in the other, the object precedes it, as in (24b).

(24) a. ó` gà e-rí rin ahù`

3S FUT NOM-eat food that
‘S/he is going to eat that food.’
b. ó` gà rín ahù` n@-ri
3S FUT food that NOM-eat
‘S/he must (certainly) eat that food.’ [?< ... WILL eat ...]
(Manfredi 1997: 99)

The two structures are associated with a subtle meaning distinction which
can be interpreted in terms of a difference in information structure: while
the form with a postverbal object is a plain future, the one with a preverbal
object renders an “(epistemic or deontic) obligative future.” Like the previ-
94 Tom Güldemann

ous case of auxiliary perfects, this can arguably be analyzed as a combina-

tion of tense-aspect and information structure, namely as a future with fo-
cus on the truth value of the proposition; from this configuration the modal
reading of obligation can be derived.

2.6. Information structure

A final set of structures displaying variable object position in Benue-Congo

languages is overtly and explicitly associated with an alternating informa-
tion structure. To date, these structures have been reported exclusively from
Bantoid languages which are spoken in a fairly restricted area along the
Southern stretch of the border between Nigeria and Cameroon, that is,
Mambila from Mambiloid (Perrin 1994), Naki from Beboid (Good 2005),
Aghem from Grassfields (Watters 1979), and Nen from Mbam (Mous
1997, 2005).
I will start with Aghem because it is the case first reported in the litera-
ture and has been researched fairly extensively. While the overall focus
system is morphologically and syntactically far more complex than can be
discussed here, the basic relevant facts, according to Watters (1979), can be
summarized as follows. The linear position immediately after the verb is
the site for focused constituents while the position before the verb can be
used for pragmatically less salient participants; this mechanism is largely
insensitive to different semantic relations. Thus, the sentence in (25a) is
used when the exclusive focus is on the direct object, and the time adverb is
given information. This pragmatic relation between the two non-subject
participants is reversed in (25b); here, the time adverb is the new informa-
tion, and the object is defocused. If both participants are extrafocal, both
occur before the verb, which then ends up being clause-final, as in (25c).

(25) a. fíl á mç$ á'zç@ç zí kí-bE@

friends 3P PST yesterday eat CL-fufu
‘The friends ate FUFU (not yams) yesterday.’
b. fíl á mç$ bE@-'kí zí á'zç@ç
friends 3P PST fufu-CL eat yesterday
‘The friends ate fufu YESTERDAY (not two days ago).’
c. fíl á máà bE@-'kç@ á'zç@ç zí
friends 3P PST.PF fufu-CL yesterday eat
‘The friends DID TOO eat fufu yesterday.’
(Watters 1979: 149, 148, 150)
 Preverbal objects and information structure in Benue-Congo 95

A largely comparable situation has been described by Perrin (1994) for

Mambila. A postverbal object is associated with focus, as shown in (26a).
Extrafocal constituents are placed before the verb, as happens with the time
adverb in (26b). It is also possible that both these constituents are focused,
as in (26c); in such cases, the last one would be more salient. An overt de-
focalization of the object is illustrated in (26d) with exclusive focus on the
time adverb and in (26e) with focus on the verb.

(26) a. mè ƾgeé naâ cngn

1S buy PST cloth
‘I bought cloth.’ or ‘It was cloth that I bought.’
b. mè léílé ƾgeé naâ cngn
1S yesterday buy PST cloth
‘It was cloth that I bought yesterday.’
c. mè ƾgeé naâ cngn léílé
1S buy PST cloth yesterday
‘I bought cloth yesterday.’
d. mè cngn ƾgeé naâ léílé
1S cloth buy PST yesterday
‘It was yesterday that I bought cloth.’
e. mè naâ cngn ƾge
1S PST cloth buy
‘I bought cloth.’ [assertive verb focus] (Perrin 1994: 233)

The situation in Nen appears to be quite different at first glance but it turns
out to be comparable to the two previous cases. According to Mous (1997,
2005), the preverbal position for an object is unmarked and conveys asser-
tive focus on this participant, as in (27a). The postverbal object, which must
be preceded by a particle such as á, as in (27b), is associated with contras-
tive focus.

(27) a. àná mònE@ índì

3S:PST money give
‘S/he gave money.’
b. àná índì á mònE@
3S:PST give ? money
‘S/he gave MONEY.’ (Mous 1997: 126)
96 Tom Güldemann

As opposed to Aghem and Mambila, the preverbal object in Nen is not

extrafocal. However, on a more abstract level, there is still a common de-
nominator across the three languages in that the preverbal object is prag-
matically less salient vis-à-vis its postverbal counterpart.

2.7. Summary

Table 2 gives a summary of the grammatical profile of O-V structures in

the Benue-Congo languages treated above, as far as the information is
available in the source. If a language has multiple O-V constructions which
differ according to the features surveyed here, they are treated in separate
lines.
Two major observations can be made. First, many cases are character-
ized by the irrelevance of preverbal auxiliaries/inflections and by the possi-
bility of subjecting more than one object or a non-object participant to the
word order shift. These cannot, without further assumptions, be claimed to
be instances of the grammaticalization scenarios mentioned in section 1 and
thus call for another explanation.
Second, O-V order in Benue-Congo is in general conditioned by
RECURRENT grammatical factors. Some relate to a property of the affected
object, in that it must be a pronoun, or refer to a generic concept, or have an
extrafocal status. Others relate to the type of predicate in a wider sense: it is
negated, has a certain tense-aspect feature, or is characterized by an auxil-
iary in general.

Table 2: Contexts of preverbal objects in Benue-Congo

___________________________________________________________
Language Specific Mul- Non- Nega- Spe- Non- Pre-
(Section object tiple object tion cific finite verbal
number) feature objects TAM verb aux./
phrase infl.
__________________________________________________________________
Nupe (2.4) YES YES ? YES
(2.5) Non-focal NO ? PERF YES
Idoma (2.4) ? ? NO (YES)
(2.5) NO YES PST YES
Stand. Yoruba (2.4) (Generic) ? NO (YES)
OBwoBroB Yoruba (2.5) NO ? PERF YES
Stand. Igbo (2.4) NO NO (YES) YES
Avu` Igbo (2.5) ?Non-focal NO NO (FUT) YES YES
Ibibio (2.2) (Pronoun) NO NO
 Preverbal objects and information structure in Benue-Congo 97

Leggbo (2.4) YES YES YES

(2.3) YES YES YES
Kana (2.2) Pronoun ? NO YES
Kaje (2.2) Pronoun ? ? YES
Mambila (2.6) Non-focal YES YES
Vute (2.2) (Pronoun) NO NO (IPFV)YES
(2.3) YES ? (YES) (PFV) YES
Tikar (2.5) ? NO IPFV YES
(2.5) YES NO YES
Naki (2.6) Non-focal YES YES
(2.3) YES ? (YES) YES
Aghem (2.6) Non-focal YES YES
Bafut (2.3) YES YES YES YES
Nen (2.6) Less focal YES NO
Ewondo (2.2) Pronoun YES YES YES
Proto-Bantu (3) Pronoun (YES) (YES) YES
__________________________________________________________________
Note: Blank = feature not relevant, (...) = restricted relevance, ? = no information

A look at other languages with object order alternations reveals that

such factors are not restricted to Benue-Congo but can be identified else-
where – in particular, including languages which are genealogically and/or
areally unrelated. One context of preverbal objects noted above for Benue-
Congo but also found outside this group is progressive and imperfective
clauses, for example, in some languages of Moru-Mangbetu (Central Su-
danic, Andersen 1984, Watson 1997) and Gbe (Kwa, Lefebvre and Brous-
seau 2002); it is also found in Kaansa and Cefo (Gur, Elders p.c.). Negation
triggers a different object position in some Gurunsi languages (Gur, Elders
p.c.) as well as in parts of Surmic (Unseth 1986) – like Moru-Mangbetu, a
family which is unrelated to Benue-Congo. To have a different position for
pronominal vis-à-vis nominal objects is cross-linguistically quite common
(cf. object clitics in Romance); it is also attested in other West African lan-
guages, like some of the Oti-Volta group (Gur, Elders p.c.) and most of the
Atlantic family (Childs forth.). In some Potou-Tano (Kwa) languages and
in Mandarin Chinese, the preverbal object is associated with the status of
definiteness and discourse-givenness (Stewart 1963, Lord 1982). This has
not been described explicitly for Benue-Congo; it is, however, noteworthy
that the verbal object prefixes in Savannah Bantu, which arose from an
earlier S-(Auxiliary)-O-V pattern, correlate with a definite postverbal ob-
ject (cf., e.g. Creissels 2000: 235). Finally, information structure plays a
98 Tom Güldemann

role in shifting object position in Doyayo (Adamawa, Elders 2004) and in

the unrelated Southern Cushitic group (Mous 1993, Kießling 1994).

3. Information structure as a potential common denominator of O-V

At first glance, these multiple factors seem fairly heterogeneous. I will,

however, propose in the following that a common denominator does exist. I
will start out from cases where O-V has the most straightforward function,
namely where it correlates with some specific information structure. In
languages like Aghem, Mambila, and Naki, the preverbal object is explic-
itly extrafocal, and the focus is on another constituent or on a clause fea-
ture. As mentioned, the situation in Nen is similar on a more abstract level:
the preverbal object is not extrafocal but it is still less focal than its post-
verbal counterpart. I will now argue that a decrease in pragmatic salience of
the object can also be identified for other, more grammaticalized contexts
of O-V structures.
A good case for this assumption can be made with regard to preverbal
object pronouns. The use of a pronoun is largely correlated with discourse
anaphora and implies the status of its referent as given/presupposed. This
means that a pronoun object is not part of the new asserted information
(recall in this respect the case of Kana in Section 2.2: as soon as the object
pronoun is emphatic, O-V no longer applies). A similar argument can be
made for preverbal definite objects; definiteness also implies a discourse-
given status, hence, such an object is extrafocal or less focal.
The situation for preverbal generic objects is slightly different. Here, the
object is not necessarily given information; it is, however, also not asserted
independently but instead demoted so as to become a semantic subpart of a
complex information unit. This analysis also applies in part to objects that
occur within non-finite verb phrases. The non-asserted status in a kind of
“incorporation” structure can correlate with a lower discourse saliency (see,
e.g., Mithun 1984).
At least since Hyman and Watters’ (1984) work on so-called “auxiliary
focus”, it has been known that certain features of the predication, such as
polarity, aspect, modality etc., can interact in a systematic way with the
information structure of a clause. The authors propose that these specific
predication operators have inherent focus. I follow this position and assume
that some of the factors that trigger or favor O-V are related to this phe-
nomenon.
 Preverbal objects and information structure in Benue-Congo 99

That negation, in particular of the “metalinguistic” type (cf. Horn 1989),

is an operator with inherent focus has been argued by a number of authors
such as Givón (1978), Heine and Reh (1983), Marchese (1983), and
Güldemann (1996, 1999). Thus, the fact that preverbal objects recurrently
co-occur with negation can be taken as a reflex of their less focal status in
this context.
Other inherently focused features, according to Hyman and Watters
(1984), are the progressive and the perfect. For the progressive, this has
been discussed in more detail by Güldemann (2003) for Bantu languages
and some other geographically and genealogically unrelated cases. Prever-
bal objects tied specifically to the progressive have not yet been found in
Benue-Congo. However, certain phenomena which come close to this do
exist (see also below regarding the historical relation between progressives
and imperfectives). Recall from Section 2.5 that clauses in Tikar with im-
perfective tă generally have O-V order. However, the situation is slightly
more complex with “semi-intransitive verbs”, which are characterized by
optional locative complements. The position of such complements is vari-
able and this is associated with a semantic distinction that implies a direct
correlation between O-V and the progressive. If the complement follows
the verb, the clause has a habitual reading, as in (28a); if it occurs in the
marked preverbal position, the clause renders a progressive state of affairs,
as in (28b).

(28) a. à tă kE$n fumban

3S IPFV leave PROP
‘Il a l’habitude de partir à Foumban.’
b. à tă fumban kE$nni
3S IPFV PROP leave
‘Il est en train de partir à Foumban.’ (Stanley 1991: 114)

Within the present framework, the cases of the perfect in Nupe, OBwoBroB
Yoruba, and possibly other languages of the area, as well as the “obliga-
tive” future in Southern Igbo treated in section 2.5 are more straightfor-
ward. They can be analyzed as configurations of information structure
where a predication operator is in focus; this causes the object to be prag-
matically less salient and to occur outside its canonical postverbal position.
I will conclude the discussion with those cases where preverbal object
position is relevant for auxiliary periphrasis more generally and an exclu-
sive explanation in terms of information structure is problematic. One
100 Tom Güldemann

could, of course, first investigate the question of whether periphrasis has a

general tendency to focus on the auxiliary category and thereby defocus
participants. There are, however, other possible scenarios for how a lan-
guage can come to apply an O-V order to many or even all periphrastic
clause types, irrespective of whether the object has a decreased information
value or not. For one thing, it is conceivable that the grammatical pattern
O-V with preverbal predication operators becomes the target of analogical
generalization and gradually expands in the language. A second possibility
is that an inherently focused verb category develops into another category
which is less sensitive to a particular information structure. In view of the
results of grammaticalization research, this hypothesis is particularly attrac-
tive for an imperfective with general O-V, because this category frequently
originates, according to e.g. Bybee, Perkins and Pagliuca (1994: 140–9),
from a progressive, which I argue is inherently focused. In both scenarios,
analogy and grammaticalization, the marked word order would be retained
but the original explanation in terms of less focal objects would no longer
need to apply.
The general argument for the V-O > O-V alternation in the relevant lan-
guages is tied to an important point which should be stated in a more ex-
plicit fashion. I have repeatedly pointed out that preverbal position of the
object or, more generally, of a non-subject participant is the marked pattern
and that this is associated with a less salient information status of this ele-
ment. Conversely, then, the position after the verb is the unmarked position
for focus on a term like an object, adjunct, or adverb in the sense of Dik
(1997); the focus is mostly assertive but in a few cases (also) contrastive, as
in Mambila and Nen. Consequently, in the cases treated in section 2.6,
where information structure is synchronically the factor transparently con-
trolling the word order shift, the alternation not only applies to a single
direct object but to multiple objects as well as to non-objects. This syntacti-
cally expressed pragmatic opposition lies at the heart of my hypothesis for
the object order alternation: in a V-O clause, the object is (part of) the as-
sertive focus while in an O-V clause the object is defocused or at least less
salient vis-à-vis a postverbal object.
The present account of the relevant word order alternations in terms of
information structure does not contradict in principle the explanation of O-
V structure in terms of grammaticalization given in Section 1, that is, an
origin of some language-specific structures in a concrete syntactic configu-
ration of auxiliary + nominalized complement or verb serialization. Rather,
it is compatible with and/or entails these historical scenarios insofar as they
 Preverbal objects and information structure in Benue-Congo 101

ultimately cause the object to occur early in the clause and hence to iconi-
cally precede other, informationally more salient elements.
The present proposal is also not meant to exclude the existence of other
factors that may contribute to the existence of O-V structures. To mention
only one other promising line of research for the future: the relatively fre-
quent occurrence of pronominal items before the verb is likely to be related
to their overall phonetic weight: as light elements they may tend to precede
heavier ones like lexical verbs.
While the order alternation regularly affects a wide range of constituents
in some languages (see, e.g. section 2.6), the predominant situation in Be-
nue-Congo is that it has a clear bias towards the object. Why should it be
like that, given that an object is generally viewed as a grammatical relation
that refers primarily to a semantic role implied by the verb? What is the
special connection between the syntactic concept of object and the prag-
matic notion of assertive clause focus that would cause pragmatically sensi-
tive word order manipulations to target the object in particular?
My basic idea is that all clause types are inherently associated with
some type of information structure. This also applies to so-called “un-
marked” clauses in fixed-word-order languages, which are often neglected
in the discussion of focus phenomena. My more specific claim is that these
clauses are unmarked not just in terms of basic syntactic configuration but
also with respect to information structure. That is, they are “in-situ” focus
constructions with a default configuration of assertive focus (= Dik’s
(1997) “completive” focus). In V-O languages with a grammaticalized
subject relation that conflates the semantic agent-role complex and the
pragmatic function of topic, this makes it possible to assume an even more
specific situation for basic clauses. All other things being equal, the default
in an intransitive S-V sentence is assertive focus on the predicate, and in a
transitive S-V-O sentence assertive focus on the object, possibly including
the predicate. In the latter case, even if the verb is also new, the object can
still be granted a higher information value in that the iconic linear progres-
sion is from less to more salient information. Moreover, and, given that in
true O-V languages the preverbal object still seems to be more prominent
than the final verb, possibly more importantly, nominal referents may be
more central to the encoding of information structure than predicates.
To the extent that the basic transitive clause of a language is a salient
construction type, it can be argued that the grammatical relation between
object and verb is inherently tied not only to certain semantic roles but also
to a pragmatic function. The default position of an object is closely associ-
102 Tom Güldemann

ated with its status as newly asserted information. This correlation can be
exploited by a language: it can place the object in an unusual position in
contexts where it has a non-canonical information status, i.e. where it is
pragmatically given or less salient vis-à-vis a more marked focus on an-
other constituent or a predication operator. This option can be seen as a
direct function of at least two language-specific factors: (a) the degree to
which a language allows for word order flexibility, and (b) the degree to
which it highlights pragmatic function in addition to semantic role in the
grammatical object relation.
In proposing this fairly abstract common denominator for the relevant
word-order alternations in general and for preverbal objects in particular, I
do not want to claim that, synchronically, the explanation accounts for all
cases in a straightforward way. Clearly, the recurrent occurrence of an ob-
ject in preverbal position does not directly reflect its extrafocal information
status but results initially from a particular grammatical construction.
Moreover, these conditioning constructions are quite heterogeneous, even
in one and the same language (cf., e.g. Nupe and similar languages where
the ‘take’-serializations and phrases involving a verbal noun generate two
quite different types of preverbal objects). Thus, while O-V under negation
in the relevant languages comes fairly close to overt defocalization, auxil-
iary periphrases are overall different in that they often trigger O-V for
purely grammatical reasons. Thus, control auxiliaries such as ‘know’,
‘want’, ‘start’, ‘be at’, etc. may simply require a non-finite verb phrase,
which in the language may happen to require a preceding object comple-
ment. Also, in a core serialization, the patient object of ‘take, get, hold’
automatically ends up before the second verb and any additional partici-
pant. In both cases, the information status of the object vis-à-vis other ele-
ments is only indirectly reflected and may be an effect rather than the im-
mediate cause.
However, information status sometimes seems to be relevant even in
such problematic cases. The Standard Yoruba examples (15) and (16)
showed that the semantics of control verbs can bear on whether an object
appears before or after the controlled content verb. The phrase ‘learn to
VERB X’ often implies an activity that, once learned, is carried out rou-
tinely and then is associated preferably with a non-individuated, non-
specific, and thus pragmatically less salient object. The phrase ‘want to
VERB X’ does not seem to have this habitual connotation and thus is more
open with respect to the pragmatic status of the lower object. While the
 Preverbal objects and information structure in Benue-Congo 103

former requires O-V order in the non-finite verb phrase, the latter allows
for both O-V and V-O.
A different situation can hold in the case of objects which are semanti-
cally entailed by the content verb: these are least likely to attract the clause
focus and hence are the maximal opposite of a specific and/or referential
object. In the case of ‘know to VERB X’ in Standard Igbo, this would seem
to make a defocusing manipulation “superfluous” – as in (14), where V-O
is possible besides normal O-V, in contrast to (13), where the semantically
more specific object must precede the verb.
The non-focus status of cognate-like objects might also be at the basis of
the alternation in (23) from Nupe: here, the V-O pattern simply asserts a
complex predicate while the O-V pattern arguably focuses on the truth of
the predicate, which demotes the object even further in terms of salience.
Since O-V patterns of some sort can pervade entire language groups, the
present proposal has potential historical implications for earlier language
states, and the question emerges as to what situation merits reconstruction
for what chronolect. Gensler (1994, 1997) has gone some way in this direc-
tion by proposing to reconstruct the syntagm S-Auxiliary-O-V-Other, be-
sides S-V-O-Other, back to Proto-Niger-Congo. The fuller range of data
and their discussion in this paper throw new light on this issue. I would
agree with Gensler in general that O-V is, alongside V-O, a historically
older pattern in Niger-Congo and not just the ever-recurring outcome of a
set of grammaticalization paths that just start out from V-O and its con-
comitant structures. At the same time, a reconstruction which entails such a
morphosyntactic alternation should also address and answer questions re-
garding the distribution of and the ultimate motivation for this variability,
and this on the basis of the particular syntactic profile of each modern
group considered.
Drawing from the above discussion, I would like to propose the follow-
ing refinements to Gensler’s hypothesis. A reconstruction of a generalized
S-(Auxiliary)-O-V-Other syntagm is likely to overstate the case for most
sub-groups of Niger-Congo. A case in point is Proto-Bantu. The Northwest
Bantu languages with preverbal pronouns like Ewondo and the modern
morphotactic pattern in Savannah Bantu languages, where object affixes in
a complex agglutinative verb usually precede the verb stem (cf. Meeussen
1967: 96–9), do not warrant such a far-reaching hypothesis. Instead, it
would be more faithful to the available evidence to assume for the relevant
proto-language a situation similar to that in Kana, Kaje, and Ewondo where
104 Tom Güldemann

all or at least many auxiliaries triggered the positional shift of pronominal

objects only.
Such a situation with a clear grammatical restriction on O-V would
plausibly have evolved from an earlier ancestor language with a similar, yet
different profile. This can be modeled in a preliminary fashion on the em-
pirical data presented here from Benue-Congo (irrespective of whether this
group turns out after more research to be a valid genealogical sub-group
within Niger-Congo). In order to account for the variation across Benue-
Congo, I posit that an earlier language state displayed a basic V-O syntax
which, however, was different from that of “canonical” S-V-O languages in
a few important respects. Sentence positions in this early language state,
including the position after the verb, had besides their association with a
typical semantic profile a greater pragmatic load than is usually recognized
by standard syntactic analyses. This feature would have assigned to the
postverbal object a stronger association with the role of assertive focus of
the clause. This could well have gone hand in hand with a weaker syntactic
bond between the verb and its object. Apart from possible external factors
(e.g. contact with languages of a different word-order type), these proper-
ties would have facilitated the development of an optional O-V pattern
which was associated with defocusing the object.
Other groups in Greenberg’s Niger-Congo like Adamawa-Ubangi, Kwa,
Gur, and Kru display an overall similar situation regarding a flexible object
position (cf. Gensler and Güldemann 2003). These, together with Benue-
Congo, form what could be viewed as the more secure core of a genealogi-
cal unit (cf. Stewart 1976 and Bennett and Sterk 1977 regarding such con-
cepts as “Volta-Congo” and “Central Niger-Congo”, respectively). Thus,
the situation briefly outlined above may possibly have already taken place
in a far older chronolect. Clearly, more research on this domain is neces-
sary. I would venture, however, that a realistic and tangible morphosyntac-
tic reconstruction of a proto-language is only possible if the whole range of
synchronic variation in the daughter languages is taken into account and an
attempt is made to relate the different patterns by means of observable
functional correlations.

Notes

This paper was presented previously at the following occasions: at the

conference “Topic and Focus: Information Structure and Grammar in Af-
 Preverbal objects and information structure in Benue-Congo 105

rican Languages”, University of Amsterdam (03/12/2004); at the “Work

in Progress” series, MPI-EVAN Leipzig (14/12/2005); and at the “Jour
fixe - Afrikalinguistisches Kolloquium” series, Universität Leipzig
(15/12/2004). Thanks are due to the respective audiences for interesting
discussions and to Orin Gensler and two reviewers for comments on a
previous draft. The following abbreviations are used in the paper: CL
noun class, DEF definite, EMPH emphatic, FUT future, H human, HAB
habitual, INF infinitive, IO indirect object, IPFV imperfective, IRR irre-
alis, NEG negative, NOM nominalization, O object, P plural, PERF per-
fect, PF predication focus, PFV perfective, POSS possessive, PROP
proper name, PST past, REM remote, S singular or subject, TA(M) tense-
aspect-(modality), V verb. Elements of examples which are focused on in
the discussion are bold-faced; question marks in the glossing indicate
elements whose function is unclear to me. My comments within quoted
material are given in square brackets.
1. Accounts within the generative framework generally assume some form of
constituent movement which I don't follow here (see Koopman (1984) on
the phenomenon in Kru languages as one representative example).
2. The elements which differ in their category status but are related to each
other historically are co-indexed by sub-script letters. The term “Other”
refers to postverbal participants other than direct objects.
3. Cf., for example Heine and Claudi’s (2001: 64–8) discussion of the
Mande case. Following Kastenholz (2003) and pace Claudi (1994), this
family shows very little internal evidence for a transition from an earlier
V-O to a modern O-V stage; rather the S-Auxiliary-O-V-Other pattern can
be reconstructed to the proto-language. A major reason for entertaining an
original V-O order and its later change to O-V by means of grammaticali-
zation seems to be Mande’s assumed genealogical affiliation with Niger-
Congo - a hypothesis that still awaits the presentation of convincing evi-
dence.
4. Benue-Congo is used in the modern sense of “New Benue-Congo” (see,
e.g. Bender-Samuel (ed.) 1989) without, however, any commitment to the
hypothesis that this is a genealogical entity.
5. Green and Igwe (1963: 167–71) distinguish these two non-finite verb
forms by means of the terms “participle” (with prefixal a-/e-) and “infini-
tive” (with prefixal i-/iB-).

References

Abraham, Roy C.
1951 The Idoma language. London: University of London Press.
106 Tom Güldemann

Adive, John R.
1989 The verbal piece of Ebira. (Publications in Linguistics 85). Dallas/
Arlington: Summer Institute of Linguistics and University of Texas.
Andersen, Torben
1984 Aspect and word order in Moru. Journal of African Languages and
Linguistics 6: 19–34.
Ayuninjam, Funwi F.
1998 A reference grammar of Mbili. Lanham/ New York/ London: Uni-
versity Press of America.
Bendor-Samuel, John T. (ed.)
1989 The Niger-Congo languages. Lanham: University Press of America.
Bennett, Patrick R., and Jan P. Sterk
1977 South Central Niger-Congo: a reclassification. Studies in African
Linguistics 8: 241–273.
Bybee, Joan L., Revere Perkins, and William Pagliuca
1994 The evolution of grammar: tense, aspect, and modality in the lan-
guages of the world. Chicago: The University of Chicago Press.
Childs, G. Tucker
forth. It’s everywhere! The S-AUX-O-V-OTHER syntagm in Atlantic.
Studies in African Linguistics.
Chumbow, Beban, and Pius N. Tamanji
1994 Bafut. In Typological studies in negation, Peter Kahrel and René van
den Berg (eds.), 211–236. (Typological Studies in Language 29).
Amsterdam: John Benjamins.
Claudi, Ulrike
1993 Die Stellung von Verb und Objekt in Niger-Kongo-Sprachen: Ein
Beitrag zur Rekonstruktion historischer Syntax. (Afrikanistische
Monographien 1.) Köln: Institut für Afrikanistik, Universität zu
Köln.
1994 Word order change as category change: the Mande case. In Perspec-
tives on grammaticalization, William Pagliuca (ed.), 191–231. (Cur-
rent Issues in Linguistic Theory 109). Amsterdam: John Benjamins.
Creissels, Denis
2000 Typology. In African languages: an introduction, Bernd Heine and
Derek Nurse (eds.), 231–258. Cambridge: Cambridge University
Press.
Dik, Simon C. [ed. by Kees Hengeveld]
1997 The theory of functional grammar, part 1: the structure of the clause.
Berlin/ New York: Mouton de Gruyter.
Déchaine, Rose-Marie, and Victor Manfredi (eds.)
1997 Object positions in Benue-Kwa: papers from a workshop at Leiden
University, June 1994. (Holland Institute of Generative Linguistics
Publications 4). The Hague: Holland Academic Graphics.
 Preverbal objects and information structure in Benue-Congo 107

Elders, Stefan
2004 Distributed predicative syntax in Doyayo: constituent order alterna-
tions and cliticization. In Proceedings of the 4th World Congress of
African Linguistics, New Brunswick 2003, Akinbiyi Akinlabi, and
Oluseye Adesola (eds.), 189–197. Köln: Rüdiger Köppe.
Gensler, Orin D.
1994 On reconstructing the syntagm S-Aux-O-V-Other to Proto-Niger-
Congo. Berkeley Linguistics Society (special session) 20: 1–20.
1997 Grammaticalization, typology, and Niger-Congo word order: pro-
gress on a still-unsolved problem. Journal of African Languages and
Linguistics 18 (1): 57–93.
Gensler, Orin D., and Tom Güldemann
2003 S-Aux-O-V-Other in Africa: typological and areal perspective. Paper
presented at the 4th World Congress of African Linguistics, Work-
shop “Distributed predicative syntax (S P O V X)”, Rutgers Univer-
sity, June 21, 2003.
George, Isaac
1971 The á-construction in Nupe: perfective, stative, causative, or instru-
mental? In Papers in African Linguistics, Chin-Wu Kim and Herbert
Stahlke (eds.), 81–100. (Current Inquiry in Language and Linguistics
1). Edmonton/ Champaign: Linguistic Research, Inc.
Givón, Talmy
1975 Serial verbs and syntactic change. In Word order and word order
change, Charles N. Li (ed.), 149–188. Austin/ London: University of
Texas Press.
1978 Negation in language: pragmatics, function, ontology. In Pragma-
tics, Peter Cole (ed.), 69–112. (Syntax and Semantics 9). New York/
San Francisco/ London: Academic Press.
Good, Jeffrey
2004 When arguments become adjuncts: split configurationality in
Leggbò. Unpublished manuscript: University of California Berkeley.
2005 A fixed-position focus construction in Naki. Paper presented at the
International Conference on Focus in African Languages, ZAS Ber-
lin, October 6–8, 2005.
Green, Margaret Mackeson, and Georgewill Egemba Igwe
1963 A descriptive grammar of Igbo. (Veröffentlichungen des Instituts für
Orientforschung 53). Berlin: Deutsche Akademie der Wissen-
schaften.
Güldemann, Tom
1996 Verbalmorphologie und Nebenprädikationen im Bantu: Eine Studie
zur funktional motivierten Genese eines konjugationalen Subsystems.
(Bochum-Essener Beiträge zur Sprachwandelforschung 27). Bo-
chum: Universitätsverlag Dr. N. Brockmeyer.
108 Tom Güldemann

1999 The genesis of verbal negation in Bantu and its dependency on func-
tional features of clause types. In Bantu historical linguistics: theo-
retical and empirical perspectives, Jean-Marie Hombert and Larry
M. Hyman (eds.), 545–587. (CSLI Lecture Notes 99). Stanford: Cen-
ter for the Study of Language and Information (CSLI).
2003 Present progressive vis-à-vis predication focus in Bantu: a verbal
category between semantics and pragmatics. Studies in Language 27
(2): 323–361.
forth. The Macro-Sudan belt: towards identifying a linguistic area in north-
ern Sub-Saharan Africa. In Africa as a linguistic area, Bernd Heine
and Derek Nurse (eds.).
Heine, Bernd
1980 Language typology and linguistic reconstruction: the Niger-Congo
case. Journal of African Languages and Linguistics 2: 95–112.
Heine, Bernd, and Mechthild Reh
1983 Diachronic observations on completive focus marking in some Afri-
can languages. Sprache und Geschichte in Afrika 5: 7–44.
Heine, Bernd, and Ulrike Claudi
2001 On split word order: explaining syntactic variation. General Linguis-
tics 38 (1): 41–74.
Heine, Bernd, and Derek Nurse (eds.)
2000 African languages: an introduction. Cambridge: Cambridge Univer-
sity Press.
Horn, Laurence R.
1989 A natural history of negation. Chicago: Chicago University Press.
Hyman, Larry M., and Daniel J. Magaji
1971 Essentials of Gwari grammar. (Occasional Publications 27). Ibadan:
Institute of African Studies, University of Ibadan.
Hyman, Larry M., and John R. Watters
1984 Auxiliary focus. Studies in African Linguistics 15 (3): 233–273.
Ikoro, Suanu M.
1996 The Kana language. (CNWS Publications 40). Leiden: Research
School CNWS, Leiden University.
Iwara, Alexander
1982 Phonology and grammar of Lokaa. M.A. thesis: University of Lon-
don.
Kandybowicz, Jason, and Mark C. Baker
2003 On directionality and the structure of the verb phrase: evidence from
Nupe. Syntax 6 (2): 115–155.
Kastenholz, Raimund
2003 Auxiliaries, grammaticalization, and word order in Mande. Journal
of African Languages and Linguistuics 24 (1): 31–53.
 Preverbal objects and information structure in Benue-Congo 109

Kießling, Roland
1994 Eine Grammatik des Burunge. (Afrikanistische Forschungen 13).
Hamburg: Research and Progress.
Koopman, Hilda J.
1984 The syntax of verbs: from verb movement rules in the Kru languages
to universal grammar. Dordrecht: Foris.
Lefebvre, Claire, and Anne-Marie Brousseau
2002 A grammar of Fongbe. (Mouton Grammar Library 25) Berlin:
Mouton de Gruyter.
Lord, Carol
1982 The development of object markers in serial verb languages. In Stud-
ies in transitivity, Paul J. Hopper and Sandra A. Thompson (eds.),
277–299. (Syntax and Semantics 15). New York/ San Francisco/
London: Academic Press.
Manfredi, Victor
1997 Aspectual licensing and object shift. In Object positions in Benue-
Kwa: papers from a workshop at Leiden University, June 1994,
Rose-Marie Déchaine and Victor Manfredi (eds.), 87–122. (Holland
Institute of Generative Linguistics Publications 4). The Hague: Hol-
land Academic Graphics.
Marchese, Lynell
1983 On assertive focus and the inherent focus nature of negatives and
imperatives: evidence from Kru. Journal of African Languages and
Linguistics 5 (2): 115–129.
Meeussen, Achille E.
1967 Bantu grammatical reconstructions. In Africana Linguistica 3, 79–
121. (Annalen Wetenschappen van de Mens 61). Tervuren: Ko-
ninklijk Museum voor Midden-Afrika.
Mithun, Marianne
1984 The evolution of noun incorporation. Language 60 (4): 847–894.
Mous, Marten
1993 A grammar of Iraqw. (Kuschitische Sprachstudien 9). Hamburg:
Helmut Buske.
1997 The position of the object in Tunen. In Object positions in Benue-
Kwa: papers from a workshop at Leiden University, June 1994,
Rose-Marie Déchaine and Victor Manfredi (eds.), 123–137. (Holland
Institute of Generative Linguistics Publications 4). The Hague: Hol-
land Academic Graphics.
2005 The innovative character of object-verb word order in Nen (Bantu
A44, Cameroon). In Studies in African comparative linguistics with
special focus on Bantu and Mande, Koen Bostoen and Jacky Ma-
niacky (eds.), 411–424. (Collectie Menswetenschappen 169). Ter-
vuren: Royal Museum for Central Africa.
110 Tom Güldemann

Nkemnji, Michael
1995 Heavy pied-piping in Nweh. Ph.D. thesis: Department of Linguistics,
University of California Los Angeles.
Perrin, Mona J.
1994 Rheme and focus in Mambila. In Discourse features of ten lan-
guages of West-Central Africa, Stephen H. Levinsohn (ed.), 231–
241. (SIL Publications 109). Arlington: University of Texas.
Redden, James E.
1979 A descriptive grammar of Ewondo. (Occasional Papers on Linguis-
tics 4). Carbondale: Department of Linguistics, Southern Illinois
University.
Stanley, Carol
1991 Description morpho-syntaxique da la langue tikar (parlée au
Cameroun). Société International de Linguistique.
Stewart, John M.
1963 Some restrictions on objects in Twi. Journal of African Languages 2:
145–149.
1976 Towards Volta-Congo reconstruction (inaugural lecture). Leiden:
Leiden University Press.
Tamanji, Pius N.
2002 Negation, verb movement and word order in Bafut. Journal of West
African Languages 29 (1): 45–64.
Thwing, Rhonda, and John R. Watters
1987 Focus in Vute. Journal of African Languages and Linguistics 9 (2):
95–121.
Unseth, Pete
1986 Word order shift in negative sentences of Surma languages. Afri-
kanistische Arbeitspapiere 5: 135–143.
Urua, Eno E.
1997 Object movement in Eastern Lower-Cross. In Object positions in
Benue-Kwa: papers from a workshop at Leiden University, June
1994, Rose-Marie Déchaine and Victor Manfredi (eds.), 189–206.
(Holland Institute of Generative Linguistics Publications 4). The
Hague: Holland Academic Graphics.
Watson, Richard
1997 OV order in Ma’di? Occasional Papers in Sudanic Linguistics 7,
103–113. Nairobi: Summer Institute of Linguistics.
Watters, John R.
1979 Focus in Aghem. In Aghem grammatical structure, Larry M. Hyman
(ed.), 137–197. (Southern California Occasional Papers in Linguis-
tics 7). Los Angeles: Department of Linguistics, University of
Southern California.
 Preverbal objects and information structure in Benue-Congo 111

Williamson, Kay, and Roger Blench

2000 Niger-Congo. In African languages: an introduction, Bernd Heine
and Derek Nurse (eds.), 11–42. Cambridge: Cambridge University
Press.
Wolff, Ekkehard, and Ludwig Gerhardt
1977 Interferenzen zwischen Benue-Kongo- und Tschad-Sprachen. In
Zeitschrift der Deutschen Morgenländischen Gesellschaft, Supple-
ment 3: 1518–1543.
Focus strategies and the incremental development
of semantic representations: Evidence from Bantu

Lutz Marten

Abstract

Bantu languages, such as Nsenga, Swahili, and Tumbuka, exhibit word-order

variation associated with specific discourse-pragmatic contexts, such as topicaliz-
ing or focusing, both at the left and at the right periphery while expressing the
same semantic or truth-conditional content. By employing the formal architecture
of Dynamic Syntax (Cann et al. 2005), the paper proposes an analysis of topic and
focus constructions that reflects the distinction between discourse-pragmatic func-
tion and truth-conditional meaning. Topic and focus interpretations are purely
pragmatic notions, which arise from the combination of contextual information and
independently available syntactic processes such as *Adjunction and LINK struc-
tures. The analysis thus dispenses with syntactically primitive notions of topic and
focus.

1. Introduction

In this paper, I discuss different word orders found in different Bantu lan-
guages which are associated with specific discourse-pragmatic contexts,
such as topicalizing or focusing a particular constituent, both at the left and
at the right periphery, but which express the same semantic or truth-
conditional content. I will argue that the distinction between discourse-
pragmatic function and truth-conditional meaning is important and should
be reflected in the syntactic analysis of topic and focus constructions. In
particular, I will show that Dynamic Syntax (Kempson et al. 2001, Cann et
al. 2005), which models how hearers build semantic representations from
the time-linear string of words encountered in context, provides the tools to
express this distinction formally. One consequence of the Dynamic Syntax
analysis is that notions like ‘topic’ and ‘focus’ are indeed pragmatic notions
but are not part of the syntactic vocabulary of natural language (Kempson
et al. 2004). Rather, syntactic configurations can be exploited for the ex-
114 Lutz Marten

pression of topic and focus without the need to postulate dedicated syntac-
tic topic and focus projections, as is often done in Principles and Parame-
ters approaches to syntax (e.g. Rizzi 1997), or as primitive predicates in a
feature structure matrix, as in Lexical Functional Grammar (e.g. Bresnan
and Mchombo 1987). In this, I hope that the analysis presented in this paper
raises an alternative to current analyses of topic and focus which will con-
tribute to a better understanding of these notions.
The paper is organized as follows: in Section 2, a short background dis-
cussion of the relevant notions and approaches to word order, topic and
focus is provided, and a brief introduction to the tools of Dynamic Syntax
is given. Section 3 is dedicated to the discussion of presentational focus,
identificational focus, and background topics at the right periphery, while
Section 4 looks at focus and topic at the left periphery. Section 5 presents
conclusions from the analysis presented and indicates problems and direc-
tions for further research.

2. Background

Before starting the discussion of word order and information structure in

Bantu, a short background discussion is provided in this section about the
position of the analysis I am going to develop with respect to wider work
on topic and focus, as well as of the model of Dynamic Syntax and the
tools which are relevant in the present context.

2.1. Syntax and information structure

Information structure has become a central topic in syntax in the last two
decades or so. That discourse-pragmatic functions like topic and focus play
an important role for word order is especially clear when looking at Bantu
languages. For example, Bresnan and Mchombo (1987) observe for
Chichewa, that, in the presence of an object clitic functioning as an incor-
porated pronoun, all permutations of S, O, and V of a transitive clause are
possible: 1

(1) a. Njûchi zi-ná-wá-lum-a a-lenje [Chichewa]

10.bees SC10-PAST-OC2-bite-FV 2-hunters
‘The bees bit them, the hunters.’
b. Zináwáluma alenje njûchi (VOS)
 Focus strategies and semantic representations: Evidence from Bantu 115

c. Alenje zináwáluma njûchi (OVS)

d. Zináwáluma njûchi alenje (VSO)
e. Njûchi alenje zináwáluma (SOV)
f. Alenje njûchi zináwáluma (OSV)

Bresnan and Mchombo’s (1987) main interest was to show that subject and
object clitics in Chichewa can be analyzed as incorporated pronouns. 2 And
while they discuss the relation of subject and object clitics with topicalized
full NPs quite extensively, the discussion of focus is restricted to a few
remarks. However, from their discussion it is clear that they assume that
NPs can be co-indexed with focus and topic predicates at f-structure and
that topics may stand in an anaphoric agreement relation with the subject
and object clitics. Similar observations about word order freedom in Xhosa
are found in du Plessis and Visser (1992: 13), who also draw attention to
the relation between different word orders and prosodic information, a
point which is further discussed in Downing et al. (2004, 2005), who, again
using Chichewa, show that different word orders are related to specific
patterns of phonological phrasing. However, no fully worked-out analysis
of patterns like the one illustrated in (1) has yet been proposed. In the fol-
lowing sections, I will outline an analysis of different word orders using the
Dynamic Syntax tools of LINK and *Adjunction and show how these relate
different word orders to the context in which they are felicitous and to their
possible discourse-pragmatic functions. Before doing this, however, I give
a brief introduction to Dynamic Syntax in the next section.

2.2. Dynamic Syntax

Dynamic Syntax (Kempson et al. 2001, Cann et al. 2005) models the proc-
ess by which hearers construct semantic representations of content from
words in context as a model of linguistic competence. Semantic representa-
tions, or ‘logical forms’, corresponding to the hearer’s representation of
what she thinks is the intended message of the speaker are formally given
in the model as annotated trees which transparently show the predicate
argument structure of the proposition: 3
116 Lutz Marten

(2) Daudi likes Muna.

Tn(0), Ty(t), Fo(like’(muna’)(daudi’)), ¸

Fo(daudi’), Ty(e) Ty(e o t), Fo(like’(muna’))

Fo(muna’), Ty(e) Fo(like’), Ty(e o (e o t))

The tree shows that verb phrase interpretation is a function of the interpre-
tation of the main predicate and the interpretation of the object and that the
interpretation of the sentence, the main proposition at the top node Tn(0), is
a function of applying the interpretation of the subject to the interpretation
of the VP. Note, however, that the trees do not show natural language syn-
tax but purely semantic composition: the tree-nodes are decorated with
semantic information from words but not by the words directly, and the
structure of the tree only reflects semantic composition, e.g. predicate-
argument structure, without projecting word order as part of the tree. The
fact that the predicates are on the right-hand side branches and arguments
on the left-hand side branches is a matter of convention and not related to
natural language syntax and word order. In fact, the whole point of the Dy-
namic Syntax enterprise is that natural language syntax reflects the way
humans are able to build complex semantic structures like the one in (2)
from a linear string of sounds. 4 Generalizations about syntax, like word
order, grammaticality, and well-formedness, are expressed through the
process of tree growth from a minimal tree as a starting point through a
succession of partial trees to a fully annotated logical form: the dynamics of
the system lie in the incremental mapping from linearly-ordered words to
structured semantic representations, during which trees ‘grow’ as a result of
syntactic transition rules or lexical actions. 5
For illustration, here is a sample derivation for the string in (3):

(3) Daudi likes Muna.

At the outset of the derivation, a minimal tree is assumed with just one
node and no branches. The node is annotated with Tn(0), indicating that it
is the root node, and ?Ty(t), indicating that at this node a requirement holds
for an expression of type Ty(t), i.e. a proposition. This expresses hearers’
 Focus strategies and semantic representations: Evidence from Bantu 117

justified expectation for information of propositional type which may inter-

act with currently held assumptions (cf. Sperber and Wilson 1995). Finally,
the diamond indicates that the node is the current node (somewhat trivially,
since there is only one node so far):

(4) Tn(0), ?Ty(t), ¸

However, since in most cases not the whole proposition is communicated at

once, syntactic rules license the introduction of subtasks in this situation:

(5) Tn(0), ?Ty(t)

?Ty(e), ¸ ?Ty(e o t)

The rationale behind this move is that the satisfaction of both subtasks
leads automatically to the satisfaction of the overall goal and that the sub-
tasks may be easier to accomplish than the task at the root node. The fact
that the argument node (on the left-hand side) becomes the current node,
rather than the predicate node on the right, is a parametric value of SVO
languages. 6
At this stage, the first lexical information is scanned, namely informa-
tion from Daudi. Lexical information in Dynamic Syntax is modeled as
procedural and as directly interacting with the tree annotations:

(6) ‘Daudi’ IF ?Ty(e)

THEN put(Fo(daudi’), Ty(e))
ELSE abort

The IF statement in the lexical information from Daudi states that the word
can be introduced into the derivation if there is a current node with a re-
quirement ?Ty(e). If this is so, then at that node ‘Fo(daudi’)’ and ‘Ty(e)’
can be added. On the other hand, if Daudi is parsed when the current node
does not have a requirement for Ty(e), the parse ends. In the case at hand,
the condition of the IF clause is met, as the current node in (5) has a re-
quirement for Ty(e), and the tree can be further developed:
118 Lutz Marten

(7) Tn(0), ?Ty(t)

Fo(daudi’), Ty(e) ?Ty(e o t), ¸

The next step is the expectation of the development of the predicate node
and the parsing of likes (ignoring tense and agreement for the moment):

(8) ‘like’ IF ?Ty(e o t)

THEN make(<p1>), put(Fo(like’), Ty(e o (e o t))),
go(n), make(<p0>), go(<p0>), put(?Ty(e))
ELSE abort

The information from like shows that lexical information does not only
decorate existing nodes but may also build new nodes. The actions of the
THEN clause result in the building of a new predicate node and a corre-
sponding argument node with a requirement ?Ty(e), which becomes the
current node:

(9) Tn(0), ?Ty(t)

Fo(daudi’), Ty(e) ?Ty(e o t)

?Ty(e), ¸ Fo(like’), Ty(e o (e o t))

The next word is Muna, which comes with lexical information similar to
that of Daudi, and can fulfill the requirement at the current node:

(10) Tn(0), ?Ty(t)

Fo(daudi’), Ty(e) ?Ty(e o t)

Fo(muna’), Ty(e), ¸ Fo(like’), Ty(e o (e o t))

 Focus strategies and semantic representations: Evidence from Bantu 119

In the final tree of the derivation, all information established during the
parse is accumulated, and all requirements are fulfilled:

(11) Daudi likes Muna.

Tn(0), Ty(t), Fo(like’(muna’)(daudi’)), ¸

Fo(daudi’), Ty(e) Ty(e o t), Fo(like’(muna’))

Fo(muna’), Ty(e) Fo(like’), Ty(e o (e o t))

The tree is identical to the one in (2), corresponding to the logical form
associated with the string Daudi likes Muna.

2.3. *Adjunction and LINK structures

So far, the tree development, i.e. the order of transitions, matched the SVO
word order of our example (recall that the arrangement of the nodes in the
tree does not reflect word order). However, this is not always the case, as
there is word order variation both between and within languages. For ex-
ample, information may be presented early or late, or in other words, at the
left or the right periphery. One way of modeling this is by employing struc-
turally underspecified tree relations established by introducing unfixed
nodes and dominated only by the root node by *Adjunction (named after
the Kleene* operation, the reflexive-transitive closure over tree nodes):

(12) Muna, …
?Ty(t), ¸

<n*>Ty(t), Ty(e),
Fo(muna’)

In (12), the information from Muna is projected onto an unfixed node,

which means that the node will be part of the eventual tree but that, at the
time of the introduction of the word, it is not yet clear at which position it
120 Lutz Marten

will be exactly. *Adjunction can be employed for the analysis of clause-

initial wh-words or for fronted NPs such as Muna in (13):

(13) Muna, Daudi likes.

After the introduction of the information from Muna at an unfixed node, the
tree will be developed as usual. The unfixed node will remain part of the
tree until a suitable stage in the derivation can be found at which the infor-
mation at the unfixed node can be incorporated into the tree (which has to
be found within the current tree). For example, a requirement might be
introduced at the object node from the verb where the unfixed node can
merge:

(14) Muna, Daudi likes.

?Ty(t)

Ty(e), ?Ty(e o t)
<n*>Ty(t), Ty(e), Fo(daudi’)
Fo(muna’)
?Ty(e), ¸ Fo(like’),
Ty(e o (e o t))

The eventual tree for (13) will be identical to the trees in (2) and (11), ex-
pressing the fact that, semantically, (3) and (13) are identical. On the other
hand, the two analyses differ, not in the result but in the steps of transitions
which were involved in deriving the final tree. It is this difference in deri-
vation which expresses the pragmatic and information structure differences
between the two examples, such that in (14), for example, *Adjunction can
be used to express focus on the preposed NP. I will discuss this point in
more detail below in relation to *Adjunction and the corresponding Late
*Adjunction, which can be used to introduce information late.
A second mechanism for introducing information outside of canonical
position is called LINK transition. It allows for the building of two parallel
trees, linked by a shared term, in which one tree can be exploited to provide
a particular context for the other. LINK structures are used in Dynamic
 Focus strategies and semantic representations: Evidence from Bantu 121

Syntax for example for the analysis of relative clauses, conjunction and
topic constructions:

(15) Muna, Daudi likes her.

In examples like this one, the initial term is projected onto a linked tree
which provides the background for the main tree to be developed. For-
mally, the tree to be developed carries a requirement that the formula value
of the linked term be part of the new tree (?<p*>Fo(muna’)), thus ensuring
that the new tree is built within the context set up by the LINK structure
and that a term is shared across the LINK structure. Note that the LINK
transition does not provide a copy of the formula value but merely intro-
duces a requirement that such a copy should be part of the tree to be built,
and so the presence of this information in the tree following the LINKed
node has to be ensured by some other means, for example by the presence
of a ‘resumptive’ pronoun:

(16) Muna, …
Fo(muna’), Ty(e)
LINK

Tn(0), ?Ty(t), ?<p*>Fo(muna’), ¸

The main tree is then built as previously, but this time the object position is
filled by a pronoun:

(17) Muna, Daudi likes her.

Fo(muna’), Ty(e)
LINK

Tn(0), ?Ty(t), ?<p*>Fo(muna’)

Ty(e), ?Ty(e o t)
Fo(daudi’)

Ty(e), Fo(U), Fo(like’),

?x(Fo(x), Ty(e o (e o t))
Female(x)), ¸
122 Lutz Marten

The lexical specification of the pronoun is the same for both its ordinary
use and the ‘resumptive’ use as in this example: it specifies its type and
introduces an underspecified formula value with the metavariable U and the
requirement that the metavariable be enriched to a full formula value (re-
stricted by Female(x) to be female, following the gender specification in
her). 7 This enrichment is a pragmatic process and may involve formula
values from the context. The reason for the ‘resumptive’ interpretation is
that any choice other than Fo(muna’) as formula value in object position
will not satisfy the requirement at the root node that Fo(muna’) be part of
the tree. However, once Fo(muna’) is chosen as the formula value, the tree
will end up as being identical to (2) and (11) (except for the linked ‘topic’
node) – again, indicating that all of the examples (3), (13), and (15) are
semantically identical but different in pragmatic meaning, which is ex-
pressed as a function of the interaction between the different transitions
involved in the establishment of the final trees and the context. It is a gen-
eral feature of Dynamic Syntax that grammaticality (and felicity) are de-
termined by the set of transitions, and not by the final tree alone. With these
tools at hand, I will now return to word order and information structure in
Bantu, starting with focus (and topic) at the right periphery.

3. The expression of focus and topic at the right periphery

The tendency to place new or focused information late in the utterance has
often been observed, and examples of this tendency can also be found in
Bantu languages, for example in presentational and identificational focus
constructions. In presentational focus constructions in Nsenga, for example,
the unmarked SV order is reversed, and the subject is introduced after the
verb: 8

(18) a. À-léndò à-fwík-à [Nsenga]

2-visitors SC2.PAST-arrive-FV
‘(The) guests have arrived.’
b. À-fwík-à à-léndò [Nsenga]
SC2.PAST-arrive-FV 2-visitors
‘There have arrived (the) guests.’

VS order expressing presentational focus is freely available in Nsenga and

is not restricted to a subset of verbs (such as unaccusative verbs); it is also
 Focus strategies and semantic representations: Evidence from Bantu 123

found with transitive verbs, with the subject following the verb and the
object:

(19) a. Kàtíshà ó-wéléng-à má-bùkù [Nsenga]

Katisha SC1.PAST-read-FV 6-books
‘Katisha was reading books.’
b. Ó-wéléng-à má-bùkù Kàtíshà [Nsenga]
SC1.PAST-read-FV 6-booksKatisha
‘Katisha was reading books.’ (‘There is reading books Katisha.’)

Inversion structures like (18b) and (19b) are part of a number of construc-
tions expressing new information or emphatic focus (also including locative
inversion and subject-object reversal) and are used to emphasize the new-
ness of (the referent of) the subject in the discourse. An example is pro-
vided in (20) from Swahili (from Bearth 1995: 198), where the two new
participants of the story are introduced through the use of VS structures:

(20) ... watu wa-ka-pand-a fiwi. Ka-j-a

... people SC2-CONS-plant-FV beans SC1.PERF-come-FV
Talafa... A-ka-toke-a Mzee Mgomba ... [Swahili]
Tafala… SC1-CONS-appear-FV Mzee Mgomba …
‘... the people were planting beans.’ ... ‘There came Talafa.’ ...
‘Then arrived Mzee Mgomba.’ ...

The analysis of these structures involves the interplay between the pro-
nominal nature of Bantu subject clitics and structural underspecification, in
particular Late *Adjunction. Following Bresnan and Mchombo (1987), I
assume that subject clitics may function like pronouns, encoding a meta-
variable as formula value which can be updated from context (as well as
restrictions as to the permissible kinds of substitutions, e.g. for class 2, that
the substituend be human (Human(x)) and a group (PL(x)), cf. Cann et al.
2005, Marten and Kempson 2002 for further discussion of the analysis of
Bantu agreement in DS). Thus, if it is clear that we are talking about guests,
no overt subject is necessary:

(21) À-fwík-à [Nsenga]

SC2.past-arrive-FV
‘They (i.e. the guests) have arrived.’
124 Lutz Marten

VS inversion structures can be used for focussing the post-verbal subject as

well as in ‘afterthought’ topic constructions, discussed further below. VS
structures involving focus arise in situations where it is not clear from the
context who the intended referent of the subject clitic is. There may be
various reasons for this. For example, the referent may be discourse new, as
in (20), or may be provided as a response to a clarification question, giving
rise to identificational focus (example (26) below). These different uses are
a function of the particular context and receive the same structural, syntac-
tic analysis in Dynamic Syntax, namely that the meta-variable projected
from the subject clitic cannot be assigned a value, and the task to provide a
suitable formula value remains open even after the predicate node is built:

(22) À-fwík-à .... (‘they came’)

Tn(0), Tns(Past), ?Ty(t)

‘??????’

Ty(e), Fo(U) Ty(e o t), Fo(fwík’)

?x(Fo(x), PL(x),
Human(x)), ¸

The lack of an appropriate referent from the context (indicated by the ques-
tion marks in the bubble) means that the parse cannot be completed after
the introduction of the verb and that further information is necessary. This
information is provided by the post-verbal subject, which is introduced into
the parse by Late *Adjunction, i.e., the application of *Adjunction not at
the outset but at the final stage of the parse when all nodes are type-
complete, but with an outstanding requirement for a formula value only.

(23) À-fwík-à (‘they came ...’)

 Focus strategies and semantic representations: Evidence from Bantu 125

Tn(0), Tns(Past), ?Ty(t)

Tn(00), Ty(e), Fo(U), Ty(e o t), Fo(fwík’)

?x(Fo(x), PL(x),
Human(x)), ¸

<n*>Tn(00), ?Ty(e), ¸

Like *Adjunction, Late *Adjunction allows the building of an unfixed

node, but this time the unfixed node is dominated by the argument node
with the outstanding requirement (‘Tn(00)’). It is decorated with a require-
ment for a Ty(e) expression, a requirement which the information from
àléndò can fulfill (for reasons of space, only the unfixed node is shown):

(24)

<n*>Tn(00), Fo(àléndò’), Ty(e), ¸

As a final step, the unfixed node is merged at subject position, and the final
tree has all requirements completed:

(25) À-fwík-à à-léndò (‘They came the visitors.’)

Tn(0), Tns(Past), Ty(t), Fo(fwík’(àléndò’)), ¸

Ty(e), Fo(àléndò’) Ty(e o t), Fo(fwík’)

The final tree of both the SV and the VS orders is identical, reflecting their
identical predicate-argument structure. However, the intermediate trees
leading to the derivation of the final tree differ, and the difference between
the two versions lies only in the steps which have been taken to reach it.
The claim is that, semantically, the two utterances in (18) (as well as those
in (19)) are identical, and this is reflected in the identical final trees, but
126 Lutz Marten

that they differ in pragmatic felicity. In particular, the VS order in the focus
examples here works only in a context where the subject clitic cannot be
fully interpreted from the context, and the postverbal subject is focused.
The use of *Adjunction for VS structures means that the left and the
right periphery are analyzed by the same structural means and that asym-
metries between left and right periphery are a function of the incremental
nature of structure building and context (cf. Cann et al. 2004). The advan-
tage of this analysis is that it distinguishes between structural aspects of the
left and right periphery, which are modelled as uniform, and the asymmetry
between the two peripheries, which is explained in the present analysis as
resulting from the difference in contextual information available at the out-
set and at the end of the parse.
As mentioned above, the same structural analysis can be given to post-
verbal identificational focus, as in the following question-answer pair from
Chichewa. Downing et al. (2005), from which the example is taken, com-
ment on this that there is identificational focus on àlèéndó in the answer, as
it gives a choice from a known list of possibilities: 9

(26) Q: À-ná-m-dyéts-á nsóòmbà ndàáni ? [Chichewa]

SC2-PAST-OC1-eat.CAUS-FV fish who
‘Who fed him fish?’
A: À-ná-m-dyèts-à nsóòmbà àlèéndó
SC2-PAST-OC1-eat.CAUS-FV fish 2-guest
‘The visitors fed him fish.’

The structural analysis in Dynamic Syntax for examples like these is identi-
cal to the analysis of presentational focus: the subject at the right periphery
is introduced by Late *Adjunction. What is different in (26) from (20) is
that a set of potential referents is available in the context, giving rise to an
identificational reading.
Notice that the analysis provides a formal, syntactic reflex of an obser-
vation often made in the literature on focus, namely that the focused ele-
ment provides a value to an open proposition or gives rise to ‘alternative’
propositions (e.g. Lambrecht 1994, Rooth 1996, Bearth 1999). In the
*Adjunction analysis, there is transparently a stage in the derivation at
which the open proposition Ȝx[fwík’(x)] is entertained and a stage at which
Ȝx[fwík’(x)](àléndò’) is entertained. Yet, semantically, in terms of truth
conditions, both SV and VS end up as Fo(fwík’(àléndò’)). This analysis
differs from analyses with designated focus (and topic) projections (as in
 Focus strategies and semantic representations: Evidence from Bantu 127

P&P) or with primitive TOP and FOC attributes (as in LFG). The Dynamic
Syntax claim is that structural configurations (such as late unfixed nodes)
can be exploited for specific pragmatic effects, arising in the process of tree
construction in a given context, but that information theoretic notions are
not part of the eventual (semantic) representation (LF) or any other repre-
sentational level (such as LF’, proposed by Vallduví 1990). Information
structure and propositional structure can thus be seen as intertwined but
distinct aspects of structure building in natural language.
It is worth pointing out that there is another instance of VS order, and
that is so-called afterthought constructions:

(27) À-fwík-à, à-léndò [Nsenga]

SC2.PAST-arrive-FV 2-guests
‘They have arrived, the guests.’

(28) Q: À-ná-m-tàání nyàání à-lééndó ? [Chichewa]

SC2-PAST-OC1-do.Q baboon 2-visitors
‘What did the visitors do to the baboon?’
A: À-ná-m-dyèts-à nsóòmbà ! à-lèéndó
SC2-PAST-OC1-eat.CAUS-FV fish 2-guests
‘The visitors fed him fish.’ (Downing et al. 2005)

Afterthought constructions differ from focus constructions in their prosody,

as they involve an intonational break between verb and subject (27) or are
marked by tonal downstep (indicated by ‘!’) as in (28), and in their prag-
matic felicity, as here the referent of the subject is not introduced as new
into the context but is given as a reminder or clarification of the context; it
is thus more (background) topical than focused. As the referent is not dis-
course new, the difference to the focus constructions discussed above is
that here the subject clitic is interpreted from the context and that the post-
posed subject provides information to ensure that it is interpreted correctly.
The emerging tree is thus different, in that, in afterthought constructions,
the subject clitic is completed and the post-verbal subject needs to match
the interpretation assigned to it; these constructions are thus in a sense the
inverse of hanging topic constructions, discussed above, where a LINK
relation ensures the identity of the left-dislocated topic and the ‘resumptive’
pronoun. Taking this parallelism as a starting point, the Dynamic Syntax
analysis of afterthought topic constructions does not involve Late
*Adjunction but the building of a LINK structure after the proposition has
128 Lutz Marten

been built, with the requirement that the formula value of the linked node,
i.e. Fo(àléndò’) in (27), be part of the proposition, which is fulfilled if the
hearer has picked the right referent from context when interpreting the sub-
ject clitic.
Like the use of *Adjunction, LINK structures provide a single structural
means for introducing information at the left and right periphery, either as
providing a context for an assertion yet to be developed or as clarifying an
intended context as an afterthought. But these different functions are solely
the result of contextual information available at the different stages in struc-
ture building reached when the linked node is built, and do not need to be
stipulated as part of the structural specification of the LINK transition, as
will also be seen when looking at the left periphery in the next section.

4. The expression of focus and topic at the left periphery

In addition to the focus (and topic) effects on the right periphery, informa-
tion structure can be expressed at the outset of the parse, at the left periph-
ery, involving *Adjunction and initial LINK structures. However, there is a
further point to be raised with respect to the left periphery, and that is the
role of subjects and subject clitics in Bantu. The pronominal nature of
Bantu subject clitics seems to indicate an analysis where subject clitics are
projected directly onto the subject node of the emergent tree structure, or as
locally unfixed nodes (Marten and Kempson 2002, Marten 2005). In either
case, this would allow for the option that overt subject NPs are in general
introduced at linked or unfixed nodes, an assumption supported by the
word order flexibility of Bantu languages illustrated earlier. But this im-
plies that the initial unfixed node is not available for other constituents
when it is filled by the subject (as there can only be one unfixed node at
any one time), and, furthermore, that pragmatic effects are reduced since
the introduction of unfixed nodes is a standard strategy for introducing
subjects. Still, initial focus examples can be found (Downing 2005: 6): 10

(29) Ngô:ma ti-zamu-limilir-a namach!ê:ro [Tumbuka]

9.maize SC1PL-FUT-weed-FV tomorrow
‘Maize we will weed tomorrow.’
 Focus strategies and semantic representations: Evidence from Bantu 129

(30) Ma-bû:ku wa-ka-p!ás-a !wâ:na [Tumbuka]

6-books SC2-PAST-give-FV children
‘They gave the children books.’

On the example in (29), Downing comments that maize is being contrasted

with some other possible crop and on (30) that it is an answer to ‘What did
they give to the children?’, indicative of contrastive and new information
focus. Furthermore, prosodic evidence (the initial NP is in a separate pho-
nological phrase) as well as the absence of an object clitic (otherwise
strongly preferred with left-dislocated objects) support the status of the
construction as focus related. In the Dynamic Syntax analysis, the initial
constituent is projected onto an unfixed node by *Adjunction, as shown in
the introduction. It is interesting that in both examples no overt subject is
expressed, leaving the unfixed node to be used by the focused object.
Another example of initial focus is the answer in (31) from Chichewa:

(31) Q: Ndàání á-ná-m-dyéts-á nsóòmbà [Chichewa]

Who SC1-PAST-OC1-eat.CAUS-FV fish
‘Who fed him fish?’
A: À-lèéndó à-ná-m-dyèts-à nsóòmbà
2-guests SC2-PAST-OC1-eat.CAUS-FV fish
‘The visitors fed him fish.’ (Downing et al. 2005)

The subject àlèéndó is in its own phonological phrase, indicating new in-
formation focus, as is also indicated by the context, and contrasts with the
subject àlèndó in (32) through the absence of the high tone on the penulti-
mate syllable and penultimate lengthening, the two criteria Downing et al.
(2005) identify as indicative of phonological phrase boundaries: 11

(32) Q: Kù-nà-chítík-á chìyáànì ? [Chichewa]

SC17-PAST-happen-FV what
‘What happened?’
A: À-lèndó à-ná-m-dyèts-à nsóòmbà
2-guests SC2-PAST-OC1-eat.CAUS-FV fish
‘The visitors fed him fish.’

That the use of initial *Adjunction is possible for the expression of fo-
cus is further supported by examples with multiple focus, where both
*Adjunction and Late *Adjunction are employed in the same structure:
130 Lutz Marten

(33) Pa-mu-pâ:nda zi-ka-d!úk-a mb!û:zi [Tumbuka]

16-3-wall SC10-PAST-jump-FV 10.goats
‘Over the wall jumped goats.’ (‘The goats jumped over the wall.’)
(And something else jumped over something else.)
(Downing 2005: 7)

In (33), both the locative NP pa-mu-pâ:nda and the subject mb!û:zi are
contrasted with something else, are prosodically marked as constituting
separate phonological phrases, and are hence, according to Downing
(2005), focused. Both NPs are dislocated, at least under the reasonable
assumption that the locative phrase is part of the VP, and so, in Dynamic
Syntax terms, (33) can be analyzed as involving *Adjunction for the loca-
tive NP, which merges with a fixed node supplied by the predicate, and
Late *Adjunction for the subject, which is introduced as unfixed with re-
spect to the subject node which lacks a full interpretation from the context.
The examples discussed here show that *Adjunction can be used for the
expression of focus in Bantu, even though, for the reasons outlined at the
outset of this section, other strategies are more common.
One of these other strategies is, in fact, to employ LINK structures,
more usually associated with topichood, to introduce new information:

(34) Q: Ba-ntfwana, ba-ba-nik-e-ni ? [Swati]

2-children SC2-OC2-give-PAST-what
‘What did they give to the children?’
A: Tin-cwadzi, ba-ti-nike ba-ntfwana
10-books SC2-OC10-give-PAST 2-children
‘Books, they gave (them) to the children.’

Both the question and the answer in (34) have a left dislocated NP which is
co-referential with an object clitic. Bantfwana in the question appears to be
a discourse topic and may as such be analyzed as linked to the main tree,
where the information from the linked node is introduced through the ob-
ject clitic -ba-. However, tincwadzi in the answer provides pragmatically
new information, but it is co-referenced with an object clitic in the verb,
which is commonly analyzed as being cross-referenced to topics, not to
focused elements. From the Dynamic Syntax perspective, the structure may
be analyzed as involving either a LINK structure, implying that there is no
one-to-one correspondence between structural operations and pragmatic
function, or as an unfixed node under the assumption that object clitics in
 Focus strategies and semantic representations: Evidence from Bantu 131

Swati can be merged with information from full NPs. This latter option
appears to be more in line with the argument presented here, but a full dis-
cussion will have to await a more in-depth Dynamic Syntax study of the
function of object clitics in Swati, and in Bantu more widely.

5. Conclusion

In this paper, I have shown how different structural possibilities at the right
and left periphery are exploited for the expression of focus (and topic) in
Bantu. On a theoretical level, my main concern was to show that the model
does not assign any syntactic meaning to pragmatic notions like topic and
focus, in contrast to most alternative analyses. From the Dynamic Syntax
perspective, pragmatic effects arise from the particular building steps in-
volved in constructing the semantic representation associated with the ut-
terance but are not reflected in the final representation itself, thus providing
a formal reflex of the distinction between propositional, semantic structure
and pragmatic information structure. Furthermore, an important part of the
Dynamic Syntax analysis is the way in which utterances are tied to the con-
text, as identical structural analyses can represent different pragmatic read-
ings (e.g. the difference between identificational and presentational focus)
as a function of different contexts. From a functional perspective, this
might be seen as an instance of the versatility of natural language, which
expresses an infinite range of meanings by limited structural means.
On the Bantu side, there are a number of questions outstanding, which
have to be addressed on another occasion, partly due to reasons of space,
and partly due to the fact that the relevant Dynamic Syntax analyses are
still in progress. Amongst those, two in particular deserve a brief mention.
First, all structures I have discussed show the expression of focus at the
clausal periphery. However, it is well known that the immediate postverbal
position in Bantu is associated with focus (e.g. Bearth 1999). The extension
of the Dynamic Syntax analysis to such examples presupposes a Dynamic
Syntax analysis of the Bantu VP, which has yet to be fully developed. 12
Second, with respect to presentational focus structures, another set of
data needs to be mentioned, and that is presentational focus constructions
with locative subject clitics:
132 Lutz Marten

(35) a. Kwà-fwík-á à-léndò [Nsenga]

SC17.PAST-arrive-FVconj 2-visitors
‘There have arrived guests.’
b. *A-lendo kwa-fwik-a [Nsenga]
2-visitors SC17.PAST-arrive-FV
Intd.: ‘There have arrived guests.’

By assumption, the postverbal subject is introduced, like in the cases dis-

cussed in section 3, by Late *Adjunction. But the function of the subject
clitic is more difficult to analyze, as it may be taken as a full locative clitic,
as an expletive element, or as something in between the two. In addition,
the verb form in (35) is marked as ‘conjoint’ by the high tone on the final
vowel, in contrast to verbs in VS order with subject agreement as discussed
here, presumably indicating that the following NP is part of the core clause.
I will leave a more detailed discussion of examples like (35) to a future
occasion, as it would require more space than is available here.
To conclude, I hope to have shown the interest and relevance to ques-
tions of information structure of the new framework of Dynamic Syntax,
and of data from Bantu, which, with about 400 different languages, remains
a valuable and largely untapped resource for linguistic theory formation.

Notes

I would like to thank Laura Downing, Ruth Kempson, Nancy Kula, Al

Mtenje, Laura Mutti, Clara Simanga, Nhlanhla Thwala and two reviewers
for comments and discussion of the points made in the article. Swati ex-
amples were provided by Nhlanhla Thwala. Nsenga examples were pro-
vided by Clara Simanga during a research visit at the Centre of Language
Studies of the University of Malawi in April 2004, whose hospitality is
hereby gratefully acknowledged. Parts of this research were supported by
AHRB award B/RG/AN8675/APN16312.
1. The following abbreviations are used in the glosses: SC = subject con-
cord; OC = object concord; FV = final vowel; CONS = consecutive tense;
PERF = perfect; CAUS = causative; FUT = future; Q = question particle.
Numbers refer to noun classes.
2. More precisely, Bresnan and Mchombo argue that the object clitic is al-
ways an incorporated pronoun while the subject clitic can also function as
a marker of grammatical agreement. Discussion of this point would lead
too far afield, but see e.g. Demuth and Johnson (1989), Marten and
Kempson (2002).
 Focus strategies and semantic representations: Evidence from Bantu 133

3. Tree annotations in this example are: Tn – treenode identifier, Ty – logical

type, Fo – formula value, ¸ (diamond) – indicating the current node. The
type value ‘e’ stands for ‘entity’ and corresponds roughly to NPs, ‘t’
stands for ‘truth-evaluable’, that is, a proposition, corresponding to a sen-
tence, and combinations of the two indicate functions, e.g. ‘e o t’, a func-
tion from an entity to a proposition, corresponding to VP interpretation, or
to intransitive verbs. More annotation will be introduced in the course of
the discussion. The type value is similar to the types used in type-logical
grammar, and, correspondingly, formula values can be thought of as
lambda terms. Like in Montague or Categorial Grammar (e.g. Steedman
2000), logical form is the only level of (syntactic and semantic) represen-
tation.
4. Or, more correctly, Dynamic Syntax is concerned with how information
from the suitably phonologically parsed input-string is built up into larger
structures. The parsing perspective adopted here is shared with models
like Government Phonology (Kaye 1989) in phonology and with several
pragmatic theories, in particular Relevance Theory (Sperber and Wilson
1995).
5. Note that, although the perspective adopted is related to parsing, the Dy-
namic Syntax model is not a parsing model in the traditional sense, as cru-
cially, Dynamic Syntax does not presuppose an independently defined
competence model. Rather, the claim is that the dynamics of building se-
mantic representations is all there is to syntax.
6. For more discussion of pointer movement see Cann et al. (2005).
7. In addition, locality restrictions on interpretation take care of binding ef-
fects, but I ignore those here; see Cann et al. (2005).
8. The Nsenga past tense marker à fuses with the preceding subject concord
as follows: SC2 à + à = à; SC1 ú + à = ó; SC17 kù + à = kwà.
9. Chichewa examples are tone marked as in the source. In the following ex-
amples (as well as in examples (28) and (31)), there appear to be tonal dif-
ferences between the questions and answers. Although Downing et al.
(2005) do not comment on this, the difference may be related to the ex-
pression of the different clause types.
10. It is sometimes claimed that initial focused constituents are impossible in
Bantu since Bantu languages express focus on the right periphery. How-
ever, the Dynamic Syntax analysis seems to be preferable in that it allows
initially focussed constituents, but also has something to say about why
they are rare.
11. I am grateful to Al Mtenje for making this point clear to me.
12. But see Marten (2002) for some Dynamic Syntax discussion of verb
phrase structure and Bantu applicatives.
134 Lutz Marten

References

Bearth, Thomas
1995 Wortstellung, Topik und Fokus. In Swahili - Handbuch, Gudrun
Miehe and Wilhelm J.G. Möhlig (eds.), 173–205. Köln: Köppe.
1999 The contribution of African linguistics towards a general theory of
focus. Update and critical review. Journal of African Languages and
Linguistics 20: 121–156.
Bresnan, Joan, and Sam Mchombo
1987 Topic, pronoun, and agreement in Chichewa. Language 63: 741–
782.
Cann, Ronnie, Ruth Kempson, Lutz Marten, Masayuki Otsuka, and David Swin-
burne
2004 On the left and on the right. In Peripheries: Syntactic Edges and
their Effects, David Adger, Cécile de Cat, and George Tsoulos (eds.),
19–47. Dordrecht: Kluwer.
Cann, Ronnie, Ruth Kempson, and Lutz Marten
2005 The Dynamics of Language: An Introduction. Oxford: Elsevier.
Demuth, Katherine, and Mark Johnson
1989 Interaction between discourse functions and agreement in Setawana.
Journal of African Languages and Linguistics 11: 21–35.
Downing, Laura
2005 The prosody of some focus-related enclitics in some Southern Bantu
languages. Paper presented at the 5th Bantu Grammar: Description
and Theory Meeting, SOAS. Handout available on-line at
http://mercury.soas.ac.uk/users/lm5/bantu_project.htm.
Downing, Laura, Al Mtenje, and Bernd Pompino-Marschall
2004 Prosody and information structure in Chichewa. ZAS Papers in Lin-
guistics 37: 167–186. Available on-line at http://mercury.soas.ac.uk/
users/lm5/bantu_project.htm.
2005 Non-accentual prosodic cues to focus in a tone language: the case of
Ntcheu Chichewa. Paper presented at Between Tone and Stress,
University of Leiden, 16–18 June 2005.
Du Plessis, J.A., and M. Visser
1992 Xhosa Syntax. Pretoria: Via Afrika.
Kaye, Jonathan
1989 Phonology: A Cognitive View. Hillsdale, NJ: Lawrence Erlbaum.
Kempson, Ruth, Ronnie Cann, and Jieun Kiaer
2004 Topic, focus and the structural dynamics of language. Ms. Available
on-line at http://semantics.phil.kcl.ac.uk/ldsnl/papers/.
Kempson, Ruth, Dov Gabbay, and Wilfried Meyer-Viol
2001 Dynamic Syntax: The Flow of Natural Language Understanding.
Oxford: Blackwell.
 Focus strategies and semantic representations: Evidence from Bantu 135

Lambrecht, Knut
1994 Information Structure and Sentence Form. Cambridge: Cambridge
University Press.
Marten, Lutz
2002 At the Syntax-Pragmatics Interface: Verbal Underspecification and
Concept Formation in Dynamic Syntax. Oxford: OUP.
2005 Passive, locative inversion and subject-object reversal. Ms. SOAS.
Marten, Lutz, and Ruth Kempson
2002 Pronouns, agreement, and dynamic construction of verb phrase in-
terpretation: A Dynamic Syntax approach to Bantu clause structure.
Linguistic Analysis 32: 471–504.
Rizzi, Luigi
1997 The fine structure of the left periphery. In Elements of Grammar,
Liliane Haegeman (ed.), 281–337. Dordrecht: Kluwer.
Rooth, Mats
1996 Focus. In The Handbook of Contemporary Semantic Theory, Shalom
Lappin (ed.), 271–297. Oxford: Blackwell.
Sperber, Dan, and Deirdre Wilson
1995 Relevance: Communication and Cognition. 2nd ed. Oxford: Black-
well.
Steedman, Mark
2000 The Syntactic Process, Cambridge, Mass.: MIT Press.
Vallduví, Enric
1990 The Information Component. Unpublished PhD dissertation, Univer-
sity of Pennsylvania.
Part III

Ex-situ and in-situ strategies of focus

marking
Ex-situ focus in Kikuyu

Florian Schwarz

Abstract

This paper discusses ex-situ focus constructions in Kikuyu, focusing on the particle
ne which thereby plays a crucial role. The two analyses of ne proposed in the lit-
erature, the focus phrase analysis (Clements 1984; Schwarz 2003) and the cleft
analysis (Bergvall 1987), are compared in detail. I argue that the focus phrase
analysis is more successful in accounting for a number of central properties of
focus constructions with ne. Among other things, it accounts for focus projection
and the relation between in-situ and ex-situ focus. At the same time, these points
constitute serious problems for the cleft analysis.

1. Introduction 1

Kikuyu 2 has in-situ and ex-situ focus constructions and wh-constructions.

Both the focus and the wh-constructions exhibit the same syntactic pattern.
This chapter is primarily concerned with the syntactic analysis of ex-situ
focus constructions.
A crucial player in the ex-situ focus constructions is the particle ne. It
appears in a number of places, namely before the focused phrase (or the
fronted wh-phrase), in simple copula constructions, and in the immediately
preverbal position in certain declarative sentences. Accounting for its dis-
tribution and function is a central task for any analysis of the ex-situ focus
and wh-constructions. Much of this chapter is therefore concerned with the
syntactic analysis of the ne-constructions. The two analyses that have been
proposed to account for the distribution of ne are the focus phrase analysis
(Clements 1984; Schwarz 2003) and the cleft analysis (Bergvall 1987). By
comparing the two analyses, I argue that the focus phrase analysis, though
not without problems, is by far more promising than the cleft analysis,
which faces a number of serious problems.
The chapter is organized as follows: Section 2 provides a brief overview
of the crucial properties of ne. Section 3 presents the focus phrase analysis
and further data supporting it, concerning focus projection and the relation
140 Florian Schwarz

between in-situ and ex-situ focus. The cleft analysis and the problems it
faces are discussed in section 4. Section 5 gives a conclusion.

2. Properties of ne

The particle ne appears in three environments: in ex-situ focus and wh-

constructions, in simple copula constructions, and preverbally in regular
declarative sentences. Examples of these are presented in section 2.1. Ne is
also subject to some crucial distributional restrictions, which are discussed
in section 2.2.

2.1. Environments in which ne appears

An example of the first environment in which ne appears, namely ex-situ

focus and wh-constructions, is given in (1): 3

(1) a. PG³ MGG #DFWN C³ TC³ P[W³ KT³ '?

FM- what A. SM- T- drink- ASP- FV
‘What did Abdul drink?’
b. PG OCG #DFWN C³ TC³ P[W³ KT³ '
FM 6.water A. SM- T- drink- ASP- FV
‘Abdul drank WATER.’

In the ex-situ question in (1a), the focus marker ne combines with the ques-
tion word kee in the sentence initial position. The sentence in (1b) is a pos-
sible answer to this question, and here the object mae ‘water’, which is
focused due to the preceding question 4 , appears in the same position as the
question word in (1a), adjacent to ne.
All questions, except for subject questions, also have an in-situ version,
which does not contain ne, as shown in (2):

(2) a. #DFWN C³ TC³ P[W³ KT³ ' MGG?

A. SM- T- drink- ASP- FV what
‘What did Abdul drink?’
b. #DFWN C³ TC³ P[W³ KT³ ' OCG
A. SM- T- drink- ASP- FV 6.water
‘Abdul drank WATER.’
 Ex-situ focus in Kikuyu 141

The discussion in this paper will mostly concern the ex-situ cases, but the
relationship between in-situ and ex-situ constructions will be relevant in the
discussion of focus projection in section 3.3.
One important point about the ex-situ constructions is that ne can be
preceded by other material (both in focus and wh-constructions), as was
first noted by Schwarz (2003). Examples of this are given in (3) where a
topicalized subject (3a) and a topicalized adverbial clause (3b) appear be-
fore the fronted object with ne: 5

(3) a. #DFWN PG OCG C³ TC³ P[W³ KT³ '

A. FM 6.water SM- T- drink- ASP- FV
‘Abdul drank WATER.’
b. OD'T' [C P[QODCPG OCG #DFWN C³TC³ P[W³KT³ '
 in-front 9.A 9.house FM 6.water A. SM-T- drink-ASP- FV
‘In front of the house, Abdul drank WATER.’

These examples will play a crucial role in the argument for the focus phrase
analysis developed in the following section.
The second construction involving ne is that of a simple copula clause.
An example is given in (4a):

(4) a. CDFWN *(PG) Ø OQ³ TWVCPK

A. FM COP 1- teacher
‘Abdul is a teacher.’
 b. CDFWN (PG) C³ C³ TG OQ³TWVCPK
 A. FM SM- T- be 1- teacher
‘Abdul was a teacher.’

The obligatory presence of ne in (4a) might suggest that ne itself can func-
tion as the copula. However, once we consider cases that are not in the third
person present tense form, ne is no longer obligatorily present (although it
still can precede the copula verb), as can be seen in (4b). Instead, the cop-
ula verb stem re appears with the usual inflectional morphology. 6 The
analysis commonly adopted for this pattern is that the underlying form of
(4a) contains a phonologically null form of the copula verb, as indicated in
(4a) (cf. Bergvall 1987, Clements 1984, Schwarz 2003). A central question
in this respect is why ne is obligatory when the verb is phonologically null.
I propose a tentative answer to this at the end of section 3.3.
142 Florian Schwarz

According to this analysis, ne appears in the immediately preverbal po-

sition in copula constructions. These are then a special case of the last envi-
ronment in which ne appears, namely preverbally in regular declarative and
interrogative sentences (which only differ in intonation), as is illustrated in
(5):

(5) a.#DFWN (PG)³ C³ TC³ P[W³ KT³ ' OCG

A. FM- SM- T- drink- ASP- FV 6.water
‘Abdul drank water.’
b.#DFWN (PG)³ C³ TC³ P[W³ KT³ ' OCG?
A. FM- SM- T- drink- ASP- FV 6.water
‘Did Abdul drink water?’

While ne is not obligatory in these sentences as far as the syntax is con-

cerned, leaving it out changes the interpretation with respect to what is in
focus. I will discuss this in more detail in section 3.3.

2.2. Distributional restrictions of ne

There are several restrictions on the distribution of ne: it can only appear
once per clause, its distribution in embedded clauses is limited, and it can-
not co-occur with the regular verbal negation marker ti. The first point is
illustrated in (6):

(6) a. *PG OCG #DFWN PG³ C³ TC³ P[W³ KT³ '
FM 6.water A. FM- SM- T- drink- ASP- FV
‘Abdul drank water.’

This sentence is fine when ne only appears in one of the two positions, but
ungrammatical as soon as it appears in both.
The second restriction is that ne cannot appear in certain embedded
clauses. For example, the sentence in (7a) becomes ungrammatical when ne
is added in the relative clause (see (7b)):

(7) a.OQ³TWVCPK Q³ &QO³KT³ ' K³ $WMW

1- teacher SM- read- ASP- FV 5- book
 PG³ C³ TC³ P[W³ KT³ ' OCG
FM- SM- T- drink- ASP- FV 6.water
‘The teacher who read a book drank water.’
 Ex-situ focus in Kikuyu 143

b.* OQ³ TWVCPK PG³ Q³ &QO³ KT³ ' K³ $WMW
1- teacher FM- SM- read- ASP- FV 5- book
 PG³ C³ TC³ P[W³ KT³ ' OCG
FM- SM- T- drink- ASP- FV 6.water

Clauses that are embedded by a bridge verb (i.e. think, know, say, etc.), on
the other hand, do allow ne. This is not surprising since such clauses be-
have in many ways like matrix clauses.
The last restriction concerns co-occurrence with verbal negation. The
regular negation marker ti appears in the verbal complex between the sub-
ject marker and the tense marker. When ti is present in this position, insert-
ing ne into the sentence leads to ungrammaticality:

(8) a. #DFWN C³ VK³ TC³ P[W³ KT³ ' OCG
A. SM- NEG- T- drink- ASP- FV 6.water
‘Abdul didn’t drink water.’
b.* #DFWN PG³ C³ VK³ TC³ P[W³ KT³ ' OCG
A. FM- SM- NEG- T- drink- ASP- FV 6.water
Intended meaning: ‘Abdul didn’t drink water.’
c.*PG OCG #DFWN C³ VK³ TC³ P[W³ KT³ '
 FM 6.water A. SM- NEG- T- drink- ASP- FV
Intended meaning: ‘Abdul didn’t drink water.’

Note, however, that there is an alternative negation marker ta which can

occur on the main verb with ex-situ focus constructions. I will come back to
this when discussing the problems of the focus phrase analysis.

3. The focus phrase analysis

The main challenge posed by the data presented in the preceding section is
to account for the different types of occurrences of ne in a unified manner
while also making the correct predictions about its distributional restric-
tions. The focus phrase analysis deals with this challenge by assuming that
ne appears in a syntactic focus phrase within an extended CP-projection
(Brody 1990; É. Kiss 1998; Rizzi 1997). The position that ne appears in is
always the same then, and the different constructions involving ne are de-
rived by having different elements move to the focus phrase. The general
structure that this account is based on is the following:
144 Florian Schwarz

(9) FP
ei
SpecFP F'
 XPF ei
F YP
[+F] 6
… XPF …

A strong feature in the head of the focus phrase triggers the movement of
an XP bearing a focus feature to the specifier of the focus phrase. There are
two slightly different possible theoretical implementations with respect to
ne. First, ne might be the head of the focus phrase (a common assumption
for focus markers). We then have to say that it cliticizes onto the material in
its specifier to get the right word order, namely ne XP (cf. Muriungi 2004
for a proposal along these lines for the closely related language Kitharaka).
Alternatively, we could say that the focus feature on the XP gets spelled out
as ne when it appears in the specifier of the focus phrase. Most of the fol-
lowing is compatible with either of these accounts, and I will simplify the
representations by putting ne XP in the specifier of the focus phrase.

3.1. Accounting for the different occurrences of ne

Based on the idea that ne appears in a syntactic focus phrase, how can we
account for the different constructions involving ne in detail? Let us first
turn to the ex-situ focus construction. Take the example of (3b) above, re-
peated here as (10a).

(10) a. #DFWN PG OCG C³ TC³ P[W³ KT³ '

A. FM 6.water SM- T- drink- ASP- FV
‘Abdul drank WATER.’

The fronting of the object mae ‘water’ can be captured by moving it to the
specifier of the focus phrase. Furthermore, the subject abdul is topicalized
so that it occurs in the sentence initial position. We can then represent the
derivation (with many simplifications) as follows: 7
 Ex-situ focus in Kikuyu 145

(10) b. CP
ei
SpecCP FP
 #DFWN ei
SpecFP IP
PGOCG ei
SpecIP VP
#DFWN ei
V DP
CTCP[WKT' OCG

When the subject remains in its base position, we get the ne-initial order
found in (1b). The case of preverbal ne is derived in a similar fashion. In
this case, the entire IP moves to the focus phrase, and the subject moves on
to the same topic position as in (10b). Moving the entire IP into the focus
phrase is motivated by the fact that immediately preverbal ne expresses
focus on the entire sentence (as will be discussed in more detail below).
The sentence in (5a), repeated below, can then be analyzed as in (11b) 8 :

(11) a.#DFWN PG³ C³ TC³ P[W³ KT³ ' OCG

A. FM- SM- T- drink- ASP- FV 6.water
‘Abdul drank water.’

b. CP
ei
SpecCP FP
 #DFWN q
SpecFP
PG+ IP IP
ei 6
SpecIP VP #DFWNCTCP[WKT'OCG
#DFWN ei
V DP
 CTCP[WKT' OCG
One difference between this case and the ex-situ focus case in (10) is that
the topicalization of the subject is obligatory here. This does not fall out of
the theory at this point. The only explanatory speculation that I can offer in
this respect at the moment is that if the subject was not topicalized, the
structure would be string identical to the ex-situ focus construction in (1b),
146 Florian Schwarz

and hence the formal marking of different foci would be less perspicu-
ous.
Assuming that copula constructions are a special case of preverbal ne,
their analysis will be as in (11b) with the only difference that the head of
the verb phrase is phonologically null in the third person singular case.

3.2. Accounting for the distributional restrictions

The preceding section has shown how the focus phrase analysis can ac-
count for the different occurrences of ne. Now we need to make sure that
we can also account for the distributional restrictions.
First, why is it that ne only appears once per clause? According to the
focus phrase analysis, this is simply because ne is tied to a particular syn-
tactic phrase which only appears once per clause.
The second question is why ne cannot appear in relative clauses. The
focus phrase assumed in the above analysis is part of the extended C-
system. There seems to be good evidence indicating that the fully extended
C-system is not present in relative clauses. For example, topicalization is
not possible in relative clauses either, which is just what we expect if the
topic position is also part of the extended C-system. The absence of ne in
relative clauses is then simply a reflection of the absence of the extended C-
system.
Finally, we have to explain why ne cannot co-occur with the regular ne-
gation marker. The answer to this question is not so obvious. One tempting
possibility might be to say that ne and the negation ti appear in the same
syntactic position. However, this is hard to reconcile with their surface
distribution (see (8a)). Another possibility is to say that negation is some-
how inherently linked to focus (cf. Hyman 1999), but it is unclear how to
spell this out in detail in the present framework. Perhaps one option would
be to say that the negative head is capable of checking the focus feature on
the fronted element and thus make the focus phrase unnecessary. Unfortu-
nately, I cannot explore this option in more detail here. So, for present pur-
poses, it must suffice to say that while a more detailed answer has to be
developed by further research, there is no reason to believe that this issue
poses a problem that is particular to the focus phrase analysis.
 Ex-situ focus in Kikuyu 147

3.3. Focus projection: In-situ vs. ex-situ focus

Up to this point, I have only discussed simple focus constructions where the
focused object appears ex-situ. However, looking at a larger variety of fo-
cus constructions, including in-situ focus and cases of focus projection,
lends further support to the analysis developed here. Furthermore, these
cases distinguish the focus phrase analysis from the cleft analysis, which
cannot account for the facts presented here, as will be discussed in the next
section.
According to the focus phrase analysis, ne marks focus, and we expect
to find complex patterns with respect to what exactly is in focus semanti-
cally given a particular formal marking of focus, just as we find such cases
of so-called focus projection in pitch accent languages. Let us start with
cases where the entire verb phrase is focused. Assuming, as above, that we
can force a particular focus structure on a declarative sentence by putting it
in the context of a question, the following question answer pair illustrates a
case of VP-focus:

(12) a. PG³ CVGC CDFWN GM³ KT³ '?

FM- what/how A. (SM)-do- ASP- FV
‘What did Abdul do?’
b. PG OCG CDFWN C³ P[W³ KT³ '
FM 6.water A. SM - drink- ASP- FV
‘Abdul [drank WATER]F.’

Formally marking the object for focus is apparently sufficient for focusing
the entire verb phrase semantically. This is exactly the type of focus projec-
tion we find for pitch-accent languages like English. The same is true for
in-situ focus; i.e. the sentence in (2b), where the object is focused in-situ,
could also express focus on the verb phrase, e.g. as an answer to an in-situ
version of the question in (12a). 10
Next, let us turn to sentence focus, which is what we find in so called
out of the blue-contexts or as answers to questions like What happened?

(13) [Context: Abdul drank non-purified water and got sick. A just got
back and wants to know from B what happened.]
A: PG³ MGG MG³ QTW?
FM- what CL- bad
‘What’s wrong?’ or ‘What happened?’ [literally: ‘What is bad?’]
148 Florian Schwarz

B:#DFWN PG³ C³ TC³ P[W³ KT³ ' OCG

A. FM- SM- T- drink- ASP- FV 6.water
‘Abdul drank water.’

In this context, we find ne in the immediately preverbal position marking

sentence focus. The fact that the entire IP is in focus here motivated the
analysis in (11b) where the IP moves to the focus phrase. Note that neither
in-situ nor ex-situ focus marking of the object (as in (1b) and (2b)) can
express sentence focus. This is different from English where focus projects
from the stressed object all the way up to the sentence level.
Apart from marking sentence focus, preverbal ne can also express a par-
ticular emphasis on the truth of a statement (i.e. verum focus), as is fre-
quently discussed in traditional grammars of Kikuyu. The following exam-
ple (adopted with a few changes from Armstrong (1940:297)) illustrates
this use of preverbal ne:

(14) A: ‘Where did you put it?’

B: ‘I put it in the granary.’
A: ‘I didn’t see it there – are you sure?’
B: PG³ P³ FC³ K¢³ C K³ MQQODG
 FM- SM- put- ASP- FV CL- granary
‘I DID put it in the granary.’

Sentence focus and verum focus are the only two functions that preverbal
ne has, contrary to claims made in the literature that it could express focus
on the predicate, i.e. verb or verb phrase focus (Güldemann 1996). 11 In
order to express narrow focus on the verb, the verb has to appear in its in-
finitival form in the ex-situ position with ne as well as in its base position in
the inflected form:

(15) A: CDFWN PG³ C³ &GM³ KT³ '?

A. FM- SM- laugh- ASP- FV
‘Did Abdul laugh?’
B: C5C CDFWN PG³ MQ³ TGT³ C C³ TGT³ KT³ '
no A. FM- SM(INF)- cry- FV SM- cry- ASP- FV
‘No. Abdul CRIED.’

Although a full analysis of this has to await another occasion, one might be
able to account for the two occurrences of the verb within the copy-theory
 Ex-situ focus in Kikuyu 149

of movement. 12 The sentence in (15) would then be an example of a situa-

tion where both copies are pronounced, presumably because after morpho-
logical reanalysis the higher copy of the verb becomes invisible to Kayne’s
LCA and, therefore, to deletion (Nunes 2004). 13
In summary, focus marking on the object (either in sentences without ne
and with the object in-situ or in sentences with ex-situ focus and ne) can
express focus on the object or on the verb phrase, and ne in preverbal posi-
tion can express sentence focus or verum focus. While the technical details
of focus projection have to be worked out in future work, the fact that there
is focus projection in constructions with ne fits in very naturally with the
focus phrase analysis.
If we make just one additional assumption, namely that every sentence
has to have a focus, several further facts that would otherwise be surprising
(in particular from the viewpoint of the cleft analysis, as will be discussed
in the next section) fall out of the theory.
First, a seemingly odd fact about Kikuyu is that “when a sentence con-
sists of an affirmative finite verb only (e.g. nƭokire, he came), nƭ [= ne; FS]
is indispensable” (Barlow (1951: 34)). 14 Under the current analysis, this
follows because there is no object that could introduce any in-situ focus
marking, and hence, the only possibility for introducing a focus into the
sentence (apart from having ex-situ subject focus with ne) is to have pre-
verbal ne.
Second, there is an interaction between the availability of in-situ focus
and ne. The in-situ focus on the object (or the verb phrase) in (2b) is no
longer available if preverbal ne is introduced. This is, of course, exactly
what we expect if we analyze ne as a focus marker which triggers move-
ment of the focused constituent (either the object or the entire IP) to the
focus phrase.
Finally, we are able to explain the obligatory presence of ne in third per-
son present tense copula constructions if we make the additional assump-
tion that the focus feature on the focused XP in in-situ focus constructions
is in some way licensed by the lexical verbal head. It is commonly assumed
that phonologically null heads have limited licensing capacities. Since the
third person present tense form of the copula is phonologically null, it can-
not license in-situ focus on the object and hence, the only way to introduce
a focus in such copula sentences is to let ne do the job.
150 Florian Schwarz

3.4. Remaining problems

Although the focus phrase analysis makes promising predictions with re-
spect to both of the two different environments in which ne appears, its
distributional restrictions and the facts connected to focus projection dis-
cussed in the preceding subsection, there are two remaining problems. The
first concerns the details of the technical implementation of ne mentioned at
the beginning of this section. I have nothing more to say about this here.
The second problem concerns a number of morphological changes that
appear on the verb phrase when it is preceded by an ex-situ focus or wh-
construction: negation changes from ti to ta, the third person subject marker
changes from a to o, and the post-verbal downstep is deleted (for details,
see Clements 1984). These changes appear to be identical to the ones we
observe in relative clauses. This has been taken as support for the cleft
analysis since on that analysis we are in fact dealing with relative clauses
(see Bergvall 1987, and the discussion in the next section). However, this
argument is not as straightforward as it may seem at first sight. As Bergvall
herself points out (Bergvall 1987: 114), once we consider more complex
constructions involving multiple clauses, with the focused element originat-
ing in the lowest clause, these changes affect different domains: the subject
marker only changes in the lowest clause, negation only changes in the
highest embedded clause, and the tonal changes affect all embedded
clauses. These phenomena presumably are general effects of A'-movement,
which need an independent account. Therefore, they do not pose a problem
that is particular to the focus phrase analysis, and the competing cleft
analysis has to account for them independently as well.

4. The cleft analysis

I now sketch out the cleft analysis (Bergvall 1987) and discuss some of the
problems it faces. Its starting point is the occurrence of ne in copula con-
structions although it does not assume that ne is the copula. As above, cop-
ula constructions are taken to be a special case of the preverbal occurrence
of ne. The cleft analysis differs from the focus phrase analysis in the case of
ex-situ focus and wh-constructions. The latter are taken to be yet another
variant of the preverbal occurrence, where the verb is the phonologically
null form of the copula which is part of a cleft.
 Ex-situ focus in Kikuyu 151

Another difference between the two analyses lies in the role that is as-
signed to ne. Since cleft constructions have a well known impact on focus
structure, it is unnecessary to assign ne the role of a focus marker. What
role does ne play then? According to Bergvall’s cleft analysis, it is an asser-
tion marker that appears in the head of the IP. I will come back to this point
below after introducing the analysis in some more detail.

4.1. Accounting for the different occurrences of ne

With ne generated in the head of the IP (see (16b)), the immediately pre-
verbal cases can be straightforwardly accounted for. Assuming that the
subject appears in the specifier of the IP (or, alternatively, in the specifier
of CP), the word order of the preverbal case can be derived without any
difficulties:

(16) a. #DFWN PG³ C³ TC³ P[W³ KT³ ' OCG

A. FM- SM- T- drink- ASP- FV 6.water
‘Abdul drank water.’

b. IP
ei
SpecIP I'
 #DFWN ei
I VP
PG ei
V DP
CTCP[WKT' OCG

As in the focus phrase analysis, copula constructions are accounted for in

exactly the same way as the preverbal case by assuming that a phonologi-
cally null form of the copula appears as the head of the verb phrase.
Finally, ex-situ focus and wh-constructions are analyzed as bi-clausal,
with a phonologically null expletive subject and a phonologically null cop-
ula in the first clause. In a slightly simplified form, the analysis as proposed
by Bergvall then is as in (17b) (Bergvall 1987: 123):

(17) a. PG OCG #DFWN C³ TC³ P[W³ KT³ '

FM 6.water A. SM- T- drink- ASP- FV
‘Abdul drank WATER.’
152 Florian Schwarz

b. IP
ei
SpecIP I'
 G ei
I VP
PG tgp
V DP CP
Ø OCGK 6
Opi CDFWNCTCP[WKT' ti

The index and the operator in the lower CP link mae to its base position.
According to this analysis, all occurrences of ne are reduced to the prever-
bal case, and the focusing effect of the ex-situ construction is attributed to
the cleft construction. Crucially, this dissociates ne per se from focus.

4.2. Accounting for the distributional restrictions

Given the analysis sketched above, how does the cleft approach account for
the distributional restrictions? With respect to the limitation to one ne per
clause, it says that each independent clause can only make one assertion
(and hence can only contain one ne, which is taken to be a marker of asser-
tion). The absence of ne in embedded clauses, in particular in relative
clauses, is explained by the fact that relative clauses (at least restrictive
ones) are presupposed, and their content is therefore not part of what is
asserted. Finally, the complementary distribution of ne and the negation
marker ti is explained by saying that ti is a marker of assertion as well (with
opposite polarity) so that it would not make sense to have both, a positive
and a negative assertion marker, in one clause.

4.3. Problems

Note that all of the points concerning distributional restrictions rest on the
assumption that ne is a marker of assertion. This characterization of ne is
problematic, given that ne routinely occurs in questions and other types of
speech acts that are not assertions.
In addition to this issue concerning the role of ne, the cleft analysis faces
a number of further problems. First, it cannot account for topics preceding
ex-situ focus constructions. Second, it is hard to reconcile with the fact that
 Ex-situ focus in Kikuyu 153

ex-situ wh-constructions can be part of multiple wh-questions. Third, it

cannot account for certain cases of multiple ne’s in complex clauses.
Let us turn to the first point in more detail. Topicalized elements can
precede the ex-situ focus constructions with ne, as shown in (3) above. The
cleft analysis, however, which assumes that ex-situ constructions do in-
volve a relative clause, falsely predicts this to be impossible, given that
topicalization out of relative clauses is impossible, as the data in (18) illus-
trate for the PP topic in front of the house:

(18) a. P[KPC QQTGC Y³ nP³ KT³' K³$WMWOD'T' [C P[QODC

mother DEM SM- see- ASP-FV 5-book infront 9.A 9.house
‘the mother who saw the book in front of the house’
b. * Q\LQD RRUHD OD'T' [CP[QODCY³nP³KT³' K³$WMW
mother DEM infront 9.A 9.house SM-see-ASP-FV 5-book
c. * OD'T' [CP[QODCQ\LQD RRUHD Y³ nP³ KT³ ' K³$WMW
infront 9.A 9.house mother DEM SM- see- ASP-FV 5-book

Furthermore, topicalization beyond clause boundaries appears to be impos-

sible, as shown in (19), which speaks against any bi-clausal treatment of ex-
situ focus constructions.

(19) a. CDFWNPG- W¢- KT- ' CVG P[KPC

A. FM- say- ASP- FV that mother
PG- nnn- KT³ ' K³$WMWOD'T' [C P[QODC
FM- see- ASP- FV 5-book in-front 9.A 9.house`
‘Abdul said his mother saw the book in front of the house.’
b. DEGXO PG- W¢- KT- ' DWH OD'T' [C
A. FM- say- ASP- FV that infront 9.A
 P[QODC Q\LQD PG³ nnP³ KT³ ' K³$WMW
 9.house motherFM- see- ASP- FV 5-book

c.# OD'T' [C P[QODC CDFWN PG³ W¢³ KT³ '

infront 9.A 9.house A. FM- say- ASP- FV
 CVG P[KPC PG³ nnP³ KT³ ' K³$WMW
that mother FM- see- ASP- FV 5-book

In (19b), the prepositional phrase is fronted within the embedded clause,
and the meaning is the same as in (19a). The sentence in (19c), on the other
hand, can only be understood in such a way that ‘in front of the house’ is
154 Florian Schwarz

the location where the saying took place and not where the mother saw the
book. Given this restriction on topicalization, the cleft analysis makes false
predictions about topicalization in ex-situ focus constructions (cf. (3)).
The second problem concerns the fact that ex-situ wh-constructions with
ne can form part of a multiple wh-question, as illustrated in (20):

(20) PQQ Y³ 'PF³ KTK³ ' MGG?

FM.who SM- sell- ASP- FV what
‘Who sold what?’

This is problematic because, cross-linguistically, cleft based questions can

normally not form part of multiple wh-questions. 15 Unless we can find
good reason to believe that we are dealing with an exception here, this
point speaks against the cleft analysis.
Finally, there are constructions involving multiple clauses which can
contain multiple ne’s. In particular, Kikuyu allows for long-distance wh-
extraction out of certain clauses. When the wh-word is extracted from the
lowest clause, ne can appear both with the fronted wh-word and preverbally
in the lowest clause:

(21) a. PG³ MQ PIQ¢G C³ W¢³ KT' CVG MCOCW

FM- where N. SM- say- ASP that K.
 PG V n³ nP³ KT' MCPCM'
FM- SM- see- ASP kanake
‘Where did Ngoge say that Kamau saw Kanake ?’
[‘Where, according to Ngoge, did Kamau see Kanake?’]
b.* PIQ¢G C³ W¢³ KT' CVG PG³ MQ MCOCW
N. SM- say- ASP that FM- where K.
 PG V n³ nP³ KT' MCPCM'
FM- SM- see- ASP kanake

However, this is only possible if the wh-word moves all the way to the
highest clause, and not if it remains in the lower clause, as indicated in
(21b). This is problematic for the cleft analysis because it explains the re-
striction that ne can only occur once per clause in semantic terms by saying
that each clause can be marked for assertion only once (however this is to
be understood for the question cases; the same point would apply to the
corresponding case of focus fronting). But semantically, the wh-word be-
longs to the lower clause so that in this respect both ne’s should be counted
 Ex-situ focus in Kikuyu 155

as belonging to the same clause. Therefore, the cleft analysis falsely pre-
dicts (21a) to be ungrammatical.
Note that this construction is not at all problematic for the focus phrase
analysis. Assuming successive cyclic movement, we expect the wh-word to
move through the specifier of the focus phrase of the lower clause. Appar-
ently, moving the wh-word through this position is compatible with having
an overt ne in the lower focus phrase. 16 This analysis gains further support
from the fact that in the closely related language Kitharaka the morpheme
equivalent to ne is obligatorily present in the lower phrase, a fact which has
been taken as an argument for an analysis in terms of successive cyclic
movement by Muriungi (Muriungi 2004).
In addition to these major problems, the cleft analysis also cannot ac-
count for the facts about focus projection and the related issues in section
3.3. According to the cleft analysis, the focusing effect of ex-situ focus
constructions is due to the syntactic configuration of the cleft, and ne only
plays its general role as an assertion marker in these cases. Beyond such
special constructions that directly affect focus structure, we thus have no
reason to expect interactions between ne and focus on this account. But as
we saw above, the possibility of in-situ focus depends on the absence of ne.
I do not see how this can be accounted for if we assume that ne is an asser-
tion marker.
Concerning the fact that we find focus projection with the ex-situ focus
construction (namely focus on the verb phrase when the object is fronted
with ne), there also is a problem for the cleft analysis, as clefts typically do
not allow focus projection.
Finally, there does not seem to be a way to account for the seemingly
odd facts discussed at the end of section 3.3. Why is ne obligatory in sen-
tences that only consist of an intransitive verb? Surely not because these
always have to be emphatically marked for assertion, but that is all that the
cleft analysis could say about this. And why is ne obligatory with third
person present tense copula constructions? Again, the role that the cleft
analysis assigns to ne, namely that of a marker of assertion, does not pro-
vide any help in explaining this.
Taken together, these problems seem to provide a good case against the
cleft analysis. 17 Furthermore, as already mentioned above, it is unclear
what the status of ne on this analysis could reasonably be, given that the
assertion marker analysis is incompatible with its presence in questions and
other types of speech acts.
156 Florian Schwarz

5. Conclusion

I have discussed Kikuyu focus constructions involving ne and compared

two analyses, the focus phrase analysis and the cleft analysis. I have argued
that the former makes out better, both with respect to the syntax and the
semantics of the construction. Although there are plenty of questions for
further research, I hope to have convinced the reader that it is most promis-
ing to approach these from the viewpoint of the focus phrase analysis.

Notes

1. Parts of this paper have been presented at SOAS (London), ZAS (Berlin),
and at the workshop ‘Topic and Focus: Information Structure and Gram-
mar in African Languages’ (Amsterdam). I would like to thank all the par-
ticipants for helpful comments and discussion, and in particular Enoch
Aboh, Rajesh Bhatt, Lisa Cheng, Laura Downing, Katharina Hartmann,
Angelika Kratzer, Victor Manfredi, Yukiko Morimoto, Brigitte Reineke,
Anna Szabolcsi, Sabine Zerbian, and Malte Zimmermann. Special thanks
are due to Manfred Krifka for crucial guidance while I was working on
my Master’s thesis on which this paper is based. Special thanks are also
due to my Kikuyu consultant, Sam Kinuthia. I gratefully acknowledge
support for this research from the ZAS.
2. Kikuyu is an SVO Bantu language spoken in Kenya. Its label in Guthrie’s
(1967) classification system is E50.
3. All data has been elicited from my consultant, unless otherwise indicated.
The following abbreviations are used in the glosses: FM: focus marker,
SM: subject marker, T: tense, ASP: aspect, FV: final vowel, A: associa-
tive, COP: copula, NEG: negation, DEM: demonstrative. Numerals pre-
ceding nouns indicate the noun-class. Tones are not marked. Although the
more detailed study of tonal effects in relation to focus is an important
topic for further research, I believe that the syntactic points made in this
chapter hold independently of such possible effects. Details concerning
tense and aspect are omitted. See Johnson (1980) for a comprehensive
discussion of Kikuyu tense and aspect.
4. I assume throughout that focus can be reliably manipulated by different
question contexts. This is independent of the issue of whether a theory of
focus ultimately needs to make reference to question-answer correspon-
dence.
5. I use topicalization in a syntactic sense here, without making any direct
claims about its discourse properties. In syntactic terms, there is evidence
for an additional Topic Phrase between the CP and the FP, since the sen-
 Ex-situ focus in Kikuyu 157

tence in (11) could be embedded by a bridge verb and would then be pre-
ceded by the complementizer ate which presumably occurs in C0. There-
fore, the topicalized element cannot appear in the specifier of CP. This is,
of course, perfectly consistent with the standard analysis of the extended
left periphery (Rizzi 1997).
6. The stem re is actually ambiguous: Apart from the copula meaning, it also
has a possessive meaning (i.e. (4b) can also mean ‘Abdul has a teacher’).
The one place where the possessive and the copula paradigms diverge is
in the third person present tense form, where the null form unambiguously
has the copula meaning, whereas re only has the possessive meaning.
7. Movement is indicated by crossing out elements of a syntactic chain that
aren’t pronounced. The IP-level is ignored to keep things simple. The
topicalized subject is represented in the specifier of CP to keep things
simple. See footnote 5 on the need for a distinct topic phrase inside of the
CP.
8. One potentially problematic aspect of this analysis, pointed out to me by
Rajesh Bhatt, is that the apparent possibility of movement out of the
moved IP is somewhat unexpected. A possible alternative analysis would
leave the IP in its base position and have ne assign focus to it from the
head of FP.
9. Interestingly, however, the order FM-S-V-O can express sentence focus in
the closely related language Kitharaka (Muriungi 2004).
10. An interesting question that was pointed out to me by Katharina Hartmann
is why focus on the verb phrase cannot be expressed by moving the VP to
the focus phrase. Perhaps this is blocked by the alternative option of just
moving the object, which is more economical.
11. Again, there is an interesting contrast with Kitharaka, where the order S-
FM-V-O can express focus on the verb as well as sentence focus (Muri-
ungi 2004).
12. Thanks to Lisa Cheng for bringing this to my attention.
13. As one of the reviewers points out, more needs to be said about this. For
example, it is unclear, given this brief description, why the object in ex-
situ focus constructions is not pronounced in both positions.
14. I assume that this holds both for intransitive verbs and transitive verbs that
only have an object marker and no overt object noun phrase, but my data
on this are incomplete.
15. Thanks to Anna Szabolcsi for pointing this out to me.
16. This fact might speak in favor of the second analysis of ne above, which
assumes it to be generated in the head of the focus phrase and then cliti-
cizes it onto the XP in its specifier, since otherwise we would have the
feature on the focused XP spelled out twice.
17. Yet another problem that the cleft analysis probably has to face is the ab-
sence of tense in the cleft copula. I do not have the relevant data to make
158 Florian Schwarz

this point, but Bergvall’s discussion of the interpretation of tense in clefts

(Bergvall 1987: 130–132) suggests that there are no clefts with past tense
copulas, which is unexpected on her account (thanks to Rajesh Bhatt for
pointing out this issue to me).

References

Armstrong, Lilias E.
1940 The Phonetic and Tonal Structure of Kikuyu. London: Oxford Uni-
versity Press.
Barlow, Arthur Ruffell
1951 Studies in Kikuyu Grammar and Idiom. Edinburgh: Blackwood &
Sons.
Brody, Michael
1990 Some Remarks on the Focus Field in Hungarian. UCL Working Pa-
pers 2: 201–225.
Bergvall, Victoria Lee
1987 Focus in Kikuyu and universal grammar. Ph.D. diss., Harvard Uni-
versity.
Clements, George N.
1984 Binding Domains in Kikuyu. Studies in the Linguistic Sciences 14:
37–56.
É. Kiss, Katalin
1998 Identificational Focus versus Information Focus. Language 74: 245–
273.
Güldemann, Tom
1996 Verbalmorphologie und Nebenprädikation im Bantu. Eine Studie zur
funktional motivierten Genese eines konjungationalen Subsystems.
Bochum: Brockmeyer.
Guthrie, Malcolm
1967 The Classification of Bantu Languages. London: Dawsons.
Hyman, Larry M.
1999 The Interaction between Focus and Tone in Bantu. In The Grammar
of Focus, Georges Rebuschi and Laurice Tuller (eds.), 151–177.
Amsterdam, Philadelphia: John Benjamins.
Johnson, Marion R.
1980 A semantic description of temporal reference in the Kikuyu verb.
Studies in African Linguistics 11: 269–320.
 Ex-situ focus in Kikuyu 159

Muriungi, Peter
2004 Wh-movement in Kitharaka as focus movement. Paper presented at
the Workshop: Syntax, Semantics and Pragmatics of Questions at
ESSLLI 11.
Nunes, Jairo
2004 Linearization of Chains and Sideward Movement. Cambridge: MIT
Press.
Rizzi, Luigi
1997 The Fine Structure of the Left Periphery. In Elements of Grammar,
Liliane M. Haegeman (ed.), 281–337. Dordrecht: Kluwer Publica-
tions.
Schwarz, Florian
2003 Focus Marking in Kikuyu. In Questions and Focus, Regine Eckardt
(ed.), 41–118. ZASPIL 30. Berlin: ZAS.
Focus in the Force-Fin system: Information
structure in Cushitic languages

Mara Frascarelli and Annarita Puglielli

Abstract

This paper investigates the Force-Fin system in two Cushitic languages – Somali
and Afar – in which the interpretation of discourse grammar categories depends on
specific morphosyntactic conditions. In particular, the analysis deals with the acti-
vation of the “Focus field” and constitutes an argument for the assumption of a
[+foc] feature in the C-domain, playing a crucial role in the interpretation of differ-
ent focus-related categories. The relevant discussion also provides substantial sup-
port for a cartographic approach to Information Structure and shows the existence
of AGREE relations between (some) functional features in the left periphery of the
sentence. Finally, a crucial connection is shown between Focus and the illocution-
ary Force of the sentence and a structural distinction between matrix and embedded
C-domains is therefore proposed.

1. Introduction

This paper investigates information structure and the Force-Fin system (cf.
Rizzi 1997) in two Cushitic languages, Somali and Afar. In particular, the
activation of the focus field will be analyzed in different clausal types,
showing the existence of a crucial connection between focus-related cate-
gories and the illocutionary force of a sentence.

1.1. The fine architecture of the C-domain

In a cartographic approach (cf. Rizzi ed. 2004), the original CP-node (a

recursive phrase, targeted by different functional categories) has been re-
analyzed as an array of functional projections each dedicated to a specific
function related to information structure. The C-domain thus provides an
interface between the propositional content (the IP-node) and specific dis-
162 Mara Frascarelli and Annarita Puglielli

course roles. 1 In particular, Rizzi (1997) proposes that the left periphery of
the sentence is included between force and finiteness:
“Complementizers express the fact that a sentence is a question, a declara-
tive, an exclamative, a relative, a comparative, an adverbial of a certain
kind, etc., and can be selected as such by a higher selector. This information
is called the specification of Force […] the C-system expresses a specifica-
tion of finiteness, which in turn selects an IP system with the characteristics
of finiteness: mood distinctions, subject agreement licensing nominative
case, overt tense distinctions.” (Rizzi 1997: 283, 284)
Force and finiteness can be expressed on a single head “in simple cases”
but are forced to split in marked constructions, namely with “the activation
of the Topic-Focus field” (Rizzi 1997: 314). The Force-Fin system thus
includes (at least) the FocP projection, that is the syntactic locus for [+foc]
interpretation (cf., among others, É. Kiss ed. 1995, Rizzi 1997, Frascarelli
2000), and different TopPs in which topic constituents are located accord-
ing to their specific discourse properties 2 (Frascarelli and Hinterhölzl
2007). The C-domain also includes a functional projection connected with
the interrogative force of the selected clause, called IntP, located in a posi-
tion that is higher than FocP (cf. Rizzi 2001). The Force-Fin system can
thus be represented as follows (the asterisk indicates recursion):

(1) [ForceP [AboutTopP [IntP [FocP [FamTopP* [FinP [IP

Languages like Somali and Afar, in which information structure is morpho-

syntactically marked, support a cartographic approach and provide evi-
dence for some refinement. In particular, data show that the activation of
[+foc] plays a crucial role in the interpretation of other discourse catego-
ries, so that an interaction must be posited between functional heads in the
C-domain.

1.2. Somali and Afar: Basic properties 3

Somali is a polysynthetic language (in the sense of Baker 1996). This con-
dition entails that T-roles are only assigned through incorporation into the
verbal head (the so-called “Morphological Visibility Condition”, MVC).
Hence, argument structure is realized by means of clitic pronouns that are
disposed in the Verbal Complex (VC) in a rigid SOV order (Puglielli 1981,
Svolacchia and Puglielli 1999) while full DPs are merged in non-argument
position and connected to the sentence by means of resumptive pronouns,
 Focus in the Force-Fin system: Information structure in Cushitic languages 163

which bind constituents to their thematic roles. Somali is also a focus-

prominent language, so that one constituent must be overtly identified as
the focus of the sentence. In particular, focused DPs occur in preverbal
position followed by the focus marker (FM) baa: 4

(2) Axmed gurigii SHALAY buu (baa+uu) pro 5

Axmed house.ANAPH yesterday FM.SCL3SGM OCL3
nooga (= na+u+ka) qaaday.
OCL1PL-for-from take.PAST.3SGM
‘YESTERDAY Axmed took it from the house for us.’
[lit.: ‘Axmed, the house, YESTERDAY he took it from it for us.’]

The pronominal argument analysis in Somali is supported by a number of

standard diagnostics, e.g. obligatoriness of clitic pronouns in the VC ((2) is
ungrammatical if the SCL uu or the OCL na is omitted), absence of non-
finite clauses and of DP anaphors (cf. Svolacchia and Puglielli 1999). In
particular, the non-argument position of full DPs (independent of focus) is
shown below by means of ‘disjoint reference effects’:

(3) SHALAY bayk jabisay Maryank mindideedk

yesterday FM.SCL3SGF break.PAST.3SGF Maryan knife-her
‘*Shek broke Maryank’s knife YESTERDAY.’

As we can see, a coreferential reading is possible in (3) between the subject

clitic ay and the R-expression Maryan (while this interpretation is excluded
in English). This shows that full DPs in Somali do not sit in A-position and
must therefore be considered as adjuncts.
Afar is an SOV language as well. However, its morphosyntactic proper-
ties are quite different from Somali, so that a comparison of the two lan-
guages is extremely effective for the issues of the present study. Afar is, in
fact, an inflectional pro-drop language in which DPs carry argument role
and pronouns are “strong” elements realized as objects of verbs and postpo-
sitions (cf. (4b)) 6 :

(4) a. Ali mootar-at bilu-k yemeete

Ali car.ABS-by Bilu.ABS-from come.PAST.3SGM
‘Ali came by car from Bilu.’
164 Mara Frascarelli and Annarita Puglielli

b. anu kaa ko-h ruube

I him you-to send.PAST.1SG
‘I sent him to you.’

c. xaa’b-e ‘koo ka’l-e

leave.COMP you stop.PAST.1SG
‘I prevented you from leaving.’

As is shown, the grammar of Afar allows non-finite subordinate clauses

(4c), and pronominal DPs are not obligatory for morphological visibility
(4a). In section 6, we will see that clitic resumption is not present in topic
constructions as well – an additional, major difference with respect to So-
mali.

2. Focus as predicate

In recent works, Frascarelli and Puglielli (henceforth, F&P 2005a, 2005b,

to appear) have argued for a Small Clause (SC) analysis of focus construc-
tions in focus marking languages. Drawing from diachronic and compara-
tive studies (cf. Hetzron 1980, Lamberti 1983) and on the basis of syn-
chronic evidence, F&P show that focus markers (FM) were originally
copular forms, which may have included a 3rd person (clitic) pronoun. Fo-
cus constructions therefore imply the presence of a SC in which the focused
DP is the (specificational) predicate while presupposed information is car-
ried by a relative clause (sitting in subject position). To provide an illustra-
tion, consider the following focus construction in Somali: 7

(5) CALI baa [Op soomaali ah ]REL.

Cali FM (=COP) Somali be.RED
‘CALI is (that is) Somali.’

A sentence like (5) takes as a presupposition the existence of “(some) per-

son who is Somali”, which is expressed through a free relative 8 (defining a
set). Since it is a headless relative, an “external” specification is needed for
the relative operator: this is provided by the DP that moves from predicate
position (in the SC) to Spec,FocP (to check the [+foc] feature). The focus
thus gains scope over the SC and specifies a value for the operator-variable
chain in the relative clause:
 Focus in the Force-Fin system: Information structure in Cushitic languages 165

(6) a. [FocP [Foc’ baa [SC [DP [CP OP [IP tOP soomaali ah ]]] [DP Cali ] ]]]Æ
b. [FocP CALI [Foc’ baa [SC [DP [CP OP [IP tOP soomaali ah ]]] tCali ]]]

Hence, focus is not merged as a part of the vP-phase containing the verb in
the relative clause. It is simply “reinterpreted” as a part of it after identifica-
tion of the operator contained therein. In other words, focus provides a
value for a variable that is not its own copy. 9
This analysis accounts for the connection between focus and relativiza-
tion that was pointed out by many scholars (since Schachter 1973), derives
focus strategies and discourse-semantic properties (Frascarelli 2005), and
explains a number of focus-related phenomena such as the so-called Antia-
greement Effect (cf. Ouhalla 1993, Frascarelli 1999, F&P 2005a, 2005b).
This is illustrated below (from Somali):

(7) Hilibka NIMANKÁAS baa cunayá.

meat.ART men-those.ABS FM eat.PROG.RED
‘THOSE MEN are eating the meat.’

(8) a. * Hilibka NIMANKÀASU baa cunayá.

meat.ART men-those.NOM FM eat.PROG.RED
b. * Hilibka NIMANKÁAS baa cunayaan.
meat.ART men-those.ABS FM eat.PROG.3PL
c. * Hilibka NIMANKÁAS bay cunayá.
meat.ART men-those.ABS FM.3PL eat.PROG.RED

When the subject is focused, it cannot show NOM case but the “default” ABS
case (as used for citation and predicative DPs). Furthermore, it triggers the
presence of the so-called “reduced paradigm” (RED) 10 and, finally, it makes
it impossible for the (focused) subject to be resumed by a pronoun.
In a theory in which the Focus DP is not an argument, but a predicate,
antiagreement effects can be accounted for without stipulations: what we
understand as the “subject” is only reinterpreted as such after the identifi-
cation of the relative operator. Antiagreement can thus be attributed to the
presence of an empty subject (i.e. the variable) in the relative clause. Since
Somali is not a pro-drop language, this obtains a reduced (i.e. kind of parti-
cipial) form of agreement: 11

(7’) [TopP hilibka [FocP NIMANKÁAS [Foc’ baa [SC [DP [CP OP [IP tOP
cunayá ] tnimankáas ]]]]]]
166 Mara Frascarelli and Annarita Puglielli

The morphosyntactic properties of focus constructions in Afar (a non fo-

cus-prominent language) provide important support for the present analysis.
Indeed, the focus shows ABS case independent of its role, and the verb
shows a “nominalizer suffix” whose nature is not clear in the literature (cf.
Bliese 1981). According to the present analysis, these properties obtain a
comprehensive explanation: the verb is embedded in a relative clause, and
the “verb suffix” –m is therefore a COMP (in final position, consistent with
the head-final quality of the language). 12 Consider the following:

(9) [FocP ÀWKA [Foc’ (y) [SC [DP [CP OP [IP tOP huurí-h adda-l
boy.ABS 3SG boat-of inside-in
kullumta habte-m]]] tàwka ]]]
fish leave.PAST.3SGF-COMP.NOM
‘THE BOY left the fish inside the boat.’
[lit.: ‘THE BOY (he is) that left the fish in the boat.’]

The relative DP is the subject of the SC that we posit (hence, marked for
NOM case) while focus is the predicative DP (showing ABS case) that
moves to FocP. The COMP head originates from the feminine noun im
(‘part/thing’) and is connected with a null operator that triggers a default
3SGF agreement. Finally, the optional y (following the focus) is the remnant
of an old copula (cf. Parker and Hayward 1985). Hence, like a FM, it is a
lexicalization of Foc°, attracting the predicative DP in its Spec (for check-
ing requirements).
Let us now investigate the interaction of focus with other discourse
categories in the Force-Fin system, starting with the functional projection
that “looks downstairs” into the propositional content, namely FinP.

3. Focus and the specification of finiteness

3.1. Force in “verb focus” constructions

When the focus field is not activated in Somali, the VC is interpreted as

new information and sentences are characterized by the presence of a spe-
cific marker, waa, traditionally considered as a “verb FM” (Puglielli 1981).
The syntactic properties of waa sentences, however, show that waa cannot
be considered as a Foc head. 13 Consider the subject resumption in the fol-
lowing sentences:
 Focus in the Force-Fin system: Information structure in Cushitic languages 167

(10) Calik MARYAM buuk / *baa arkay.

Cali Maryam FM.3SGM FM see.PAST.3SGM
‘(As for Cali), he saw MARYAM.’

(11) Calik Maryam wuuk / waa arkay.

Cali Maryam WAA.3SGM WAA see.PAST.3SGM
‘(As for Cali), he saw her, Maryam.’

In both (10) and (11), Cali – that is understood as the subject of the follow-
ing sentence – is realized as a left-hand topic. Consistent with the MVC (cf.
Section 1.2), clitic resumption (uu) is obligatory in the focus construction
(10) while it is only optional in the presence of waa (cf. (11)). This asym-
metry is explained if we assume that waa is connected with subject agree-
ment features which are not available to a FM like baa. In other words, baa
and waa do not sit in the same functional projection. Indeed, waa is strictly
connected with overt tense distinctions and NOM case marking (cf. (12)-
(13)) while baa excludes the latter and requires a reduced form of verb
inflection (cf. (8a-b) above). Moreover, no TopP projection is available
between waa and the VC (cf. (14)) while topics are allowed between baa
and the rest of the sentence (cf. (17) below):

(12) Waa sheegaa/ sheegay/ sheegayaa/ sheegayay

say.PRES.3SGM/say.PAST.3SG/say.PRES.PROG.3SGM/say.PAST.PROG.3SGM
‘He says / said /. is saying / was saying.’

(13) Wiilku / *wiilka waa soomaali.

boy.ART.NOM boy.ART.ABS Somali
‘The boy is Somali.’

(14) ( soomaali) Cali (soomaali) waa (*soomaali) ahaa (soomaali).

Somali Cali be.PAST.3SGM
‘Cali was Somali.’

Finally, waa is in complementary distribution with other sentential markers

such as ma in questions (cf. Section 4.1) and ha in imperative clauses (e.g.
ha i labin, ‘don’t’ bother me’). We thus conclude that waa is not a FM, but a
modality marker (cf. also Saeed 1984). It is therefore a lexicalization of
Fin° that activates [+decl] sentential force 14 in the selected IP. As a Fin
head, it is connected with subject agreement and tense specifications
168 Mara Frascarelli and Annarita Puglielli

through AGREE (Chomsky 2001, 2002) with AgrS°. Subject omission is

thus permitted (cf. (11), (12)). The “focus interpretation” on the VC must
therefore be considered as a “by-product” of [+decl] force in the absence of
[+foc] constituents. In other words, when the focus field is not activated,
sentential and illocutionary force coincide: the relevant declarative sentence
has no other “intention” than expressing a statement.

3.2. Force in narrow focus constructions

When the focus field is activated, the illocutionary force of the sentence is
crucially modified. The relevant sentence is merged as a SC construction,
in which new information is only carried by the predicative DP and the
verb is embedded in a relative clause (as presupposed information). This
means that in focus constructions Fin does not select an IP-node, and a
DECL marker like waa is therefore excluded.
Since AGREE is a “local relation between two adjacent heads” (cf.
Chomsky 2001), Fin° cannot be related with AgrS° and subject omission is
therefore not permitted. 15 This explains the asymmetry shown in (10)–(11),
the structures of which are given, respectively, in (15) and (16) below:

(15) [TopP Calik [FocP MARYAM [Foc’ baa+uuk [FinP [Fin’ ‡ [SC [DP[CP OP [AgrSP
tuu tOP arkay] tMaryam ]]]]]]

(16) [TopP Calik [TopP M.z [FocP[FinP [Fin’waa[+decl] [AgrSP (uu)k[AgrS’ ‡z arkay ]]]]]

[+FIN, +AGR]

Sentence like (15), however, show that Fin° (though empty) is not com-
pletely “inert” in narrow focus constructions because it serves as a licensing
head for the subject trace in Spec,AgrSP. This is in line with Rizzi’s (1997)
explanation of “anti-adjacency effects”: in Somali (like in French), Fin° can
be endowed with AgrS features in virtue of its selecting properties. This
allows subject extraction. In this respect, also consider the following sen-
tence:

(17) Shalay CALI baan anigu ku arkay, jaamacadda.

yesterday Cali FM-SCL1SG I.NOM 16 at see.PAST.1SG university.ART
‘Yesterday, I saw CALI, at the university.’
 Focus in the Force-Fin system: Information structure in Cushitic languages 169

In (17), subject extraction occurs across a topic (anigu), which evidently

does not create minimality effects. 17 In this respect, the Somali pattern is
the same as in French, but different from English (examples from Rizzi
1997):

(18) a. Je ne sais pas qui [ ton livre [ t pourrait l’acheter ]]

b.* A man who [ liberty [ t should never grant t to us]]

We suggest that the relevant asymmetry is connected with clitic resumption

and argue for a structural distinction between topic constructions, which
depends on the availability of clitic pronouns in a given language. In Fras-
carelli (2004), evidence is provided that clitic-resumed topics are merged in
the C-domain (where they form a binding chain with the clitic) while non
clitic-resumed topics are moved from an IP-internal position. This distinc-
tion accounts for topic properties from a cross-linguistic perspective: in
languages like French and Somali, topics are inserted in TopP and, as such,
they do not interfere in movement chains:

(16’) [FocP CALIk [Foc’ baa+aanj [TopP anigu [FinP t’j FIN° [SC [DP[CP OP [AgrSP tj
tOP ku arkay] tCali ]]]]]]

In a language like English (in which clitics are not available), on the other
hand, topics are moved and induce minimality effects on subject extraction.
Functional projections in the left periphery thus show different proper-
ties. Some of them are connected with operator movement while others
contain non-quantificational constituents that are merged in the C-domain.
This is a crucial distinction for the internal composition of the Force-Fin
system.

4. Focus and questions

The realization of interrogative force in Cushitic languages shows the exis-

tence of a connection between focus, finiteness, and a specific position in
which the [+int] feature is encoded for interface interpretation.
170 Mara Frascarelli and Annarita Puglielli

4.1. Yes-no questions

A yes-no question in Somali has different realizations according to what is

the “centre” of the interrogative force. Specifically, a question must be
either centered on the VC or on the focused DP, yielding different orders
and marking combinations. Consider the following sentences:

(19) a. Cali hadiyad ma / muu keenáy ?

Cali present QM QM.3SGM bring.PAST.3SGM
‘Did Cali bring a present?’
b.* Cali ma / muu hadiyad keenáy?

(20) a. Cali ma HADIYAD *baa / buu keenáy ?

Cali QM present FM FM.3SGM bring.PAST.3SGM
‘Is it a PRESENT that Cali brought?’
b. ma HADIYAD *baa / buu Cali keenáy?
c.* Cali HADIYAD buu ma keenáy?

As we can see, yes-no questions require the presence of a question marker

(QM), namely ma, which must gain scope over the constituent that is ques-
tioned. Hence, it precedes the VC in (19) and is located in a position higher
than the FocP in (20). Note in particular that ma in (19) shows the same
syntactic properties discussed for waa in [+decl] sentences (cf. Section
3.1): the subject pronoun can be dropped (19a), and no topic is allowed
between ma and the VC (19b). When ma is used in combination with baa,
on the other hand, the properties of a narrow focus construction are found
(compare (20) with (10)).
It is thus possible to conclude that ma is a [+int] marker that is merged
in different functional projections according to the specific illocutionary
force of the sentence. In particular, when the focus field is activated, the
interrogative force must gain scope over it. This forces the activation of a
higher projection in the C-domain that is connected with questioning,
namely IntP (Rizzi 2001). After operator movement of the predicate-focus,
ma is merged in Int°:

(21) [TopP Calik [IntP [Int’ ma [FocP HADIYAD [Foc’ baa+uu [FinP ‡ [SC [DP[CP
OP [AgrSP tuu tOP keenáy] thadiyad ]]]]]]]]]
 Focus in the Force-Fin system: Information structure in Cushitic languages 171

When the focus field is not activated, on the other hand, ma is merged in
Fin° and activates the [+int] feature over the embedded IP, through the
(more economical) relation AGREE with the Int° head:

(22) [TopP Calik[TopP hadiyadz [IntP [Int’ [FocP [FinP [Fin’ ma [AgrSP (uu)k [Agr’ ‡z
[+INT] keenay ]]]]]]]?

This analysis leads to the important generalization that [+int] is activated

on the functional head that takes scope over the informative part of the
sentence. 18
Yes-no questions in Afar present similar properties, thus supporting the
present analysis. Since Afar is not a focus prominent language, questions
can be realized without specific markers through prosodic means (a falling
intonation, cf. (23a)). However, an alternative option is available in which
the QM ma is used (23b):

(23) a. selteè ?
finish.PAST.2SG
b. ma selte ?
QM finish.PAST.2SG
‘Did you finish?’

Like in Somali, Afar ma activates [+int] force on the predicate it selects.

Hence, it obtains a yes-no question interpretation when an IP follows (cf.
(24)), and a wh-question when it precedes a focused DP (as in (25)):

(24) awkí ma biyaakitta ?

boy.NOM QM be sick.3SGM
‘(As for the boy), is he sick?’

(25) ma ÀWKA (y) biyaakittam ?

QM boy.ABS 3SG be sick.3SGF-COMP.NOM
‘Which BOY is sick?’

The latter interpretation, triggered by the combination of [+foc] and [+int],

leads us to explore the correlation between FocP and IntP in more detail.
172 Mara Frascarelli and Annarita Puglielli

4.2. Wh-questions

Given the information-structural parallelism between focus and wh-

constituents (they are both syntactic operators connected with new informa-
tion), FocP has been very often considered to be some sort of “multi-
functional” projection (cf. Horvath 1986, Rizzi 1997, Frascarelli 2000)
targeted by different focus-related categories (new and contrastive foci, wh-
phrases, negative polarity items, etc.). Indeed, wh-constituents and foci do
not co-occur in the same clause, and this can be explained by assuming that
they compete for the same functional position (Rizzi 1997).
This assumption, however, is only partially correct. Frascarelli (2005)
shows that the realization of focus and wh-constituents presents several
asymmetries cross-linguistically and concludes that different projections
must be posited for their interpretation in the C-domain. In particular, the
interpretation of (genuine) wh-questions depends on the interaction be-
tween Foc° and Int°. For this, the internal structure of wh-phrases in
Cushitic languages provides immediate support. Consider the following
(Somali):

(26) Cali muxuu (= ma+wax+baa 19 +uu) cunay ?

Cali QM+thing+FM+3SGM eat.PAST.1SG
‘As for Cali, what did she eat?’

As is clear, the wh-phrase maxaa (‘what’) is obtained through the combina-

tion of two functional heads ([+int] ma and [+foc] baa) with the generic NP
wax. 20 This is also the case in Afar, in which the QM ma combines with the
relative pronoun iyya (cf. (27)) and with the generic noun waxa ‘thing’ (cf.
(28)):

(27) lee MIYYA-y torobem ?

water who.ABS-3SG drink.PAST.3SGF-COMP.NOM
‘Who drank the water?’

(28) awka lee MAXA-t baahtam ?

girl water what.ABS-with bring.PAST.3SGF-COMP.NOM
‘With what brought the girl water?’

We argue that wh-constituents are not inherently specified for a discourse

feature connected with question, even though they are potential candidates
 Focus in the Force-Fin system: Information structure in Cushitic languages 173

for expressing such a discourse property. Wh-constituents take on a genu-

ine (i.e. information-seeking) interrogative force only if they combine with
[+foc]. For this reason, a chain is created between Int° and Foc° through
AGREE. This endows Int° with the [+foc] feature, inducing interrogative
force on the constituent in its scope. 21 The derivation of (26) thus is the
following:

(29) [TopP Calik [IntP [Int’ ma[+int] [FocP WAX [Foc’ baa[+foc]+uuk [FinP ‡ [SC
[DP[CP OP [AgrSP tuu tOP cunáy] twax ]]]]]]]]]
[lit.: ‘As for Cali, was there ANYTHING that he ate?’]

The [+int] feature also triggers operator-movement to Spec,IntP. This dis-

tinguishes a question focused on a DP from a (quantified) wh-question.
Consider the following contrast:

(30) a. ma NIN baa yimid ?

QM man FM come.PAST.RED
‘Is there a MAN who came?’
b. NINMAA (nin+ma+baa) yimid ?
man.QM.FM come.PAST.RED
‘WHICH MAN came?’

(31) a. Haa (NIN baa yimid). (ok for (30a); * for (30b))
‘Yes (A MAN came).’
b. CALI baa yimid. (* for (30a); ok for (30b))
‘CALI came.’

In both (30a–b), the activation of the focus field defines yimid (‘who
came’) as presupposed information. However, (30a) is a yes-no question
while (30b) is not, as is shown by the possible answers in (31a–b). This
shows that the interpretation of wh-questions is obtained through a Spec-
head relation in IntP when Foc° is activated. This analysis is further sup-
ported by wh-phrases like goormaa ‘when’ in Somali:

(32) GOORMUU (goor+ma+baa+uu) yimid ?

time.QM.FM.3SGM come.PAST.3SGM
‘When [lit: ‘in which time’] did he arrive?’
(33) [IntP GOOR [Int’ ma[+int] [FocP [Foc’ baa[+foc]+uu [FinP ‡ [SC [DP[CP OP [AgrSP
tuu tOP yimid ] tgoor ]]]]]]]]]
174 Mara Frascarelli and Annarita Puglielli

In conclusion, the interpretation of genuine wh-questions requires the SC

derivation discussed for focus in Section 2. A wh-question implies the fo-
calization of a generic NP (i.e. PERSON, THING, TIME, etc.) that is in the
scope of [+int]. Since a generic NP cannot specify a value for the operator
in the relative clause, the resulting identification is “vacuous” and the in-
formation-seeking value is thus obtained.
A final consideration concerning the syntactic nature of the interactions
shown between functional positions in the C-domain is in order. The ques-
tion might arise whether a base-generation analysis could not account for
them in a more economic way. Movement constraints and scopal proper-
ties, however, show that the operation Move (in a cartographic approach) is
needed. Evidence in this direction is provided by WCO effects (in Afar 22 )
and scope ambiguities in focus constructions, as is shown in the following
Somali sentence:

(34) DHAMMAN BUUGAAGTA bay saddexda arday gateen

all books.ART FM.3PL three.ART student buy.PAST.3PL
‘The three students bought EVERY BOOK.’

Sentence (34) allows for a double reading: either each of the three students
bought all the books (of a given list) or the books were all bought by the
three students (together). The possibility of the first reading shows that the
quantified phrase dhamman bugaagta can be interpreted in the scope of the
DP saddexda, that is to say, in the position of reconstruction at LF.

5. Focus and illocutionary force

The data discussed so far have shown that FocP is not a “multifunctional”
position: FocP is only dedicated to the syntax and interpretation of informa-
tional focus. The activation of [+foc], on the other hand, is crucial for the
interpretation of focus-related discourse categories.
In this respect, it is important to point out that (genuine) focus informa-
tion is only available in a matrix C-domain. Consider the following data
from Somali (for cross-linguistic evidence, cf. Frascarelli 2005):
 Focus in the Force-Fin system: Information structure in Cushitic languages 175

(35) * hilibka waan sheegay [ in

meat.ART DECL.SCL1SG say.PAST.1SGM that
NIMANKÁAS baa cunayá]
men-those.ABS FM eat.PROG.RED
‘I said that THOSE MEN are eating the meat.’

(36) [ Isaga [ oo isbitaalka (*baa) ku jira]]

he CONJ hospital.ART FM in stay.PAST.RED
buu dhintay
FM.SCL3SGM die.PAST.3SGM
‘He died while he was in the hospital.’

(37) Wiilka [ Maryam (*baa) la hadlayá] waa walaal-kay.

boy-ART Maryam FM with speak.PROG.RED DECL brother-my
‘The boy that is talking to Maryam is my brother.’

As we can see, no [+foc] feature is available in complement (35), adverbial

(36), or relative (37) clauses. Indeed, wh-phrases in embedded domains are
realized as generic NPs heading a relative clause:

(38) waxaan (wax+baa+aan) weyddiyay [goorta /*goormaa

thing.FM.1SG ask.PAST.1SG time-ART / time-QM.FM
Cali imaanayó]
Cali come.PRES.3SGM.DEP
‘I was wondering when Cali is going to come.’
[lit.: ‘The thing that I was asking (is) the time (that) Cali is coming.’]

Sentences like (38) show that genuine (i.e. focused) wh-phrases cannot be
realized in embedded clauses. This shows that “IntP is inherently endowed
with a wh [i.e., quantificational] feature” (Rizzi 2001: 293); not with a
[+foc] feature, however. This combination is obtained through AGREE with
Foc°, which can only occur in a matrix C-domain. In line with recent pro-
posals 23 , we thus conclude that new information is strictly connected with
the root illocutionary force.
176 Mara Frascarelli and Annarita Puglielli

6. Topic constituents and information structure in the C-domain

All along this presentation, we have provided evidence for a number of

interactions occurring between functional projections when the focus field
is activated. The presence of topic constituents, on the other hand, has
proved a marginal role in this respect: [-foc] extraposed DPs do not inter-
fere with operator-movement and the specification of force. Nevertheless, a
closer look at the syntax-discourse properties of topics can help provide a
deeper understanding of information structure in the C-domain.
First of all, it is important to note that, even though dislocation of [-foc]
constituents is a major feature of Cushitic languages, two positions are
totally excluded for topics, that is to say below FinP (cf. (39)) and between
IntP and FocP (cf. (40)):

(39) Cali Maryam ma (*Maryam) arkay ?

Cali Maryam QM see.PAST.3SGM
‘(As for) Cali, did he see her, Maryam?’

(40) Cali ma (*Cali) HADIYAD buu keenáy ?

Cali QM present FM.3SGM bring.PAST.3SGM
‘(As for) Cali, is it a PRESENT that he brought?’

In a cartographic approach of functional projections, we propose that this is

not a language-specific property but the effect of a structural condition for
interface interpretation. In Sections 3 and 4, we have shown the crucial role
of AGREE relations between Fin° and AgrS° and between Int° and Foc°. We
therefore argue that TopP projections are excluded in these positions to
avoid topic interference with UG requirements.
Data from Cushitic languages also show that TopP positions are associ-
ated with different discourse properties (cf. Frascarelli and Hinterhölzl
2007). Consider the following asymmetry in the realization of topics in
Somali yes-no questions:

(41) a. hadiyadda (*hadiyad) ma CALI baa keenáy ?

present.ART QM Cali FM bring.PAST.RED
‘(As for the present) did CALI bring it?’
b. Ma CALI baa hadiyad (*hadiyadda) keenáy?
‘Did CALI bring a present?’
 Focus in the Force-Fin system: Information structure in Cushitic languages 177

In (41), the interrogative force is centered on the focused constituent

(marked by baa) while the rest of the sentence is presupposed information.
It is thus interesting to note that on the topic (hadiyad) an opposite [defi-
niteness] requirement depending on its position is imposed, which is con-
nected with its role in the discourse (as is made clear by translations). Spe-
cifically, the high topic in (41a) must be definite and obtains an “aboutness”
role while the low topic in (41b) is indefinite and is simply part of back-
ground information (as a “familiar” element).
Somali data thus support an analysis in terms of “different types of top-
ics” (cf. note 2): even though [-foc] DPs are merged in extra-sentential
position, it is not the case that just any kind of topic can occupy any posi-
tion in the C-domain. Instead, there is a systematic correlation between
discourse roles and grammatical properties of topics, which is encoded in a
strict hierarchy of functional heads. In particular, the topic position that is
higher than the focus field defines “what the sentence is about” and cannot
be used to express “familiarity” (which is encoded in the structural domain
between Foc and Fin). For further support, consider the “broad focus” ques-
tion in (42) and its possible answers in (43):

(42) Cali muxuu sameeyay ?

Cali QM.thing.3SGM do.PAST.3SGM
‘What did Cali do?’

(43) a. Cali MARYAM buu dilay

‘Cali beat MARYAM.’
b. MARYAM buu dilay, Cali
c. *MARYAM baa Cali dilay

As we can see, when a DP (Cali, in the relevant case) is presented as the

aboutness topic in the question, it cannot be realized as a low topic in the
answer. It can either be maintained in that position or realized as a right-
hand topic (as an “afterthought”).
To conclude this section, let us briefly consider topicalization in Afar.
As is shown in (44)–(45), what is “given” information in the question must
be realized in a fixed order in the answer, which is exactly the order of
arguments in unmarked sentences (namely S>OI>OD):
178 Mara Frascarelli and Annarita Puglielli

(44) Q: ‘Who gave water to Maryam ?’

a. OMAR tee’em Maryam-a lee
Omar give.PAST.3SGF-COMP.NOM Maryam-to water
‘OMAR gave water to Maryam.’
b. Maryam-a lee OMAR tee’em.
c.* OMAR tee’em lee Maryam-a.
d.* lee Maryam-a OMAR tee’em.

(45) Q: ‘What did Omar give to Maryam ?’

a. LEE tee’em Omar Maryam-a.
water give.PAST.3SGF-COMP.NOM Omar Maryam-to
‘Omar gave WATER to Maryam.’
b. Omar Maryam-a LEE tee’em.
c.* LEE tee’em Maryam-a Omar.
b.* Maryam-a Omar LEE tee’em.

To provide an explanation, it is important to point out that in Afar clitic-

resumption is not available. According to our proposal (cf. Section 3.2),
this core-grammar property entails that topics are merged in IP-internal
position. It is thus possible to argue that the rigid ordering imposed on top-
ics is due to movement constraints (i.e. minimality). Hence, given informa-
tion is either in its basic position (so that right-hand topicalization ((44a)–
(45a)) is in fact PF marginalization; cf. Cardinaletti 2002) or, after verb
raising, the remnant VP is moved to the C-domain (possibly to Spec,GP; cf.
Poletto and Pollock 2004) to obtain left-hand topicalization ((44b)–(45b)).
Further investigation is needed to gain deeper understanding of the
structure and interpretation of topic constructions. The connection between
discourse roles and formal properties of topics is the subject of ongoing
research.

7. Conclusions

This paper has provided substantial evidence that discourse roles are con-
nected with specific positions in the C-domain and that AGREE relations
must be posited between functional heads for interpretative requirements.
We have shown the crucial role of the [+foc] feature for the activation of
different kinds of new-related information and its importance for the defini-
tion of the illocutionary force of the sentence. Focus and force are thus
strictly connected and, in this respect, a structural distinction must be pos-
 Focus in the Force-Fin system: Information structure in Cushitic languages 179

ited between matrix and embedded clauses that is centered on the presence
of a “focus field”. Topic constituents, on the other hand, are independent of
the illocutionary force of the sentence and are present in both root and em-
bedded clauses. Their different position in the C-domain is only dependent
on their specific discourse role.

Notes

We thank Istvan Kenesei, Victor Manfredi, Chris Reintges, and the audi-
ence of the Workshop on Focus in African Languages in Amsterdam (De-
cember, 3–4 2004) for helpful comments and discussion.
1. Investigation of the left periphery of the sentence has gained significant
impulse in the Minimalist framework, given the centrality of Interfaces
(the only “conceptually necessary” levels, cf. Chomsky 1995) and the
specification that Internal Merge is connected with scopal features and
discourse requirements (cf. Chomsky 2002, 2004).
2. Contra a “free recursion” analysis of topicalization (cf. Rizzi 1997), Fras-
carelli and Hinterhölzl (2007) provide substantial evidence that different
TopP projections must be posited to realize different “types of topics”. In
particular, the [aboutness] feature is realized in the highest TopP node,
while low topics are connected with [familiarity] and used for continuity
in the discourse. On this point, see also Section 6.
3. When not otherwise indicated, examples are taken from original data. In
this respect, we thank Axmed Cabdullaahi Axmed, Cabdalla Omar Man-
sur (Somali) and Mohammed Ali Mahmoud (Afar) for their help and pa-
tience in testing and discussing our data.
4. The list of the abbreviations used in the glosses is the following:
ABS = absolutive case ANAPH = anaphoric article
ART = definite article DECL = declarative (marker)
F = feminine FM = Focus Marker
M = masculine NOM = nominative case
OCL = object clitic PAST = past tense
PL = plural PROG = present progressive
QM = question marker RED = reduced paradigm
SCL = subject clitic SG = singular
5. The 3rd person object clitic (both genders and numbers) has no phonetic
realization and must be interpreted as an object pro. Evidence for this
analysis is provided by the interpretation of sentences like (i) below. As is
shown in the English translation, this empty category always obtains a
referential reading, so that a pseudo–intransitive use is excluded in this
language (for details, cf. Puglielli 1981, Svolacchia and Puglielli 1999):
180 Mara Frascarelli and Annarita Puglielli

(i) Cali baa pro arkay.

Cali FM see.past.red
‘CALI saw it/him/her/them.’
6. The ABS label is used to indicate non-nominative case, as in use among
scholars of African languages.
7. The traditional arguments for positing a relative clause analysis are dis-
cussed in F&P (2005a). A comparative analysis of the syntactic properties
of relative clauses in Somali and Afar is provided in F&P (2005b).
8. The literature on free relatives is very rich. The present approach is in line
with an analysis of the free relative as a DP that takes a complement CP
(cf. among many others, Groos and van Riemsdijk 1981, Grosu 1994). In
a sentence like (5) the DP-head is the generic (null) noun ‘PERSON’.
9. In Frascarelli (2005), this analysis is discussed in detail and proposed for a
unified theory of focus structure and interpretation, in which identification
and spell-out conditions depend on core grammar rules (the Phase Im-
penetrability Condition) and interface requirements.
10. The reduced paradigm has three forms in Somali: one for 3SGF, one for
1PL and one for all other persons (cf. Puglielli 1981).
11. It is important to underline that reduced agreement is not a specific prop-
erty of focus constructions, but a morphosyntactic effect arising whenever
the subject is the head of an operator-chain. Consider, for instance, the fol-
lowing sentence in which nimankaas is the head of a relative clause (for
the analysis of waa, cf. Section 3.1 below):
(i) nimankaas hilibka cunayá waa saaxiibboday
men-those meat.ART eat.PROG.RED friends-POSS1SG
‘Those men that are eating meat are friends of mine.’
The present analysis thus provides a unified account for anti-agreement
and related effects.
12. Although –m plausibly derives from an original lexical form (roughly
meaning ‘thing’), it cannot be considered the NP-head of the relative
clause. Evidence for its COMP status is provided by the focalization (and
relativization) of PPs. In this case, the preposition cannot be postponed to
–m, forcing a pronominal resumption strategy within the relative clause.
This is illustrated below:
(i) MOHAMMADA Ilyas maruw ka-h ye’em
Mohammed.ABS Ilyas ram him-to
give.PAST.3SG.COMP.NOM
‘Ilyas gave a ram to MOHAMED.’
13. For space limitations, the syntax of waa sentences cannot be addressed in
detail. For the sake of discussion, we simply point out that waa does not
trigger a contrastive reading of the verb, nor does it lead to anti-agreement
effects. Indeed, this property is fully expected in the present analysis. As
is discussed in F&P (2005a), waa constructions do not involve a func-
Focus in the Force-Fin system: Information structure in Cushitic languages 181

tional relation between an operator and a variable in a relative clause.

Quantifying implies, in fact, partitioning of the universe and this operation
only pertains to nominal elements. Anti-agreement effects with focused
verbs are therefore not expected cross-linguistically; possible counterex-
amples would be ascribed to a nominalization strategy for verb focusing
(which is attested in a number of languages).
14. As is discussed in Chierchia and McConnel-Ginet (1990), the sentential
force refers to the conventional pragmatic force associated with a sentence
type. The illocutionary force, on the other hand, is connected with
speaker’s intentions.
15. Indeed, the FM baa does not show any type of number/gender agreement
with the focus phrase in Spec,FocP.
16. Case is transmitted to topic DPs through binding chains (cf. Baker 1996).
17. Note that subjects can be extracted independent of the syntactic role (ar-
gument vs. adjunct) of the constituent in topic position, as is shown by the
following sentence:
(i) anigu CALI baan shalay ku arkay, jaamacadda.
‘I saw CALI, yesterday, at the university.’
18. This is consistent with the Extension Condition (Chomsky 1995, 2001)
and with the fact that, cross-linguistically, it is not possible to focus a con-
stituent and question another part of the sentence. When a question con-
tains a focused DP, this is the predicate in the scope of the [+int] force.
The rest of the sentence is presupposed information:
(i) Is JOHN going out?
[ = ‘Someone is going out (presupposition). Is it JOHN (Focus)?’]
19. The phonological process that obtains the FM –aa in clusters is a debated
question that cannot be addressed in detail. Lamberti (1983) suggests that
baa is derived from an original form *awaa and, indeed, the alternation
[w]|[b] is widely attested in the derivation of Somali verbal nouns (e.g.,
duq ‘old’ Æ duqow ‘to get old’ Æ duqoobid ‘getting old’; for further de-
tails, cf. F&P 2005a). It is therefore possible that, in clusters with the FM,
an underlying [w] is present and simplified (a common morpho-
phonological operation, as is shown by ma+wax in (26).
20. This analysis is very much in the spirit of Cheng’s (1991) proposal of wh-
items as free-choice indefinites (with the interrogative interpretation being
brought about by dedicated question particles). The non-referential inter-
pretation of the noun wax in Somali is supported by independent evidence,
as is shown by examples like the following:
(i) Wax waan arkay.
thing decl.1sg see.past.1sg (‘I saw something.’)
21. Despite the “Freezing Condition”, Rizzi (2004: note 8) does not exclude
the possibility of a “combination of features”. In particular, the author
182 Mara Frascarelli and Annarita Puglielli

suggests the possibility of a complex Q+Foc head (after movement of Q to

Foc).
22. WCO effects are excluded in polysynthetic Somali, since full DPs never
bind a trace in A-position (cf. Section 1.2). An illustration from Afar wh-
questions is given below:
(i) * kayk lee MIYYAY k torobem?
his water who.abs-3sg drink.past.3sgf-comp.nom
23. See Haegeman (2002) on adverbial clauses and Bayer (2001) on emphatic
topicalization in Bavarian.

References

Appleyard, David L.
1989 The relative verb in focus constructions: an Ethiopian areal feature.
Journal of Semitic Studies 34.2: 291–305.
Baker, Mark
1996 The Polysynthesis Parameter. Oxford: Oxford University Press.
Bayer, Josef
2001 Asymmetry in emphatic topicalization. In Studia Grammatica 52:
Adiatur Vox Sapientiae, Caroline Féry and Wolfgang Sternefeld
(eds.), 15–47. Berlin: Akademie Verlag.
Bliese, Loren F.
1981 A Generative Grammar of Afar. Arlington: The Summer Institute of
Linguistics and University of Texas.
Cardinaletti, Anna
2002 Against optional and zero clitics. Right Dislocation vs. Marginaliza-
tion. Studia Linguistica 56.1: 29–57.
Cheng, Lisa Lai-Shen
1991 On the Typology of Wh-questions. MIT, Ph.D. dissertation,
Chierchia, Gennaro, and Sally McConnell-Ginet
1990 Meaning and Grammar. Cambridge Mass., MIT Press.
Chomsky, Noam
1995 The Minimalist Program. Cambridge, Mass.: MIT Press.
2001 Derivation by phase. In Ken Hale: A life in language, Michael
Kenstowicz (ed.), 1–52. Cambridge, Mass.: MIT Press.
2002 An interview on Minimalism. In On Nature and Language, Adriana
Belletti and Luigi Rizzi (eds.), 92–161. Cambridge: Cambridge Uni-
versity Press.
2004 Beyond explanatory adequacy. In Structures and Beyond. The Car-
tography of Syntactic Structures. Vol. 3, Adriana Belletti (ed.), 104–
131. Oxford: Oxford University Press.
 Focus in the Force-Fin system: Information structure in Cushitic languages 183

É. Kiss, Katalin (ed.)

1995 Discourse Configurational Languages. Oxford: Oxford University
Press.
Frascarelli, Mara
1999 Subject, nominative case, agreement and Focus. In Boundaries of
Morphology and Syntax, Lunella Mereu (ed.), 195–215. Amsterdam:
John Benjamins.
2000 The Syntax-Phonology Interface in Focus and Topic Constructions
in Italian. Dordrecht: Kluwer Academic Publishers.
2004 Dislocation, clitic resumption and minimality: A comparative analy-
sis of left and right Topic constructions in Italian. In Romance lan-
guages and linguistic theory 2002, Reineke Bok-Bennema, Bart
Hollebrandse, Brigitte Kampers-Manhe, and Petra Sleeman (eds.),
99–118. Amsterdam: John Benjamins.
2005 Focus as predicate and Spell-out conditions. Deriving cross-
linguistic strategies from a unified account. Ms. Univ. Roma Tre.
Frascarelli, Mara and Annarita Puglielli
2005a The Focus System in Cushitic Languages. In Proceedings of the 10th
Hamito-Semitic Linguistics Meeting (Quaderni di Semitistica, 25),
Pelio Fronzaroli and Paolo Marrassini (eds.), 333–358. Firenze: Uni-
versità degli Studi di Firenze.
2005b A comparative analysis of restrictive and appositive relative clauses
in Cushitic languages. In Contribution to the IGG XXX, Laura Brugè,
Giuliana Giusti, Nicola Munaro, Walter Schweikert, and Giuseppina
Turano (eds.), 279–303. Venezia: Cafoscarina.
to appear Focus Markers and Universal Grammar. In Cushitic and Omotic
languages: Proceedings of the fourth International Conference on
Cushitic and Omotic Languages (Leiden, 10–12 april 2003), Azeb
Amha, Maarten Mous, and Graziano Savà (eds.), Cologne: Rüdiger
Köppe Verlag.
Frascarelli, Mara, and Roland Hinterhölzl
2007 Types of Topics in German and Italian. In On Information Structure,
Meaning and Form, Susanne Winkler and Kerstin Schwabe (eds.),
87–116. Amsterdam: John Benjamins.
Groos, Anneke, and Henk van Riemsdijk
1981 Matching effects in free relatives: a parameter of core grammar. In
Theory of Markedness in Generative Grammar, Adriana Belletti,
Luciana Brandi, and Luigi Rizzi (eds.), 171–216. Pisa: Scuola Nor-
male Superiore.
Grosu, Alexander
1994 Three Studies in Locality and Case. London: Routledge.
184 Mara Frascarelli and Annarita Puglielli

Haegeman, Liliane
2002 Anchoring to speaker, adverbial clauses and the structure of CP.
Georgetown University Working Papers in Theoretical Linguistics 2,
117–180.
Hetzron, Robert
1980 The Limits of Cushitic. Sprache und Geschichte in Africa 2, 7–126.
Horvath, Julia
1986 Focus in the Theory of Grammar and the Syntax of Hungarian.
Dordrecht: Foris.
Lamberti, Marcello
1983 The origin of the Focus particles in Somali. In Sprache, Geschichte
und Kultur in Afrika, Vossen von Rainer and Ulrike Claudi (eds.),
59–112. Hamburg: Helmut Buske Verlag.
Ouhalla, Jamal
1993 Subject-extraction, negation and the anti-agreement effect. Natural
Language & Linguistic Theory 11: 477–518.
Parker, E. M., and R .J. Hayward
1985 An Afar-English-French Dictionary (with Grammatical Notes in
English). London: SOAS.
Puglielli, Annarita (ed.)
1981 Studi Somali 2. Sintassi della Lingua Somala. Roma: Ministero AA
EE.
Rizzi, Luigi
1997 The Fine Structure of the Left Periphery. In Elements of Grammar.
Handbook in Generative Syntax, Liliane Haegeman (ed.), 281–337.
Dortrecht: Kluwer Academic Publishers.
2001 On the position Int(errogative) in the Left Periphery of the Clause. In
Current Studies in Italian Syntax. Essays offered to Lorenzo Renzi,
Guglielmo Cinque and Giampaolo Salvi (eds.), 287–296, Amster-
dam, New York: Elsevier.
2004 (ed.) The Structure of CP and IP: The Cartography of Syntactic
Structures. Vol. 2. Oxford: Oxford University Press.
Saeed, John Ibrahim
1984 The Syntax of Focus and Topic in Somali. Hamburg: Helmut Buske.
Schachter, Paul
1973 Focus and relativization. Language 49:19–46.
Svolacchia, Marco, and Annarita Puglielli
1999 Somali as a polysynthetic language. In Boundaries of Morphology
and Syntax, Lunella Mereu (ed.), 97–120. Amsterdam: John Benja-
mins.
Coptic relative tenses: The profile of a morpho-
syntactic flagging device

Chris H. Reintges

Abstract

A number of languages morphologically distinguish relative clauses and questions

as well as focus constructions from simple declarative clauses by means of special
inflectional morphology. This morphological pattern thus provides prima facie
evidence for these constructions as a natural class. The present paper comprises a
case-study of the morpho-syntax of so-called relative tenses in Coptic Egyptian
(Afro-Asiatic, 3rd-13th c. CE). Different from well-studied cases of special inflec-
tion, Coptic relative tenses are selected when the questioned or focused constituent
appears in-situ in a clause-internal position, yet displays the scope and interpreta-
tional properties of a displaced constituent. To accommodate these facts, an analy-
sis in terms of covert wh-movement will be proposed.

1. Introduction

Since Chomsky’s (1977) ground-breaking study on wh-movement, consid-

erable attention has been paid to the syntactic similarities between such
apparently diverse construction types as relative clauses, constituent ques-
tions, and contrastive focus structures. What these sentence types have in
common, is that they can be most insightfully analyzed as involving an
open position or “variable” that is assigned an interpretation by a scope-
taking operator. In the generative-transformational framework, the opera-
tor-variable relation has been analyzed as the outcome of a movement op-
eration that links two or more positions in the syntactic structure. The dis-
placed operator (a question word or focus phrase) is relocated to a position
in which its scope and interpretational properties are determined while its
case and thematic properties are interpreted in the original position (or in-
termediate ones along the movement path).
The evidence for the unity of operator-variable constructions is largely
indirect: since operator-variable dependencies typically involve syntactic
186 Chris H. Reintges

reordering, the point of origin of the movement process is generally pho-

netically empty. English constituent questions represent a clear example
where the operator is visible in the form of the fronted wh-phrase while the
associated position in which the wh-phrase is interpreted is not, e.g. Whom
did you meet __ yesterday evening? (where __ indicates the position from
which movement has occurred). Indeed, in the standard analysis of that-
relative clauses (Browning 1987), neither the operator nor the variable is
phonetically realized, e.g. the man [Op that you met __ yesterday evening],
where Op stands for the null counterpart of the relative pronoun whom.
There is, however, direct morphological evidence for the naturalness of
operator-variable constructions as a class. A number of genetically related
and unrelated languages distinguish relative clauses, questions, and a range
of declarative focus constructions from pragmatically neutral clauses by
means of a special type of inflectional morphology. A particularly well
studied example of this morphological pattern comes from relative aspects
in Hausa (Chadic, Nigeria). In this language, the perfective and imperfec-
tive forms divide into two paradigms: the “general” and the “relative”. The
latter is selected in relative clauses, wh-fronting questions, and focus-
fronting constructions. See Tuller (1986), Green (1997, in press), Newman
(2000), Jaggar (2001), Hartmann & Zimmermann (this volume) for a de-
tailed description and analysis. 1 (Relative aspect forms appear in boldface;
small caps indicate focus.)

(1) a. Relative clause (Hausa)

kƗ ga bƗ̖Ýî-n [dà sukà isǀ yànzu]?
2SG.M.PERF see guest-DD.PL COMPREL REL.PERF.3PL arrive now
‘Did you see the guests who just arrived?’ (Jaggar 2001, 527)
b. Wh-question
wƗ̖ dà wƗ̖ sukà zǀ ?
who.pl REL.PERF.3PL come
‘Who came?’ (Newman 2000, 488)
c. Declarative focus sentence
Amir˾kƗwƗ nƝ̖ sukà fƗrà zuwƗ̖.
Americans COP.PL REL.PERF.3PL be.first go.VN
‘(Who got to the moon first?) It was the AMERICANS who got
(there) first.’ (Jaggar 2001, 494)

The Coptic conjugation system shows a similar subdivision of tense-aspect-

mood (TAM) markers which fall into a “general” and a “relative” class. 2
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 187

Thus, consider the following sentence pair in which both the declarative
clause and the corresponding wh-questions employ the Perfect marker a,
which appears in front of the subject. (The basic word order pattern is sub-
ject-verb-object SVO.) But while pragmatically neutral declaratives like
(2a) are well-formed with the tense-aspect word alone, wh-questions like
(2b) require the presence of the relative marker ‹ntí besides the TAM
marker in order to be grammatical.

(2) a. Neutral declarative clause (general TAMs) (Coptic)

awn a ruhe šnpe ero-f J‹O p-ma ‹n-Taȕenİse
and PERF night happen to-3SG.M in DEF.SG.M-place of-Tabenêse
‘And night-fall surprised him in the area of Tabenêse.’
(V. Pachôm. 136, 24–25)
b. Wh-question (relative TAMs)
‹nt-a u šnpe ‹mmo-k pa-þoeis p-‹rro
REL-PERF what happen to-2SG.M DEF.SG.M.1PL-lord DEF.SG.M-king
‘What happened to you, our lord and king?’ (Eudoxia 36, 24)

On the face of it, Coptic relative tenses look like a clause-typing device that
distinguishes questions from declarative sentences. A clause-typing analy-
sis would, however, be at odds with the broad distribution of relative TAMs
across different sentence types. Just like in Hausa, Coptic relative TAMs
appear in relative clauses, wh-questions, and focusing sentences. 3

(3) a. Relative clause

e-p-ma [nt-a-k-kynt-f nhİt-f ]
to- -place REL-PERF-2SG.M-find-3SG.M inside-3SG.M
‘the place where you found it’ (Ac. A&P 204,145-146)
b. Wh-question
‹nt-a u ȕnk e-pe.k-hİt ?
REL-PERF what come to-DEF.SG.M-2SG.M-heart
‘What has come into your heart?’ (AP Chaîne, no.139, 31:7)
c. Declarative focus sentence
‹mmnn alla ‹nt-a-u-s‹ȕtot-f
no but REL-PERF-3PL-prepare-3SG.M
‹m-p-diaȕolos m‹n ne.f-aggelos
for-DEF.SG.M-devil with DEF.PL-3SG.M 3sm-angels
‘(Is the Purgatory prepared for us?) Not at all! It is rather
prepared FOR THE DEVIL AND HIS ANGELS.’ (B. Hom. 14, 19-21)
188 Chris H. Reintges

The presence of operator-variable dependencies in relative clauses, ques-

tions, and contrastive focus constructions provides a necessary, but not
sufficient condition for the introduction of relative tenses into the structure:
the relevant operator must also be in the appropriate configuration. As
shown by the contrast between (4a) and (4b) below, relative TAMs occur in
wh-in-situ questions, so called because the question word is located in the
same position as its non-interrogative counterpart in declarative sentences.
By contrast, no such special inflection is needed when the question word
moves to the left of the clausal subject.

(4) a. Wh-in-situ question (relative TAMs)

awo nt-a-u-ei eȕol tnn ?
and REL-PERF-3PL-come PCL where
‘From where did they come?’ (Apoc. 7, 13; ed. Budge)
b. Wh-fronting question (general TAMs)
eȕol tnn a-tet‹n-ei e-pei-ma ?
PCL where PERF-2PL-come to-DEM.SG.M-place
‘From where did you come here?’ (B. Martyr. 220, 8)

These distributional patterns highlight the two important aspects of Coptic

relative tenses: on the one hand, this special inflection unifies a class of
constructions while, on the other hand, it is subject to parametric variation
in terms of the syntactic conditions that govern its presence.
The paper is organized as follows. Section 2 sets the stage by discussing
the typological profile of special inflection. Section 3 provides some back-
ground information about Coptic clause structure with particular attention
paid to the topic/focus field. The focus of section 4 is on the categorical
properties and morphological alternations of relative TAM markers, which
will be analyzed as alternating relative complementizers. Section 5 exam-
ines the distribution of relative TAMs across different sentence types. Elabo-
rating on Reintges, LeSourd & Chung’s (2006) comparative study, section
6 presents a syntactic analysis of Coptic wh-in-situ questions as “hidden”
movement configurations. Section 7 examines the relation between mor-
pho-syntactic flagging and the overt versus covert movement distinction.
Section 8 summarizes the main results of the paper.
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 189

2. Parameters of cross-linguistic variation

The use of special inflectional morphology is attested cross-linguistically as

one of the options for the structural encoding of features associated with the
standard inventory of operator-variable constructions (e.g. Hyman &
Watters 1984, Haïk 1990, Watanabe 1996, Chung 1998). Reintges, Le-
Sourd & Chung (2006, 166ff.) identify a set of parameters of cross-
linguistic variation, which include (A) the location of special inflection, (B)
the structural information encoded, (C) the recursiveness property in long-
distance movement construction, and (D) the inventory of operator-variable
constructions thus flagged.

A. LOCATION. The morphology of special inflection is associated with

a restricted set of syntactic heads. According to Zaenen (1983), this
type of morphology surfaces on verbs, inflectional heads (TAM
markers), or left-peripheral heads like complementizers.

In Hausa, this morphological pattern is visible in the so-called person-

aspect complex (PAC), which occurs in preverbal person and is morphologi-
cally independent of it. The PAC inflection consists of a weak pronoun and a
TAM marker, which are sometimes fused together (Newman 2000, 564ff.).
Depending on how the PAC is decomposed in such cases, one might infer
that inflection is either fused with a morpheme that encodes person, num-
ber, and gender features (i.e. phi-features) or else encodes phi-features it-
self. On either analysis, Hausa represents the typologically marked case of
a language in which this morphological pattern is sensitive to phi-features.
Furthermore, moved focus/wh-phrases may be modified by a pronomi-
nal copula that recurs in predicational sentences. Green (1997, in press)
analyzes the copula as a focus marker that is merged in the head position of
a designated focus phrase. As illustrated in (5), the copula cƝ̖ (SG.F) agrees
in number and gender with the question word wƗ̖cƝ ‘who’ (SG.F).

(5) Wh-question with relative aspect and focus copula (Hausa)

wƗ̖cƝ cƝ̖ ta mutù ?
who.SG.F COP.SG.F 3SG.F.REL-PERF die
‘Who died?’ (Jaggar 2001, 514)

The semantic contribution of the optional focus copula is subtle, but it

seems to reinforce an exhaustive listing interpretation. In this respect, it
190 Chris H. Reintges

differs from fully grammaticalized focus markers (Hartmann &

Zimmermann, this volume). Green & Reintges (2005) view the focus cop-
ula as a different manifestation of special inflectional morphology. How-
ever, in having semantic import, it differs from relative aspects that merely
signal an application of wh-movement.

B. STRUCTURAL INFORMATION. In some languages, special inflection

only indicates operator-variable dependencies. In other languages,
it also indicates the categorial features as well as the case and theta-
role of the moved constituent (cf. the term “wh-agreement”).

Chamorro, an Austronesian VSO language spoken on the Mariana Islands,

has a standard inventory of wh-constructions all of which exhibit the hall-
marks of wh-movement (displacement of the wh-phrase to the left periph-
ery, obligatory gap at the extraction site, sensitivity to islands). According
to Chung (1994, 1998), wh-movement feeds two distinct morphological
operations. One operation triggers the introduction of “wh-agreement”
marking on finite verbs and adjectives. The other operation involves alter-
nations in the form of the lexical complementizer, which may lack phonetic
content (indicated as [C Ø]).
Wh-agreement registers the case of the variable left behind by wh-
movement while Operator-C agreement marks both the categorial status
and the thematic role of the moved constituent. Thus, compare the declara-
tive clause in (6a), where the verb assumes neutral inflection, with the cor-
responding wh-questions in (6b–d) below. In (6b), the nominative case of
the moved wh-subject is flagged by the wh-agreement morpheme -um- on
the finite verb. In (6c), the objective case of the wh-object is registered by
the infix -in- plus possessor agreement. In (6d), the complementizer ni
marks both the nominal category and the locative theta-role of the fronted
wh-adverb manu ‘where’.

(6) a. Neutral declarative VSO clause (Chamorro)

Ha-fa!gasi si Juan i kareta.
AGR-wash Juan the car
‘Juan washed the car.’
b. Wh-agreement; wh-subject with nominative case
Hayi [C Ø] fuma!gasi i kareta ?
who COMPQ WH[NOM].wash the car
‘Who washed the car?’
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 191

c. Wh-agreement; wh-object with oblique case

Hafa [C Ø] fina!gasése-nña si Henry pära hagu?
what COMPQ WH[OBJ].wash.PROG-AGR Henry for you
‘What is Henry washing for you?’
(Chung 1998, 236, (52)–(53a–b))
d. Operator-C agreement, wh-adverb
Manu ni mañ-ásaga siha ?
where COMPQ AGR.live.PROG they
‘Where are you living?’ (Chung 1998, 58, (80a))

The morphological effects of wh-agreement proper are induced by the trace

left behind by the moved wh-phrase, which thus behaves just like any overt
nominal in that it must be case-marked by the inflected verb.

C. RECURSIVENESS. Languages vary according to the domain in which

the designated syntactic heads are affected. In cases of long-
distance movement across a clause-boundary, special inflection
surfaces either on every designated head along the extraction path
or on the highest one (Haïk 1990, 354–62).

A salient feature of Modern Irish wh-movement constructions is the pres-

ence of the alternating complementizer particles aL (the ‘direct relative
particle’) and aN (the ‘indirect relative particle’). The COMP-particle aL is
selected in contexts in which wh-dependencies terminate in a gap while the
particle aN is chosen when such dependencies terminate in a resumptive
pronoun (McCloskey 1990, 2001). In cases of long-distance wh-movement
across a clause boundary, every complementizer that intervenes between
the moved wh-phrase and the variable in the base position assumes its char-
acteristic form aL. 4

(7) Multiple occurrence of alternating complementizer (Irish)

Céi a dúradh léithi [a cheannódh_i é]?
who COMP.aL was-said with-her COMP.aL would-buy it
‘Who was she told would buy it?’(McCloskey 2001, 94 fn. 23)

A different situation obtains in Hausa long-distance wh-movement struc-

tures where only the highest TAM inflection assumes the relative aspect
form. 5
192 Chris H. Reintges

(8) Single occurrence of relative aspect (Hausa)

wàÎànnè yârƗi HàlƯmà ta tàmbayƗ̖
which.PL children Halima 3SG.F.REL-PERF ask
[kǀ___ sun sƗ̖ci kuÎîn] ?
COMPQ 3PL.PERF steal money
‘Which children did Halima ask whether they stole the money?’
(Newman 2000, 502)

D. DISTRIBUTION OVER SENTENCE TYPES. Special inflection always

flags canonical wh-constructions, yet in some languages it also ap-
plies to yes-no questions, conditionals, temporal adverb clauses,
predicative adjuncts, as well as narrative use (Hyman & Watters
1984, 256ff.; Haïk 1990, 350f).

In Hausa, the relative perfective may appear in conditional clauses as well

as in series of past events, which constitute the foregrounded portions of the
narrative (Tuller 1986, 112ff.; Jaggar 2001, 162ff, in press). (The discourse
fragment in (9a) is from a personal “brush-with-death” narrative.)

(9) a. Conditional clause (Hausa)

in kin/kikà kintsƗ̖, zƗ mù tàfi.
if/when 2SG.F.PERF/2SG.F.REL-PERF be.ready FUT 1PL go
‘If/when you are ready, we’ll go.’ (literally: ‘If/when you have
got ready…’) (Jaggar 2001, 609)

b Narrative relative perfective

(…) mukà bnjÎè Ýǀfàr˾˾ mǀtàr˾, mukà yi saurƯ,
1PL.REL-PERF open door.of car.DD 1PL.REL-PERF do speed,
sai wani mùtûm ya bnjÎƝ̖ mîn (…)
then DEM men 3SG.M. REL-PERF open IO.1SG
mukà yi saurƯ, mukà jƝ, mukà bnjÎƝ̖ (…)
1PL.REL-PERF do speed 1PL.REL-PERF go 1PL.REL-PERF open
‘(... ) we opened the car door and moved quickly, then some man
opened (it) for me (…), and we moved quickly and went and
opened (it) (…)’ (Jaggar 2001, 163)

One might wonder whether the narrative use of relative aspects registers the
presence of an operator-variable dependency. Since there is no overt focus
constituent, the spell-out of special inflection in foregrounded chains of
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 193

events may be related to the presence of a null focus operator (Tuller 1986,
117).

3. The cartography of the Coptic left periphery

In this section, I will lay out the background assumptions about Coptic
syntax that the discussion of relative tense formation in the reminder of this
study is based on. Coptic Egyptian may be classified as a discourse-
configurational language where topic and focus prominence involve a de-
parture from the canonical SVO surface order. See Reintges (2004, chap.
10) for a more detailed discussion of the correlation between word order
and discourse structure.
The markers of tense-aspect-mood are phonological clitics, which dis-
play characteristic properties of auxiliary verbs, such as impoverished
agreement and the compatibility with several clausal positions. Most TAMs
occur in clause-initial position, preceding both the subject and the main
verb. Root modals, on the other hand, appear in a lower syntactic position,
following the subject and preceding the main verb. For the purposes of this
paper, I will not further explore the complex interaction between the two
inflectional positions but assume, following Rizzi (1997), that clause-initial
TAMs are directly merged into the ‘Finiteness Phrase’ (FINP) while root
modals are either merged or move together with the main verb to the head
of a clause-internal ‘Mood Phrase’ (MODP) (see Reintges 2001 for a more
detailed discussion).
Just as in Hungarian, Hausa, and the Kwa languages, the discourse-
configurational syntax of Coptic Egyptian involves an articulated topic-
focus field (see, among various others, É. Kiss 1998, Rizzi 1997, Green
1997, in press, Green & Reintges 2004a, 2005, Aboh 2004). To begin with,
fronted focus/wh-phrases appear following the subordinating complemen-
tizer þe ‘that’, as in (10a), or the dedicated interrogative particle eye, as in
(10b), either of which is merged into the highest functional head of the
clause, viz. the C0/Force0-node. From this it follows that the moved fo-
cus/wh-phrase is located in the specifier position of a functional projection
below C0 and above the FINP.
194 Chris H. Reintges

(10) a. ti-þ‹nu ‹mmn-t‹n [þe h‹n u ‹n-šaþe

(PRES)1SG-ask PREP-2PL COMP with what of-word
a-tent‹n-mute ero-i (…)]
PERF-2PL-say about-1SG
‘I ask you with which reason do you say about me (...).’
(Acts 10, 29)

b. eye etȕe u tet‹n-ynr‹m ‹nsn-n hns ‹nt-a-‹r

Q for what (PRES)2PL-gaze at-1PL as.if REL-PERF-1SG-make
pai h‹n te.n-kyom İ ten-m‹nt-euseȕİs ?
DEM.SG.M through DEF.SG.F.1PL-power or DEF.SG.M.1PL-piety
‘For what reason do you gaze at us as if we have done this (thing)
out of our (own) power or piety?’ (Acts 3, 12)

Topicalized constituents may either precede or follow fronted focus/wh-

phrases.

(11) a. anon de etȕe ne.n-noȕe mar-‹n-opt-‹n mawaa-n

we PCL for DEF.PL.1PL-sin OPT-1PL-lock.up-1PL SELF-1PL
‘(As for) us, BECAUSE OF OUR SINS, let’s lock ourselves up!’
(AP Chaîne no.41, 8:28–29)

b. awo n-aš ‹n-he anon t‹n-snt‹m p-wa

and in-what of-manner we (PRES)1PL-hear DEF.SG.M-one
p-wa hrai h‹n te.f-ape
DEF.SG.M-one PCL in DEF.SG.F.3SG.language
‘And how do we hear each one (of us) in his native tongue?’
(Acts 2, 8)

The availability of two topic positions can be accommodated straightfor-

wardly under Rizzi’s (1997) left-periphery analysis, according to which a
non-recursive ‘Focus Phrase’ (FOCP) is embedded into a topic field. Since
preposed time and location adverbs as well as free-standing pronouns typi-
cally appear to the right of fronted focus/wh-phrases, Green & Reintges
(2004a, 2005) propose that such ‘lower’ topics are base-generated in the
specifier position of the FINP. The relative ordering of complementizers,
topic constituents, and fronted focus/wh-phrases motivates the following
hierarchical structure of the Coptic clause.
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 195

(12) The fine structure of the Coptic left periphery

CP
3
C0 TOPP
3
TOPIC TOPP
3
Top0 FOCP
3
FOCUS/WH-XP FOCP
3
FOC0 FINP
3
TOPIC FINP
3
FIN0 MODP
TAM 3
SU MODP
ru
MOD0 VP
TAM

If we assume that moved wh-phrases or focus constituents compete for the

same position Spec, FOCP, a number of gaps in the Coptic data receive a
principled explanation. Thus, neither multiple focus fronting nor a combi-
nation of wh-movement and focus-fronting is attested.

4. The morpho-syntax of relative TAMs

The focus of this section is the morphological derivation and the syntactic
positioning of Coptic TAMs. This special morphological pattern is derived
by adding a relative complementizer to the TAM marker, which is inde-
pendent of the verb. Despite their categorical status, relative markers ap-
pear in Foc0 rather than in C0. 6
196 Chris H. Reintges

4.1. A paradigmatic split in the Coptic tense-aspect-mood system

Coptic Egyptian has a complex inflectional system, which comprises more

than twenty different verbal tenses. At the foundation of this richness of
morpho-semantic distinctions is the division of the four absolute tenses (the
Present, the Habitual, the Future, and the Perfect) into a general and a rela-
tive class, cf. table 1. (The verb nau ‘to see’ has been chosen to illustrate a
typical paradigm: a-f-nau ‘he has heard’.)

Table 1. The inflectional paradigms of Coptic affirmative tenses

PRESENT PERFECT
general relative general relative
1SG ti-nau e-i-nau 1SG a-i-nau ‹nt-a-i-nau
2SG.M k‹-nau e-k-nau 2SG.M a-k-nau ‹nt-a-k-nau
2SG.F te-nau e-te-nau 2SG.F are-nau ‹nt-are-nau
3SG.M f‹-nau e-f-nau 3SG.M a-f-nau ‹nt-a-f-nau
3SG.F s‹-nau e-s-nau 3SG.F a-s-nau ‹nt-a-s-nau
1PL t‹n-nau e-t‹n-nau 1PL a-n-nau ‹nt-a-n-nau
2PL tet‹n-nau e-tet‹n-nau 2PL a-t‹n-nau ‹nt-a-t‹n-nau
3PL se-nau e-u-nau 3PL a-u-nau ‹nt-a-u-nau
FUTURE HABITUAL
general relative general relative
1SG ti-na-nau e-i-na-nau 1SG ša-i-nau e-ša-i-nau
2SG.M k‹-na-nau e-k-na-nau 2SG.M ša-k-nau e-ša-k-nau
2SG.F te-na-nau e-te-na-nau 2SG.F šare-nau e-šare-nau
3SG.M f‹-na-nau e-f-na-nau 3SG.M ša-f-nau e-ša-f-nau
3SG.F s‹-na-nau e-s-na-nau 3SG.F ša-s-nau e-ša-s-nau
1PL t‹n-na-nau e-t‹n-na-nau 1PL ša-n-nau e-ša-n-nau
2PL tet‹n-na-nau e-tet‹n-na-nau 2PL ša-t‹n-nau e-ša-t‹n-nau
3PL se-na-nau e-u-na-nau 3PL ša-u-nau e-ša-u-nau

There is a clear morphological relationship between the two sets of ab-

solute tenses: relative tenses are derived from general tenses by adding a
relative marker e– or ‹nt– in front of the TAM expression (a null morpheme
in the present tense). The system of relative TAM formation is organized
around a marked value represented by the Relative Perfect marker ‹nt– and
the ‘elsewhere’ form e–. Besides the affirmative relative forms listed in
table 1, Coptic has two modal forms, the Conditional conjugation e-f-šan-
nau ‘if/when he sees’ and the Third Future e-f-e-nau ‘he shall see’, which
both show the relative marker e– but lack the corresponding non-relative
form.
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 197

(13) awn on ne-snİu [e-u-šan-amelei]

and PCL DEF.PL-brothers REL-3PL-COND-neglect
e-k-e-nehse ‹mmn-u h‹m p‹-nomos
REL(-PRES)2SG.M-PREP-wake PREP-3PL in DEF.SG.M-law
e-p‹-nute
to-DEF.SG.M-god
‘And also (concerning) the brothers, if they become negligent,
you shall raise them up in the law of God.’
(V. Pachôm. 94, 14–16)

Since Coptic makes productive use of relative TAMs in the modal system,
these forms can even be described as the predominant system. The alterna-
tions in the shape of the relative TAM marker provide prima facie evidence
for the correlation between special inflection and tense/finiteness.

4.2. The syntactic position of relative TAMs

Coptic relative clauses are externally headed and contain no relative pro-
noun or other subordinator besides the relative TAM marker. In expressing
the core functions of relativization (attribution, subordination), such rela-
tive markers can be classified as [+finite] relative complementizers (see de
Vries 2002, chap. 5). The relative marker itself does not encode any nomi-
nal-functional features of the relative head; all these features are present on
a resumptive pronoun in the embedded clause.

(14) nei-esowi [e-ša-u-ȕi n-toot-ui

DEM.PL-sheep REL-HAB-3PL-take from-hand-3PL
m-pe.ui-šos]
PREP-DEF.SG.M.3PL-shepherd
‘(like) sheep from which their shepherd is taken away’
(V. Pachôm 92, 30–93, 1)

It is possible to combine the relative marker e– with the relative comple-

mentizer et–, as seen in (15). When preceding a nominal subject, the
marker e– assumes a longer form ere– (Reintges 2004, 414ff. §11.1.2).

(15) p-İii [et-ere pei-šİre šİm mowt nhİt-fi]

DEF.SG.M-house COMPREL-REL DEM.SG.M-boy little die inside-3SG.M
‘the house where this little boy died’ (Ac. A&P 206, 163–164)
198 Chris H. Reintges

The multiple occurrence of such relative elements shows that Coptic rela-
tive clauses have an articulated left-periphery. While the initial comple-
mentizer et– is merged into the C0-node, the default relative marker e– /
ere– occupies a lower functional position which can be identified with the
head of the focus projection. Further empirical support for a lower left-
peripheral position of relative TAM markers comes from clitic left-
dislocation structures like (17) where the topicalized pronoun anok ‘I’ pre-
cedes the relative marker e– (cf. Cinque 1990).

(16) p-mai [ anok e-ti-na-ȕnk ero-fi ]

DEF.SG.M-place I REL-1SG-AUX-go to-3SG.M
‘the place where I am going to’ (John 8, 21)

Coptic relative tenses display a considerable number of context-dependent

alternations in the form of the relative marker. Reintges (2003, 400f.) views
these complementizer alternations as a morphological reflex of the move-
ment and the incorporation of the TAM auxiliary into the relative marker in
Foc0. This is shown in tree diagram (17). (OP stands for the abstract relative
operator.)

(17) The left periphery of Coptic relative clauses

CP
2
OP CP
2
C0 FOCP
et– 2
FOCP
2
FOC0 FINP
ere– 2
FIN0 MODP
TAM 2
SU MODP
2
T0-FOC0-movement MOD0 VP
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 199

T(ense)0-to-FOC(us)0 movement seems to be a corollary of T0-to-C0 move-

ment (Rizzi 1996, Pesetsky & Torrego 2001). Since the morphological
effects of incorporation surface on the FOC0 head, it acts as the attracting
category or probe, presumably because it has unvalued tense/finiteness
features that need to be checked. 7

5. Distribution across sentence types

This section surveys the non-relative environments in Coptic where relative

TAMs are allowed and/or grammatically required. The presence of relative
TAMs is obligatory in wh-in-situ questions as well as a range of declarative
focus constructions. In yes-no questions, on the other hand, the presence of
relative TAMs seems to be optional.

5.1. Wh-in-situ questions

In Cheng’s (1991) typological framework, Coptic can be classified as an

optional fronting language, in which wh-fronting and wh-clefting are avail-
able as marked alternatives to the canonical wh-in-situ pattern. Wh-in-situ
has a broad syntactic distribution, appearing in main and embedded clauses.
Moreover, neither wh-arguments nor wh-adjuncts show any resistance to
wh-in-situ.

(18) a. Wh-in-situ object question

e-i-na-þe u na-k ?
REL(-FUT)-1SG-AUX-say what to-2SG.M
‘What shall I say to you?’ (AP Chaîne no. 28, 5:25)
b. Wh-in-situ prepositional object question
e-k-nİu tnn, pa-son ?
REL(-PRES)-2SG.M-come where DEF.SG.M.1SG-brother
‘Where did you come from, my brother?’ (Ac. A&P 198:64–65)
c. Wh-in-situ adverb question
‹nt-a-k-ei e-pei-ma ‹n-aš ‹n-he ?
REL-PERF-2SG.M-go to-DEM.SG.M-place in-what of-manner
‘How did you get to this place?’ (B. Martyr. 206, 29)

Wh-in-situ is not restricted to matrix questions but can also occur in em-
bedded questions. When the relative marker surfaces to the right of the
200 Chris H. Reintges

subordinating complementizer þe ‘that’, the wh-in-situ phrase takes the

embedded scope and the resulting interpretation is that of an indirect ques-
tion.

(19) n-ti-sow‹n an [ þe nt-a u šnpe mmo-s ]

NEG-(PRES-)1SG-know NOT COMP REL-PERF what happen to-3SG.F
‘I don’t know what happened to her.’ (Hilaria 7, 30–31)

As we can see from (20), wh-in-situ questions with relative TAMs can be
further modified by an interrogative particle like eye.

(20) eye ere ne.t‹n-šİre nuþe eȕol h‹n nim ?

Q REL (PRES-)DEF.P.2P-son cast PCL in who
‘In whom are your sons casting out (demons)?’ (Luke 11, 19)

Coptic question particles are compatible with wh-in-situ, wh-fronting, as

well as wh-cleft constructions. In view of their tolerance towards the sur-
face position of the wh-phrase, they do not seem to be involved in any kind
of licensing relationships. In conveying positive or negative presupposi-
tions as well as corroborative focus readings, such interrogative particles
are closely connected to the modal domain (Reintges 2003, 378ff.).

5.2. Declarative focus constructions

The syntactic derivation of declarative focus structures runs entirely paral-

lel to that of wh-questions, with the clause-internal placement of the focus
phrase conditioning relative TAMs. In the longstanding research tradition on
focus, question-answer sequences have been used as diagnostic tools for
identifying the new information focus within a clause. The focus corre-
sponds to the sentence element in the response that provides a value for the
variable introduced by the wh-constituent in the question while the presup-
position underlying the question and the corresponding answer remains
constant (Chomsky 1971, Jackendoff 1972, Erteshik-Shir 1986).

(21) Q: e-r-ȕİk e-tnn ?

REL-(PRES)-2SG.F-goSTAT to-where
‘Where are you (woman) going to?’
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 201

A: e-i-ȕİk e-p‹-topos
REL(-PRES)-1SG-goSTAT to-DEF.SG.M.-shrine of
‹n Apa Mİna ‹nta-šlİl.
Apa Mêna CONJ.1SG-pray
‘I am on my way TO THE SHRINE OF APA MÊNA to pray.’
(Mêna, Mirc. 27b:22–25)

As Rooth (1992, 84) puts it, focus in an answer evokes a set of alternatives
that qualify as potential answers in the context of the question. In doing so,
it marks a contrast between the asserted answer and other potential answers
(in our example, other holy places to go to).
The same situation applies to exhaustive listing focus, which specifies
an exhaustive set of which the proposition holds true and excludes other
possibilities (Kuno 1972, É. Kiss 1998). Its prototypical syntactic frame in
Coptic is the ‘not X but Y’ construction which provides a straightforward
way of rejecting a previous utterance and offering an alternative specifica-
tion of the variable (Horn 1989, Erteschik-Shir 1997, Herburger 2000).

(22) ‹mpnr pa-šİre, ‹mp-u-toš-‹k gar

no DEF.SG.M.1SG-son NEG.PERF-3PL-appoint-2SG.M PCL
e-ti-oikonomia, alla ‹nt-a p‹-þois toš-‹k
to-DEM.SG.F-service but REL-PERF DEF.SG.M-lord appoint-2SG.M
e-u-sols‹l ‹n-ne-snİu et-waaȕ (…)
to-INDEF.SG-comfort for-DEF.PL-brothers COMPREL-be.holySTAT
‘No, my son, for you have not been destined (lit. they have not
destined you) for this career (as a hermit), but the Lord has
appointed you AS A COMFORT FOR THE HOLY BROTHERS (…)’
(B. Martyr. 216, 33–217, 1)

Relative TAMs are found in presentational focus contexts, in which a new

discourse referent is introduced “out of the blue”. In including only infor-
mation that is not construable from the previous context, presentational
focus differs systematically from the new information focus in the response
to a preceding question (Green & Reintges 2005). 8

(23) e-w‹na-f-s h‹n n-et-ka ma na-f

REL-HAVE-3SG.M-3SG.F among DEF.PL-COMPREL-put place for-3SG.M
‘He has it (a place) among THOSE WHO MAKE A PLACE FOR HIM.’
(Shenoute III 85:14)
202 Chris H. Reintges

Relative TAMs may also be used as narrative tenses, indicating that a par-
ticular event is of special relevance for the subsequent discourse (Hopper
1979, Reinhart 1984). Such foregrounded event chains typically respond to
cause/reason questions that require a discourse as a felicitous answer. In the
following discourse fragment, the relative TAM marker ere appears in a
present tense sentence with a stative verb form (kyal2w ‘to be entrusted’),
which is clearly atelic (Reintges 2004, 216f. §6.2.3, 259-60 §7.3.2.2). The
selection of a relative TAM in this context seems to be motivated by the
desire to single out the situation that is crucial for the development of the
narrative.

(24) Q: ahro-s tei-h‹ll2 [ e-s-rime n-tei-he ] ?

why-3SG.F DEM.SG.F-old.woman REL(-PRES)-3SF-weep in-DEM.SG.F-way
‘What about this old woman who is weeping in such manner?’
A: eti ere pe.s-hai onah ere
ADV REL (PRES-)DEF.SG.M-3SG.F-husband be.alive REL
h‹n-‹nka kyal2w ero-f ‹nte u-r2me
(PRES-) INDEF.PL-thing enthrustSTAT to-3SG.M of INDEF.SG-man
a-f-mu de h‹n-u-š‹p-en-š2p ‹nweš ‹n-šaþe (…)
PERF-3SG.M-die PCL suddenly without of-word
‘What about her, this old woman, who is weeping in such a man-
ner? When her husband was still alive, someone’s things were en-
trusted to him. But he died all of a sudden without a word (…)’
(AP Chaîne no. 225, 65:16–18)

The generalization that emerges from the previous discussion is that in-situ
focus in Coptic Egyptian is compatible with argument, predicate, and sen-
tence focalization, with the obtained new information or contrastive focus
interpretation being governed by the discourse context (see Green & Jaggar
2003 for parallel facts in Hausa).

5.3. Interrogative focus constructions

Coptic yes-no questions fall into two groups: those which are introduced by
interrogative particles and those that lack any morphological marking of
interrogative force. Both types of yes/no questions may condition relative
tenses, as seen in (25a–b).
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 203

(25) a. e-k-na-kynšt nsn-i e-i-na-ȕnk

REL(-FUT)-2SG.M-AUX-look after-1SG REL(-FUT)-1SG-AUX-go
e-p-tako ? mpnr, pa-þoeis!
to-DEF.SG.M-perdition NO DEF.SG.M.1SG-lord
‘Would you watch me rushing to perdition? Oh no, my Lord!’
(Ac. A&P 200, 90–91)
b. mİ e-k-na-kaa-t nsn-k
Q REL(-FUT)-2SG.M-AUX-put-1SG behind-2SG.M
etȕe u-hnȕ mmate ?
for INDEF.SG-thing only
‘Will you abandon me (lit. put me behind you) BECAUSE OF ONE
THING ONLY?’ (Ac. A&P 200, 81–82)

If relative tenses were connected to interrogative force, their broad syntac-

tic distribution across non-interrogative sentence patterns (relative clauses,
declarative focus constructions) would be left unexplained. For this reason,
it seems more promising to relate their presence in yes-no questions to fo-
cus rather than to clause-typing. In (25b), for instance, the exhaustive list-
ing interpretation of the cause adverbial etȕe u-hnȕ ‘because of one thing’
is brought about by the focus-sensitive particle mmate ‘only’ (Herburger
2000, 105ff.).
Alternative questions entail a choice between two members of a disjunc-
tive set that contradict each other. The selection of one member will there-
fore automatically eliminate the alternative option (Erteschik-Shir 1997).
As we can see from (26), the contrastive focus reading of alternative yes-no
questions may condition the presence of relative TAMs.

(26) p‹-þoeis e-k-tšn na-n ‹n-tei-paraȕolİ

DEF.SG.M-lord REL(-PRES)-2SG.M-say to-1PL PREP-DEM.SG.F-parable
þen e-k-tšn ‹mmo-s e-won nim ?
or REL(-PRES)-2SG.M-say PREP-3SG.F to-one every
‘Oh Lord, do you tell this parable TO US, or do you tell it TO
EVERYONE?’ (Luke 12, 41)

What we seem to be dealing with is a case of apparent optionality: the rele-

vant trigger for relative TAMs in yes-no questions seems to be the presence
of focus.
204 Chris H. Reintges

6. The “hidden” movement configurationality of Coptic wh-in-situ

In contrast to Hausa relative aspects and Chamorro wh-agreement, Coptic

relative tenses are restricted to focus/wh-in-situ contexts. Elaborating on
Reintges, LeSourd & Chung’s parametric analysis of special inflection
(2006), I will argue in what follows that Coptic wh-in-situ instantiate a
“hidden” movement configuration where syntactically invisible wh-
movement leaves a footprint in the relative TAM marking. The main evi-
dence for a movement approach to Coptic wh-in-situ comes from the paral-
lelism in scope and interpretation between the wh-in-situ and the wh-
fronting pattern.

6.1. Wh-movement and interrogative interpretation

The construal of a genuine question interpretation of wh-items supports a

movement analysis of Coptic wh-in-situ. Consider the following triplets of
examples which involve the wh-words nim ‘who’ and u ‘what’. In the a-
examples, the wh-word remains in-situ and relative marking surfaces in the
left periphery. As we can see from comparison with the b-examples, the
wh-item is interpreted as a question word in exactly the same way as its
fronted counterpart. This contrasts with the c-examples in which the TAM
auxiliary assumes the non-relative form and the wh-in-situ phrase is inter-
preted as a specific indefinite. Such an indefinite interpretation is not avail-
able for fronted wh-phrases (Reintges 2003, 380ff., 2004, 144ff.§ 4.2.1.).

(27) a. nt-a nim þpo na-f n-tei-hypomenİ ?

REL-PERF who achieve for-3SG.M PREP-DEM.SG.F-endurance
‘Who has achieved for himself such endurance?’ (Hilaria 12, 29)
b. nim a-f-ent-k e-pei-ma ?
who PERF-3SG.M-bring-2SG.M to-DEM.SG.M-place
‘Who brought you here?’ (KHML I 3:7–8)
c. ša-u-eime [ þe a nim p-šİre
HAB-3PL-know COMP PERF who DEF.SG.M-child
‹n-nim þrn h‹m p-agnn ]
of-who win in DEF.SG.M-contest
‘They know that so-and-so, the son of so-and-so, won the
contest.’ (B. Apoc. 134, 14–15)
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 205

(28) a. e-i-na-ti u na-k ?

REL(-FUT)-1SG-AUX-give what to-2SG.M
‘What shall I give you?’ (Genesis 30, 31)
b. u se-þi wa ero-i h‹m
what (PRES-)3PL-speak malice against-1SG in
p-e-ti-š‹p hmot anok haro-f ?
DEF.SG.M-COMPREL(-PRES-)1SG-take grace I for-3SG.M
‘What could they say maliciously against me because of what I
give thanks (lit. take grace) for?’ (1 Cor. 10, 30)
c. a-i-ti u mn u ehun e-pei-ma.
PERF-1SG-give what and what PCL to-DEM.SG.M-place
‘I gave such and such thing to this place.’(Shenoute IV 105, 16)

Since both wh-fronting structures and relative-marked wh-in-situ construc-

tions are associated with an interrogative interpretation, there is compelling
evidence for equating the co-occurrence of wh-in-situ and relative tenses
with a movement configuration. By contrast, the specific indefinite inter-
pretation of in-situ wh-words instantiates the non-movement option.

6.2. Wh-movement and scope

The scope of an overtly moved wh-phrase is contingent on its landing site.

When a wh-phrase moves overtly to the embedded focus position to the
right of the subordinator þe ‘that’, as in (29a), it takes embedded scope and
the entire construction is interpreted as an indirect question. If, on the other
hand, the wh-phrase undergoes long-distance wh-movement across a clause
boundary, as in (29b), it takes matrix scope and the resulting interpretation
is that of a direct question. 9

(29) a. ti-þ‹nu ‹mmn-t‹n [þe h‹n u ‹n-šaþe

(PRES).1SG-ask PREP-2PL COMP with what of-word
a-tent‹n-mute ero-i …]
PERF-2PL-say about-1SG
‘I ask you with what reason do you say about me...’(Acts 10, 29)
b. ‹n-aš ‹n-he ‹ntok k‹-tšn ‹mmo-s
in-which of-manner you.SG.M (PRES-)2SG.M-say PREP-3SG.F
[þe tet(n)-na-‹r r‹mhe] ?
COMP (FUT-)2PL-AUX-make free.man
‘How do you say that you will become free?’ (John 8, 33)
206 Chris H. Reintges

Parallel facts can be observed for embedded wh-in-situ questions. In finite

complement clauses, the wh-in-situ constituent generally takes embedded
scope, which yields an indirect question interpretation. The relative marker
appears after the complementizer þe ‘that’.

(30) ‹mpe-f-þoo-s [ þe ‹nt-a-f-kaa-f tnn ]

NEG.PERF-3SG.M-say-3SG.F COMP REL-PERF-3SG.M-put-3SG.M where
‘He did not tell where he had put it.’ (AP Chaîne no 235, 65:18)

However, examples are also attested where the in-situ wh-phrase scopes out
of the embedded finite clause and the entire construction is interpreted as a
direct question. In this context, the relative TAM marker surfaces at the left
edge of the higher clause.

(31) eye ntntn e-tetn-þn mmo-s ero-i [ þe ang nim]?

Q you.PL REL-2PL-sayPREP-3SG.F about-1SG COMP I who
‘Who are you saying of me that I am?’ (Mark 8, 29)

The matrix scope of embedded wh-in-situ and the presence of relative TAMs
in the matrix clause indicate that successive-cyclic movement through the
specifiers of CP has taken place. The Coptic facts therefore provide
counterevidence to Rackowski & Richards’ (2005) claim that morpho-
syntactic flagging reflects an AGREE relation within the verbal (vP) domain
from which the wh-phrase has been extracted.
A different situation obtains in infinitival wh-in-situ questions, in which
only the direct question interpretation seems to be available. Once again,
relative TAMs flag the highest clause over which the wh-in-situ takes scope.
Since they lack TAM markers, Coptic infinitival clauses do not project an
articulated left periphery, including the FINP and the topic-focus field.
Since there is no designated scope position inside the infinitival clause, the
embedded wh-in-situ phrase must move to the specifier position of the ma-
trix FOCP, where its scope is determined (McDaniel 1989, Dayal 1991).

(32) a. e-tet‹n-wnš [ e-tra-ka nim nİ-t‹n

REL(-PRES)-2PL-want to-CAUS.INF.1SG-place who for-2PL
eȕol h‹m pe-snau ] ?
PCL from DEF.SG.M-two
‘Who do you want me to release to you out of the two?’
(Matthew 27, 21)
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 207

b. e-k-wnš [ e-tre-n-snȕte na-k

REL(-PRES)-3SG.M-want to-CAUS.INF-1PL-prepare for-2SG.M
tnn ‹m-p‹-paskha e-wom-f ] ?
where PREP-DEF.SG.M-Paskha.meal to-eat.INF-3SG.M
‘Where do you want us to prepare the Pasha meal for you to eat?’
(Matthew 26, 17)

To conclude, the absence of clear semantic contrasts between wh-in-situ

and wh-fronting questions suggests that both patterns are derived by the
same movement operation. In wh-in-situ questions, this movement process
is syntactically invisible since the wh-phrase is apparently frozen in place.
Yet, covert wh-movement feeds overt morphology, triggering the spell-out
of relative TAMs.

6.3. The syntactic derivation of wh-in-situ and wh-fronting structures

The previous discussion on scope and interrogativity raises a non-trivial

question concerning the representational level at which covert wh-
movement is operative: does it take place at narrow syntax or at LF? As-
suming with Chomsky (2005, 16ff.) that LF operations have no access to
the morpho-phonological cycle, the presence of relative TAMs in Coptic
wh-in-situ questions clearly shows that covert movement operation takes
place in the narrow syntax before the derivation is shipped to PF.
Adopting the copy theory of movement, Reintges, LeSourd & Chung
(2006) derive the word order contrast between wh-in-situ and wh-fronting
constructions from different pronunciation sites of the scope-taking opera-
tor: the phonological realization of the overtly moved wh-phrase in the
target position (Spec,FOCP) produces wh-fronting structures while the pho-
nological realization of the relevant operator in the original position gives
rise to the hidden movement configurationality of wh-in-situ. Moreover, the
semantic non-distinctness of the wh-in-situ and the wh-ex-situ patterns re-
ceives an immediate explanation since the spell-out of the moved wh-
phrase is a matter of the morpho-phonological rather than the semantic
component. 10
In current Minimalist views on movement, in order for one category (the
goal) to move to the vicinity of another (the probe), the two must stand in
an agreement relation through which features on maximal projections and
features on heads are matched. Both the probe and the goal have active
(that is, uninterpretable or unvalued) features whose values can be set by
208 Chris H. Reintges

the corresponding features of the other. Subsequently, movement is imple-

mented to bring the goal into a local configuration with the probe, an opera-
tion that leaves the original copy of the goal untouched. Once the probe and
the goal are in a local configuration within a single projection, the uninter-
pretable features on both categories are erased (Chomsky 2000, 122ff.,
2001, 4f.).
If this theory is applied to the case at hand, the probe for wh-movement
is the left-peripheral focus head and the goal is the wh-phrase. With respect
to their feature specification, I assume, following Chomsky (2000, 128),
that wh-phrases have an uninterpretable feature [uwh] and an interpretable
interrogative feature [Q] that matches the uninterpretable interrogative fea-
ture [uQ] on the focus head. 11 It may very well be the case that the focus
head inherits its uninterpretable [uQ] feature from the force-indicating C0
head (Chomsky 2005, 9). If the uninterpretable wh-feature on the in-situ
wh-phrase cannot be eliminated via movement to the designated scope po-
sition, it can no longer receive an interrogative interpretation. As a result,
such wh-in-situ items in Coptic can only receive a specific indefinite inter-
pretation.
Movement of the wh-phrase to the specifier position of the focus phrase
yields the proper checking configuration. If the lowest copy of the wh-
phrase is pronounced, then the relative marker is spelled out overtly, as
schematically represented in diagram (33). (Strikethrough indicates mate-
rial that is left unpronounced. The relative marker on the TAM word is rep-
resented as REL. Arrows indicate the feature-matching relationship between
the focus head and tense.)

(33) Covert wh-movement with relative tense marking

[ C0 [FOCP WH [ FOC REL ][TP TPERF [MODP SU [MOD [VP SU] [VP V WH]]]]

feature T0-to-Foc0 movement

inheritance

If, on the other hand, the highest copy of the wh-phrase is pronounced, then
the relative marking, which would occur in the same projection, is left un-
pronounced, as schematically represented in diagram (34).
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 209

(34) Overt wh-movement without relative tense marking

[ C0 [FOCP WH [FOC REL] [TP TPERF [MODP SU [MOD [VP SU] [VP V WH ]]]]]

feature
inheritance

The analysis proposed here departs from Chomsky’s (2001) “Derivation by

Phase” model, according to which wh-movement proceeds successive-
cyclically through different derivational domains (or phases), including the
verb phrase with full argument structure and the CP. Phases are domains
that are eligible for phonological realization, which means that a constituent
has to target the left edge of each phase in order to undergo subsequent
movement.
There is, however, no independent evidence in Coptic Egyptian for the
phasal status of the vP, as argued in Rackowski & Richards (2005). In lack-
ing tense and finiteness, Coptic verbs seem to be defective categories. All
these features are morphologically manifest on the TAM word and thus de-
fine opaque domains for core syntactic processes like case and extraction
(cf. Reintges 2001). Complementizer alternations provide strong evidence
for a feature matching agreement relation between the left-peripheral FOC0
head, which acts as a probe for wh-movement, and finite tense. It is clear
that the probe enters into some agreement relationship not only with the
goal (the wh-phrase) but also with the head of FINP which is productively
involved in the formation of wh-dependencies (cf. Chomsky 2004, 116).

7. Spellout conditions on special inflection

In the canonical wh-in-situ pattern, the relative TAM overtly marks a local
specifier-head relationship between the designated functional head and the
topmost copy of the displaced wh-phrase which is not phonologically real-
ized. The question that arises is why relative TAMs are systematically ab-
sent in wh-fronting structures where the topmost copy of the moved wh-
phrase is phonologically realized. In this section, I propose an economy
explanation for the complementary distribution between wh-fronting and
relative TAMs, which registers the presence of operator-variable dependen-
210 Chris H. Reintges

cies. I will then briefly discuss Tuller’s (1986) specificity filter for Hausa
relative aspect inflection. 12

7.1. An economy condition on morpho-syntactic flagging

Adopting the late insertion view of Distributed Morphology, I assume that

the syntactic component is essentially phonology-free (Halle & Marantz
1993). The core task of the syntax is to create structural dependencies
through syntactic operations while spell-out operations are confined to the
morpho-phonology (Bobaljik 2002). Spell-out operations fall into two
classes, namely (i) the phonetic realization of copies in the course of lexical
insertion, and (ii) the conversion of a hierarchical structure into linear or-
der. The decision which copy to pronounce may be dictated by language-
specific constraints but may also fall out from the language’s parameter
setting. In short, nothing is deleted, but not everything is pronounced.
On this view, two issues concerning phonological realization of wh-
movement structures are involved: firstly, whether the higher or the lower
copy of a moved wh-phrase is pronounced, and secondly, which of the cop-
ies triggers the morphological effects of special inflection. In Reintges,
LeSourd & Chung’s (2006, 183f.) analysis, the relative marker is inserted
into the structure roughly at the point when the moved wh-phrase is pro-
nounced. In other words, in Coptic, the spell-out of wh-movement struc-
tures is governed by two language-specific morphological requirements. On
the one hand, the creation of operator-variable dependencies in the narrow
syntax must be marked overtly by the pronunciation of the topmost copy or
by morphological flagging. On the other hand, there is an economy condi-
tion preventing the over-coding or double marking of such dependencies by
the simultaneous spell-out of both the topmost wh-copy and relative TAM
marking.
This situation is reminiscent of the Doubly-filled COMP Filter in English
since either the head or the specifier of the target projection of wh-
movement can be realized, but not both (Chomsky & Lasnik 1977). Alter-
natively, one might think of an economy condition of wh-scope marking
where the scope of a moved wh-phrase is either marked overtly by the pro-
nunciation of the attracted category in the target position or by inserting a
morphological scope marker into the attracting Foc0 head.
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 211

7.2. A specificity filter for Hausa relative aspects?

The economy condition of morpho-syntactic flagging does not carry over to

Hausa, where wh-fronting requires relative aspect marking, while this is
prohibited in the corresponding Coptic structures. Conversely, relative as-
pect marking is blocked in Hausa in non-echo wh-in-situ questions but
obligatory in Coptic. Thus, Hausa is the “mirror image” of Coptic with
respect to the syntactic conditions on special inflection (Green & Reintges
2004a, 2005, Reintges & Green 2004).

(35) a. Wh-fronting with relative aspects

ìnƗ sukà jƝ ?
where 3PL-REL-PERF go
‘Where have they gone?’ (Jaggar 2001, 517)
b. Wh-in-situ with general aspects
sunƗ̖ fitôwa dàgà ìnƗ ?
3PL-IMPERF come.out.VN from where
‘Where are they coming out from?’ (Newman 2002, 496)

In Hausa, overt wh-movement feeds overt morphology while covert move-

ment leaves no morphological footprint. The prohibition against relative
aspects in wh-in-situ questions can be accounted for if wh-in-situ constitu-
ents move to their target position at LF, which has no direct access to the
morphological operations.
To exclude the combination of moved wh-phrases and non-relative as-
pects, Tuller (1986, 73 and 109ff.) postulates an interpretative filter for
Hausa which prohibits [–definite] heads from being specified by the feature
[focus]. In her system, operators in the specifier position of CP are assigned
focus. The relative aspect qualifies as a [+definite] head, witnessed by the
narrative use of the relative perfective to describe specific events in the
past. As a [+definite] head, the relative aspect in T0 is compatible with the
focus features associated with the displaced wh-phrase in Spec,CP.
The Coptic facts are problematic for Tuller’s analysis. As we have seen
in section 5.2, relative TAMs may be used as narrative tenses in much the
same way as Hausa relative aspects and would therefore be eligible as a
[+definite] head. Nevertheless, it is impossible to use this special inflection
in Coptic focus/wh-fronting contexts. This suggests that special inflection
in both languages is not controlled by semantically driven filters but rather
functions as a morpho-syntactic flagging device for operator-variable de-
212 Chris H. Reintges

pendencies. As such, relative TAMs are subject to language-particular spell-

out conditions.

8. Conclusions

Coptic provides yet another example of a language in which classical wh-

constructions (relative clauses, wh-questions, declarative focusing construc-
tions) are flagged by relative tense-aspect-mood categories. The relative
marker could be identified with a finite relative complementizer, which is
inserted into the head of a designated focus phrase in the clausal left-
periphery. The presence of operator-variable dependencies provides a nec-
essary, but not sufficient condition for the presence of morpho-syntactic
flagging. The presence of relative TAMs is required in wh-in-situ questions
while they are absent in wh-fronting structures. Coptic hosts either the
fronted wh-phrase in the specifier position or the relative marker in the
head position of the left-peripheral focus phrase, but these must not co-
occur within the same projection.
Coptic wh-in-situ questions display the typical characteristics of wh-
movement as far as scope and interrogative readings are concerned. The
hidden movement configuration of wh-in-situ has been accounted for in
terms of covert syntactic movement and the phonological realization of the
moved wh-phrase in the base position. Syntactically invisible wh-
movement feeds overt morphology in the form of relative TAM marking.
From this perspective, the difference between ex-situ and in-situ construc-
tions reduces to the question of where the moved wh-phrase is phonologi-
cally realized, viz. in the base position or in the target position of move-
ment. Since Coptic wh-in-situ is derived by wh-movement followed by the
pronunciation of the lowest wh-copy, the presence of relative tenses as the
morphology of extraction falls into place.

Notes

Many thanks to Victor Manfredi and Malte Zimmermann for detailed

written comments on a previous version of the present paper. For discus-
sion on the Coptic facts, I am indebted to Lisa Cheng, Sandy Chung,
Hamida Demirdache, Melanie Green, Phil Jaggar, and Lutz Marten. I am
also grateful to Jeroen van de Weijer for the correction of my English. The
usual disclaimers apply. The research reported here is funded by the
Coptic relative tenses: The profile of a morpho-syntactic flagging device 213

Dutch Organization for Scientific Research N.W.O. (Vidi-grant 276-70-

008).
1. The following abbreviations are used in the glosses: 1 ‘first person’; 2
‘second person’; 3 ‘third person’; AGR ‘agreement’; AUX ‘auxiliary verb’;
CAUS.INF ‘causative infinitive’; COMP ‘subordinating complementizer’;
COMPQ ‘interrogative complementizer’; COMPREL‘relative complementizer’;
COND ‘conditional’; CONJ ‘conjunctive’; FUT ‘future’; IMPERF ‘imperfec-
tive’; INF ‘infinitive’; INTERJ ‘interjection’; IOM ‘indirect object marker’;
NOM ‘nominative case’; OBJ ‘objective case’; OBL ‘oblique case’; PCL
‘particle’; PERF ‘perfect’; PL ‘plural’; PREP ‘preposition’; PRES ‘present
tense’; Q ‘yes/no question particle’; REL ‘relative marker’; SG ‘singular’;
SUBJ ‘subjunctive’; VN ‘verbal noun’; WH ‘wh-agreement’. Glosses are
given in parentheses for morphemes that have no surface-segmental
shape.
2. Coptic Egyptian is the indigenous language of late-antique and early me-
dieval Christian Egypt (from about the 3rd to the 11th century CE) and
represents the most recent stage of the Ancient Egyptian language. Coptic
is actually a dialect cluster with at least six regional varieties, two of
which gained supra-regional importance: Sahidic Coptic, the vernacular of
Upper Egypt, and Bohairic Coptic, the vernacular of Lower Egypt. The
data in this study are exclusively taken from Sahidic Coptic, which, due to
its early records and rich literature represents the main reference dialect.
3. Jaggar (2001, 162 fn. 4,5) suggests that the term “relative” be replaced
with the cover term “focus” to provide a semantic characterization of the
environments in which relative aspects are grammatically allowed and/or
required. Layton (2000, 352–5) re-analyzes Coptic declarative and inter-
rogative main clauses that contain relative TAMs (traditionally called “sec-
ond tenses”) as focusing constructions, which does, however, not include
relative constructions, temporal adverbial clauses, and secondary predi-
cates. Although Jaggar’s reductionist approach is appealing from a con-
ceptual point of view, it is not entirely clear or obvious whether a seman-
tic ‘focus’ analysis carries over to relative clause constructions, where
relative aspect marking is obligatory, or conditional and temporal adjunct
clauses, where it is optional. It is worth pointing out that the information
structure of this construction type has received alternative analyses in
terms of topicalization (see Green & Reintges 2004b, 86f. for further dis-
cussion).
4. According to McCloskey (1990), the phonological and morphological dis-
tinction between the alternating complementizers is sometimes not re-
flected in the phonological shape of the complementizer itself but rather in
the initial phonological mutation induced on the following sentence ele-
ment (typically a verb). The abbreviations aL and aN reflect the fact that
the former complementizer induces lenition on the adjacent verb whereas
214 Chris H. Reintges

the latter induces nasalization on the initial segment of a following verb.

Complementizer alternations are subject to dialect variation (McCloskey
1990, 242–3, fn.7, 2001, 68, fn.2).
5. Although this generalization holds for most Hausa speakers, Tuller (1986,
120) notes that some speakers permit relative aspect marking to surface on
every inflectional head in the path of movement. This appears to represent
a case of dialect variation since most speakers of Standard (Kano) Hausa
do not produce such constructions (Green & Reintges 2005).
6. Victor Manfredi proposes to treat the relative marker e– “not as a mor-
pheme of any kind, but rather as an epenthetic mora motivated by phras-
ing (i.e. assuming prosodic conditioning, analogous to the destressing pat-
tern of anaphora in English)”. In fact, the e– formative derives historically
from the aspectual auxiliary jw ‘to come’, which performs various clause-
chaining and subordinating functions in Pre-Coptic Egyptian. Both the
complementizer allomorphs ‹nt– and et– come from the relative comple-
mentizer ntj ‘that’, which is itself derived from the genitival linker nj ‘of’.
See Polotsky (1944) and Loprieno (1995) for a more detailed discussion
of the precursors of Coptic relative TAMs.
7. A question remains with respect to the auxiliary verb na– ‘go’, which in-
dicates future tense. I assume that the so-called First Future f-na-sot‹m ‘he
is going to hear’ is actually a present progressive. On this view, the null
morpheme of the present tense is incorporated into the relative marker e–
in the Foc0-node.
8. Indefinite subjects that correspond to presentational focus are generally
not construed with relative TAMs. The reason for this is that indefinite sub-
jects trigger the insertion of the copular verbs w‹n ‘be’ and m‹n ‘not be’
into the structure in the present and the future tense, which thus assume
the form of existential sentences (Reintges 2004, 259 §7.3.2.1).
(i) a. wen u-noky ‹n-r2me š2ne h‹m p‹-palation
be INDEF.SG-great of-man be.sick in DEF.SG.M-palace
‘A nobleman in the palace who became sick.’ (Hilaria 10, 27)
b. w‹n u-m‹nt.eȕeiİn na-taho-u
be (FUT-)INDEF.SG-misery AUX-come.upon-3p
‘A misery will come upon them.’ (V. Pachôm. 90, 28–91, 1)
The complementary distribution between copula support and special in-
flection in presentational subject focus contexts needs to be clarified in fu-
ture research.
9. Since Coptic is a dead language, it is impossible to establish that wh-in
situ questions are sensitive to islands. However, there are no attested ex-
amples of wh-in-situ questions in which the wh-phrase originates inside an
island and takes scope outside the island. In wh-fronting structures, the
long-distance movement of a topical wh-phrase out of an adjunct clause is
saved by resumptive pronominalization, as seen in (i).
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 215

(i) nimi gar ‹mmn-t‹n [e-fi-weš-ket u-pyros]

who PCL among-2PL REL(-PRES).3SG.M-want-build INDEF.SG-tower
mİ n-fi-na-hmoos an ‹n-šor‹p
Q neg(-fut)-3sg.m-aux-sit not at-first
n‹-fi-fi p-np ‹n-te.f-danapİ?
conj-3sg.m-carry def.sg.m-number of-def.sg.f.3sg.f-cost
‘Who among you, who intends to build a tower, would he not first sit
and estimate the total amount of its costs?’ (Luke 1, 28)
10. The possibility of pronouncing the lowest, as opposed to the highest copy
of a moved wh-phrase has been explored by Groat & O’Neil (1996), Pe-
setsky (1997, 1998, 2000), Bobaljik (2002), Nunes (2004), and various
others. The phonetic realization of the lowest wh-copy is widely believed
to be a last resort option, available only when the pronunciation of the
highest chain link does not result in a convergent derivation at PF (Bošk-
oviü 2002, Nunes 2004). The Coptic facts clearly show that lower copy
pronunciation may instantiate a parameterized option of a language’s wh-
interrogative paradigm (Reintges, LeSourd & Chung 2006).
11. This feature composition implements the idea that wh-movement is a
means of clause typing, which specifies a clause as a question. The feature
specification of the attracting probe, however, is a matter of debate in the
current literature. For instance, Cheng & Rooryck (2000) argue that the
C0 node has an interpretable [Q] feature and an unvalued [wh] feature.
12. I am indebted to Malte Zimmermann for clarifying the issues concerning
the spell-out conditions on special inflectional morphology addressed in
this section.

References

Aboh, Enoch Oladé

2004 The morphosyntax of complement-head sequences. Clause structure
and word order patterns in Kwa: Oxford studies in comparative syn-
tax. Oxford, New York: Oxford University Press.
Bobaljik, Jonathan D.
2002 A-chains at the PF-interface: copies and ‘covert’ movement. Natural
Language & Linguistic Theory 20: 197–267.
Boškoviü, Željko
2002 On multiple wh-fronting. Linguistic Inquiry 33: 351–383.
Browning, Maggie
1987 Null operator constructions. Ph.D. dissertation. MIT, Cambridge,
Mass.
216 Chris H. Reintges

Cheng, Lisa Lai-Shen

1991 On the typology of wh-questions. Ph.D. dissertation. MIT, Cam-
bridge, Mass.
Cheng, Lisa Lai-Shen, and Johan Rooryck
2000 Licensing wh-in-situ. Syntax 3: 1–19.
Chomsky, Noam
1971 Deep structure, surface structure, and semantic interpretation. In
Semantics: an interdisciplinary reader in philosophy, linguistics and
psychology, Danny D. Steinberg and Leon Jakobovits (eds.), 183–
216. Cambridge: Cambridge University Press.
1977 On wh-movement. In Formal syntax, Peter W. Culicover, Thomas
Wasow, and Adrian Akmajian (eds.), 71–132. New York: Academic
Press.
2000 Minimalist inquiries: the framework. In Step by step: Essays on
minimalist syntax in honor of Howard Lasnik, Roger Martin, David
Michaels, and Juan Uriagereka (eds.), 89–155. Cambridge, Mass.:
MIT Press.
2001 Derivation by Phase. In Ken Hale: A Life in Linguistics, Michael
Kenstowicz (ed.), 1–52. Cambridge, Mass.: MIT Press.
2004 Beyond explanatory adequacy. In Structures and beyond, Adriana
Belletti (ed.), 104–131. The carthography of syntactic structures.
Vol. III. Oxford: Oxford University Press.
2005 Three factors in language design. Linguistic Inquiry 36, 1–22.
Chomsky, Noam, and Howard Lasnik
1977 Filters and control. Linguistic Inquiry 8: 425–504.
Chung, Sandra
1994 Wh-agreement and “referentiality” in Chamorro. Linguistic Inquiry
25: 1–44.
1998 The design of agreement: evidence from Chamorro. Chi-
cago/London: The University of Chicago Press.
Cinque, Guglielmo
1990 Types of A'-dependencies. Linguistic Inquiry monograph 17. Cam-
bridge, Mass.: MIT Press.
Dayal, Veneeta
1991 Subjacency effects at LF: The case of Hindi wh. Linguistic Inquiry
22: 762–770.
Erteschik-Shir, Nomi
1986 Wh-questions and focus. Linguistics and Philosophy 9: 117–149.
1997 The dynamics of focus structure. Cambridge: Cambridge University
Press.
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 217

Green, Melanie
1997 Focus and copular constructions in Hausa. Ph.D. dissertation.
School of Oriental and African Studies, London. To appear in Publi-
cations of the Philological Society. Oxford: Blackwell.
Green, Melanie, and Philip J. Jaggar
2003 Ex-situ and In-situ Focus in Hausa: syntax, semantics and discourse.
In Research in Afroasiatic grammar II: Selected Papers from the
Fifth Conference on Afroasiatic Languages, Paris 2000, Jacqueline
Lecarme (ed.), 187–214. Amsterdam/Philadelphia: Benjamins.
Green, Melanie, and Chris H. Reintges
2004a Syntactic anchoring in Hausa and Coptic wh-constructions. In Pro-
ceedings of the 27th Annual Meeting of the Berkeley Linguistics Soci-
ety, Andrew Simpson (ed.), 61–72. University of California, Berke-
ley.
2004b Review article of Jaggar (2001). Lingua 114: 44–91.
2005 Syntactic conditions on special inflection: evidence from Hausa and
Coptic Egyptian interrogative and focus constructions. University of
Sussex Working Papers in Linguistics and English Language 14 at
http://www.sussex.ac.uk/linguistics/1-4-1.html.
Groat, Erich, and John O’Neil
1996 Spell-out at the LF interface: achieving a unified syntactic computa-
tional system in the Minimalist framework. In Minimal ideas: syn-
tactic studies in the Minimalist framework, Werner Abraham, Sam-
uel D. Epstein, Höskuldur Thráinsson, and Jan-Wouter Zwart (eds.),
113–139. Amsterdam/Philadelphia: Benjamins.
Haïk, Isabelle
1990 Anaphoric, pronominal and referential INFL. Natural Language and
Linguistic Theory 8: 347–374.
Halle, Morris, and Alec Marantz
1993 Distributed Morphology. In The view from building 20: essays in
honour of Sylvain Bromberger, Kenneth Hale and Samuel J. Keyser
(eds.), 111–176. Cambridge, Mass.: MIT Press.
Hartmann, Katharina, and Malte Zimmermann
this vol. Exhausitivity marking in Hausa: A reanalysis of the particle nee/cee.
Herburger, Elena
2000 What counts: focus and quantification. Cambridge, Massachusetts:
MIT Press.
Hopper, Paul J.
1979 Some observations on the typology of focus and aspect in narrative
language. Studies in Language 3: 37–64.
Horn, Laurence R.
1985 Metalinguistic negation and pragmatic ambiguity. Language 61:
121–174.
218 Chris H. Reintges

Hyman, Larry M., and John R. Watters

1984 Auxiliary focus. Studies in African Linguistics 15: 233–273.
Jackendoff, Ray
1972 Semantic interpretation in generative grammar. Current studies in
linguistics 2. Cambridge, Mass.: MIT Press.
Jaggar, Philip J.
2001 Hausa. London Oriental and African language library. Amster-
dam/Philadelphia: John Benjamins.
in press The Hausa perfective tense-aspect used in wh-/focus constructions
and historical narratives: a unified account. To appear in West Afri-
can linguistics: descriptive, comparative and historical studies in
honor of Russell G. Schuh, Larry M. Hyman and Paul Newman
(eds.), Studies in African Linguistics, special issue.
Kuno, Susumo
1972 Functional sentence perspective: A case study from Japanese and
English. Linguistic Inquiry 3: 269–320.
Layton, Bentley
2000 A Coptic grammar with chrestomathy and glossary. Porta Lingua-
rum Orientalium N.S. 20. Wiesbaden: Harrassowitz.
Loprieno, Antonio
1995 Ancient Egyptian: a linguistic introduction. Cambridge: Cambridge
University Press.
McCloskey, James
1990 Resumptive pronouns, A'-Binding, and levels of representation in
Irish. In The syntax of Modern Celtic languages, Randall Hendrick
(ed.), 199–248. Syntax and Semantics 23. New York: Academic
Press.
2001 The morphosyntax of WH-extraction in Irish. Journal of Linguistics
37: 67–100.
McDaniel, Dana
1989 Partial and multiple WH-movement. Natural Language and Linguis-
tic Theory 7: 565–604.
Nunes, Jairo
2004 Linearization of chains and sideward movement. Cambridge, Mass.:
MIT Press.
Newman, Paul
2000 The Hausa language: an encyclopedic reference grammar. New
Haven and London: Yale University Press.
Pesetsky, David
1987 Wh-in-situ: movement and unselective binding. In The representa-
tion of (in)definiteness, Eric J. Reuland and Alice G.B. ter Meulen
(eds.), 98–129. Current studies in linguistics 14. Cambridge, Mass.:
MIT Press.
 Coptic relative tenses: The profile of a morpho-syntactic flagging device 219

1997 Optimality theory and syntax: movement and pronunciation. In Op-

timality Theory: an overview, Diana Archangeli and D. Terence
Langendoen (eds.), 134–170. Oxford: Blackwell.
1998 Some optimality principles of sentence pronunciation. In Is the best
good enough?, Pilar Barbosa, Danny Fox, Paul Hagstrom, Martha
McGinnis, and David Pesetsky (eds.), 337–383. Cambridge, Mass.:
MIT Press.
2000 Phrasal movement and its kin. Linguistic Inquiry monographs 37.
Cambridge, Massachusetts: MIT Press.
Pesetsky, David, and Esther Torrego
2000 T-to-C movement: causes and consequences. In Ken Hale: A life in
linguistics, Michael Kenstowicz (ed.), 355–426. Cambridge, Mass.:
MIT Press.
Polotsky, Hans Jakob
1944 Études de syntaxe copte. Publications de la société d’archeologie
copte. Cairo.
Rackowski, Andrea and Norvin Richards
2005 Phase edge and extraction: a Tagalog case study. Ms., MIT, Cam-
bridge, Mass..
Reinhart, Tanya
1984 Principle of Gestalt perception in the temporal organization of narrative
texts. Linguistics 22: 779–809.
Reintges, Chris H.
2001 Aspects of the morphosyntax of subjects and objects in Coptic Egyp-
tian. In Linguistics in the Netherlands 2001, Ton van der Wouden
and Hans Broekhuis (eds.), 177–188. Amsterdam/ Philadelphia: John
Benjamins.
2002 A configurational approach to Coptic Second Tenses. Lingua Aegyp-
tia 10: 343–388.
2003 Syntactic conditions on special inflection in Coptic interrogatives. In
Research in Afroasiatic grammar II. Selected papers from the fifth
conference on Afroasiatic languages, Paris 2000, Jacqueline Le-
carme (ed.), 363–408. Current Issues in Linguistic Theory 241. Am-
sterdam/Philadelphia: Benjamins.
2004 Coptic Egyptian (Sahidic dialect): A learner’s grammar.
Afrikawissenschaftliche Lehrbücher 15. Cologne: Rüdiger Köppe
Verlag .
Reintges, Chris H., and Melanie Green
2004 Coptic second tenses and Hausa relative aspects: a comparative
view. Lingua Aegyptia 12: 157–177.
220 Chris H. Reintges

Reintges, Chris H., Phil LeSourd, and Sandra Chung

2006 Movement, wh-agreement, and ‘apparent’ wh-in-situ. In press. To
appear in Wh-movement moving on, Lisa Lai-Shen Cheng and Nor-
bert Corver (eds.), Cambridge, Mass.: MIT Press.
Rizzi, Luigi
1996 Residual Verb Second and the wh-criterion. In Parameters and func-
tional heads: essays in comparative syntax, Adriana Belletti and
Luigi Rizzi (eds.), 63–90. Oxford/New York: Oxford University
Press.
1997 The fine structure of the left periphery. In Elements of Grammar:
Handbook in generative syntax, Liliane Haegeman (ed.), 281–337.
Dordrecht: Kluwer.
Rochemont, Michael S.
1986 Focus in generative grammar. Amsterdam/Philadelphia: Benjamins.
Rooth, Mats
1992 A theory of focus interpretation. Natural Language Semantics 1: 75–
116.
Tuller, Laurie
1986 Bijective relations in Universal Grammar and the syntax of Hausa.
Ph.D. dissertation. University of California at Los Angeles.
Watanabe, Akira
1996 Case absorption and wh-agreement. Dordrecht: Kluwer.
Vries, Mark de
2002 The syntax of relativization. Ph.D. Dissertation. University of Am-
sterdam. LOT Dissertation Series 53. Utrecht: LOT.
Zaenen, Annie
1983 On syntactic binding. Linguistic Inquiry 14: 469–504.
Part IV

The inventory of focus marking devices

Identificational operation as a focus strategy in
Byali

Brigitte Reineke

Abstract

This article investigates the focus system of Byali, a Gur language spoken in the
north-western part of Benin. It is argued that the essential feature of focus marking
in this language is an identificational operation which takes the form of an ex-situ
construction. The latter is analyzed as containing two interrelated predications,
namely a constituent that obligatorily precedes an identifying 'to be'-verb – as such
representing a nominal predication – and a verbal process or state established in
previous discourse. The grammatical operation of interrelating the two predications
results in a biclausal focus construction. The identifying 'to be'-verb, which has
been grammaticalized as a focus marker, but palpably still bears verbal features,
establishes a relation between the focused constituent and the predicative statement
representing the background. Whereas a combination of two propositions within a
single utterance is shown for the ex-situ constructions, no such overlap can be
established for the in-situ focus. As a result, I will argue that in-situ focus construc-
tions exhibit a higher degree of grammaticalization.

1. Introduction

This paper sheds light on the focus system of Byali and tries to show that
the essential feature of the focus strategy in this language is an identifica-
tional operation clearly retrievable from the ex-situ constructions. The latter
are envisioned to contain two interrelated predications, a constituent which
obligatorily precedes an identifying 'to be'-verb – as such representing a
nominal predication – and a pre-established verbal process or state. The
operation of interrelating the two predications results in a focus construc-
tion. As Caron (2000: 28) puts it: ”l’imbrication, la présentation simultanée,
dans un seul énoncé, d’un élément identifié (avec ou sans contraste) avec
une place non-instanciée d’une relation prédicative.” The identifying 'to
be'-verb that has been grammaticalized as a focus marker but palpably still
224 Brigitte Reineke

bears verbal features establishes the relation between the focused constitu-
ent and the predicative statement. In-situ focus constructions are considered
to be more grammaticalized as thought of before.
The Byali language, along with Ditammari, Nateni, Waama, and
MD'NKO', belongs to the eastern subgroup of the Oti-Volta-languages of
the Gur family. It is spoken in the Atakora mountains in the north-western
part of Benin. All the languages of this sub-group, except for Ditammari,
display a full-fledged noun class system characterized essentially by suf-
fixation of the class marker to the noun stem. Ditammari has prefixes in
addition to suffixes, which distinguishes it from most Gur languages which
are mainly characterized by suffixes. The languages have a fully intact
agreement system, as shown by the existence of anaphoric pronouns and of
nominal class markers linked to the modifiers of the noun which indicate
their syntactic dependencies. In addition, the verbal system exhibits a
grammaticalized aspectual system characterized by the binary opposition of
perfective and imperfective aspect. Also, these languages display a verbal
derivation having semantic as well as syntactic functions. On the prosodic
level, tone plays an important part in the grammar and the lexicon. The
basic word order in a canonical sentence is SVO; only in case of pronomi-
nal objects, the order is SOV.
This paper is structured as follows: part 2. illustrates the essential char-
acteristics of the Byali focus system. As concerns the concrete focus phe-
nomena, the distinction between non-verbal focus (3.) and verbal focus (4.)
is made, and the asymmetry between non-subject focus (3.1) and subject
focus (3.2) inherent in non-verbal focus is paid special attention to. In (5.),
an interpretation of the presented facts is given. The paper ends with a
summary in (6.).

2. Overview on the essential characteristics of the focus system

To express focus grammatically, the language employs morphological,

syntactic, as well as phonological means. At the level of morphology,
grammatically marked focus is realized by a focus marker immediately
following the constituent in focus. This holds true for the labelling of term
focus in the same way as for verbal focus. Concerning the polarity, a single
specific focus marker exists for affirmative ((l)G) 1 and negative focus
(PYC), respectively. Different syntactic focus constructions manifest them-
selves in the ex-situ and in the in-situ structures. In the ex-situ construc-
 Identificational operation as a focus strategy in Byali 225

tions, the focused constituent is located in sentence initial position. Al-

though it shares such a position with the topic constituent, the focused term
is integrated into the structure of the predicate and fulfills a grammatical
function whereas the topic constituent always stands outside the predication
and, from a semantic and structural perspective, can be seen to be optional. 2
In ex-situ constructions, the focus constituent represents the in-focus part of
the sentence that is followed by the out-of-focus part after a break. In addi-
tion to this multiple marking of the focused term – i.e. syntactically by the
sentence-initial position, morphologically by the focus marker, and (in case
of the ex-situ constructions) phonologically by phrasing – the non-focal
part of the focus construction exhibits relative structure, an isomorphism
frequently encountered across languages. The out-of-focus part is marked
as pragmatically dependent in two respects: Besides the relative structure,
there is a pragmatically conditioned verbal morphology that is restricted to
the perfective aspect. The examples show that neither in ex-situ nor in in-
situ focus constructions is the perfective verb joined by the suffix
³U‹YJKEJJCUan assertive meaning, cf. (1) and (2) vs. (3). 3 The latter
example (3) is unmarked with respect to focus. Thus, the perfective verb
takes the suffix ³U‹

Pragmatically marked focus constructions, the suffix ³U‹ is absent:

(1) D¸"I‹P7PF‹" DCPCPC G

child buy.PFV banana FM
‘The child bought BANANAS.’

as an answer to a constituent question:

(2) D¸"I‹P7PF‹" DCCT‹?

child buy.PFV what
‘What did the child buy?’

No marking of focus, the suffix ³U‹ appears:

(3) D¸"I‹ P7PF‹"³ U‹ DCPCPC

child buy.PFV ASS banana
‘The child bought bananas.’
226 Brigitte Reineke

Concerning the functional differentiation between the two syntactic focus

constructions that has been observed in the literature on many languages of
the world, 4 as far as the current research goes it does not seem to apply to
Byali. It is often claimed that the ex-situ constructions represent the so-
called identificational focus where an element is chosen out of the possible
alternatives in the context. This constituent would be marked as the one
which the predication would hold for to the exclusion of all other alterna-
tives. At the same time, the focused element would stand in contrast to the
other options so that the features of exhaustivity and contrast would be
highlighted as typical for the ex-situ constructions. The in-situ type as in-
formation structure refers to the part of predication which represents the
new information, i.e. the focused constituents fill a gap within the prag-
matic information of the listener. While in Byali new information focus is
realized primarily in-situ and identificational focus is very frequently asso-
ciated with exhaustivity and contrast, neither of the two types of structures
seem to be bound unequivocally to one of the semantic-pragmatic mean-
ings. According to the judgment of the informants, both syntactic focus
constructions allow for either of the interpretations.

3. Non-verbal focus

The differentiation between non-subject and subject focus applies to non-

verbal focus, whether the focused term is the argument of a verb or an ad-
verbial phrase not governed by the valency of the verb. Such differentiation
appears to be necessary insofar as this asymmetry known from many other
languages is typical for Byali as well even though from a structural point of
view there is no difference regarding non-subject focus in sentence initial
position. While non-subject focus can be realized either ex-situ or in-situ,
subject focus obligatorily is expressed by an ex-situ construction.
For the better understanding of the following data illustrating both ex-
situ as well as in-situ constructions, some preliminary remarks shall be
given before an interpretation is carried out in section 5.
1. The relative structure that marks the non-focused part of the ex-situ
construction and which represents the background is characterized by a
high toned nasal placed in front of the verb, as well as by the class
marker of the reference noun following the verb. In (5), the relative na-
sal precedes the verb [CtOD‹"‘see.PFV’ that is followed by the class
marker of the noun in focus, D¸³¸ I‹‘child’, here in the allomorphic
 Identificational operation as a focus strategy in Byali 227

variant M‹ From this example, it is evident that the syntactic postverbal
base-position of the argument marked for focus by the focus marker,
here the object position, is empty, i.e. the position in the presupposed
predication is not filled with an anaphoric pronoun.
2. The interrogative clauses show that in Byali, in contrast to many other
languages, the question words are never followed by the focus marker.
Contrary to answers, where the focus marker obligatorily marks the fo-
cused constituents in in-situ as well as in ex-situ constructions. This
fact indicates that in Byali question words are inherently focus-marked.
3. In sentences where a focus marker follows the object in in-situ con-
structions (cf., for example, (7) and (11)), narrow focus on the object as
well as focus on the verb together with its succeeding object can occur,
so that in such a case the scope of the focus stretches over the whole of
the verbal phrase. Sentences where the object is focus-marked in-situ
are therefore ambiguous because they may as well represent a reply to
the question “What did she do?”. Their meaning can only be deter-
mined by the context.
4. The combination of negated verb forms with marked focus that is ac-
ceptable in many languages does not apply to Byali.

3.1. Non-subject focus

For the presentation of the relevant data, question/answer-pairs are used as

a reliable method for elicitation. Questions point towards the interest of the
interrogator for information. The answers focus on the entities representing
the desired information. In examples (4) to (11) where the object is fo-
cused, the WH-question together with the corresponding answer are shown
in both construction types, namely ex-situ and in-situ, respectively.

3.1.1. Object

ex-situ
(4) Q: YG"GV‹" Coffi P" [CtOD W"?
who Coffi REL see.PFV CL.REL
‘WHO is it that Coffi has seen?’

(5) A: D¸¸³ I³ G W P" [CtOD‹" M‹

child CL FM CL.SUBJ. REL see.PFV CL.REL
‘It is the CHILD that (s)he has seen.’
228 Brigitte Reineke

in-situ
(6) Q: Coffi [CtOD‹" YG"GV‹"?
Coffi see.PFV who
‘WHOM has Coffi seen?’

(7) A: W [CtOD‹" D¸¸³ I³ G

  CL.SUBJ see.PFV child CL FM
‘He has seen the CHILD.’ (DK"K"³I‹‘child’)

ex-situ
(8) Q: DCCT‹" W P" P7PF‹" ¹?
what CL. SUBJ REL buy.PFV CL.REL
‘WHAT is it that (s)he has bought?’

(9) A: DCPCPC G W P" P7PF‹" JCp

banana FM CL.SUBJ REL buy.PFV CL.REL
‘It is BANANAS that (s)he has bought.’

in-situ
(10) Q: W P7PF‹" DCCT‹?
CL.SUBJ buy.PFV what
‘WHAT has (s)he bought?’

(11) A: W P7PF‹" DCPCPC G

CL.SUBJ buy.PFV banana FM
‘(S)he has bought BANANAS.’

Negation of an object focused in-situ is realized by a combination of the

preverbal negation morpheme R‹ and a focus marker following the object. 5

(12) W R‹ P7PF‹" DCPCPC PYC

 CL.SUBJ NEG buy.PFV banana FM.NEG
 (S)he has not bought BANANAS.’

3.1.2. Adjuncts

When adjuncts are focused, the same regularities apply as in the case of
focused objects, i.e. they can occur in ex-situ as well as in in-situ construc-
tions and they are always followed by the focus marker. In ex-situ construc-
 Identificational operation as a focus strategy in Byali 229

tions, the out-of-focus part of the utterance also exhibits the features of a
relative clause. The class marker that is referring to adjuncts and which is
postponed to the verb as a discontinuous part of the relative construction is
always -m in relation to the m-noun class.
In (13) to (16), examples with a local adjunct are given:

ex-situ
(13) Q: OC"PV‹"" W P" [CtOD‹" O" UC"D³ ¹?
where CL.SUBJ REL see.PFV CL.REL book CL
‘WHERE is it that (s)he has seen the book?’

(14) A: MCT³ º [CtJ‹ G W P" [CtOD‹""

 market CL inside FM CL.SUBJ REL see.PFV
O" UC"D³¹ 
CL.REL book CL
‘It is ON THE MARKET that (s)he has seen the book.’

in-situ
(15) Q: W [CtOD‹"" UC"D³ ¹ OC"PV‹?
CL.SUBJ see.PFV book CL where
‘WHERE has (s)he seen the book?’

(16) A: W [CtOD‹"" UC"D³ ¹ MCT³ º[CtJ‹G

CL.SUBJ see.PFV book CL market CL inside FM
‘(S)he has seen the book ON THE MARKET.’

3.2. Subject focus

As already mentioned, the focusing of subjects is restricted to the ex-situ

strategy. In Byali, the asymmetry between non-subject focus and subject
focus is expressed solely by this restriction of subject focus to ex-situ con-
structions. A hint at the ex-situ realization of subject focus is constituted by
the fact that the predication is categorically taking a relative-like structure,
as is exactly the case with ex-situ focused non-subjects, cf. the sentences
given in (17) to (20). Therefore, the subject constituent marked for focus is
directly followed by the relativizing nasal. Also in this context, attention
has to be paid to the fact that because of focus being inherent to it the ques-
tion pronoun asking for the subject is not followed by any focus marker.
230 Brigitte Reineke

(17) Q: YG"GV‹"P" U+¹P [CtOD W" D¸"³ I‹?

who REL yesterday see.PFV CL.REL child CL
‘WHO is it that saw the child yesterday?’

(18) A: Coff" G P" U+¹P M‹ [CtOD W"

Coffi FM REL yesterday CL.OBJ. see.PFV CL.REL
‘It is COFFI that saw it yesterday.’

(19) Q: DCCT‹" P" FGG ¹?

what REL fall.PFV CL.REL
‘WHAT has fallen?’

(20) A: UC"D³  G P" FGG  ¹

book CL FM REL fall.PFV CL.REL
‘It is a BOOK that has fallen.’

Structurally, subject focus constructions can also serve to express sentence

focus and are therefore ambiguous. This isomorphism holds valid for many
Gur languages. Apparently, the non-existence of a topic in case of sentence
focus, which constitutes a thetic construction, determines subject focus
structure. 6
The following sentence containing a focus-marked subject is a reply to
the question “What happened?”.

(21) P UCP³JWt  NG P" UYCP‹" JWp

1POSS car CL FM REL break.PFV CL.REL
‘MY CAR BROKE DOWN.’

4. Verbal focus

To mark focus in the verbal domain grammatically, in Byali, the same fo-
cus markers are used as with term focus 7 , i.e. (l)G for affirmative and PYC
for negative focus. The focus marker immediately follows the verb.

(22) W UQWO PG

CL.SUBJ smoke.IPFV FM
‘(S)he SMOKES.’
 Identificational operation as a focus strategy in Byali 231

(23) W UºPF‹" G OCPU¸P¹

 CL.SUBJ turn on FM Computer
‘(S)he HAS TURNED ON the computer.’

If the verb in focus is negated, the combination of the preverbal negation

morpheme R‹and the negative post-verbal focus marker turns out to be
obligatory:

(24) O R‹ UºPF‹" PYC OC"PUKP¹

1SG NEG turn on.PFV FM.NEG Computer
‘I haven’t TURNED ON the computer.’ 8

With the focused sentence-initial position being restricted to nominal con-

stituents, the essential feature of verb focus is that it exclusively occurs in
in-situ constructions. In some cases, the Byali language makes use of a
nominal periphrasis strategy for marking focus on the verbal phrase by
nominalizing the verb that forms an associative construction together with
its original object (Note that verbs are nominalized by integrating them into
the m-class). Thus, the nominalization of the verb-object phrase is a pre-
condition for it to be marked for focus and at the same time for it to occur
in ex-situ constructions. However, the nominalizing strategy for marking
predicate focus only occurs when the verbal phrase is represented by a pro-
gressive construction consisting of the dummy verb D‹"³RW"¢¸‘to hold’ + a
nominalized verb.

ex-situ
(25) EQ³ W UCJCp³ P"PG W O" RW"¢‹" O"
 porridge CL preparation CL FM CL.SUBJ REL hold.IPFV CLREL
‘It is the PREPARING OF PORRIDGE that she does.’

in-situ
(26) W RW"¢¸ EQ³ W UCJCp³ P" PG
CL.SUBJ hold.IPFV porridge CL preparation CL FM
‘She IS PREPARING PORRIDGE.’

It would lead to ungrammaticality if the nominalized verb is copied and

used as a finite verb instead of RW"¢¸, cf. (27):

(27) *EQWUCJCp³PPGWO"UCJCpO"
232 Brigitte Reineke

5. Interpretation

In order to understand how focus phenomena in Byali can be interpreted,

attention shall be centered on the ex-situ constructions. The following data
will show why the identificational strategy is considered decisive for mark-
ing focus in the Byali language.
The ex-situ focus constructions contain two interrelated predications, a
constituent followed obligatorily by a 'to be'-verb with identifying meaning
and placed at the left periphery, and a verbal expression referring to a pre-
supposed process or state. The sentence-initial term is identified as a refer-
ent and at the same time as the focused term by the 'to be'- verb functioning
as focus marker; it is brought into an equational relation with the verbal
predication in which the syntactic base position of an argument is empty.
The operation of interrelating these two predications results in a focus con-
struction with a biclausal character. The in-focus part of the overall struc-
ture has to be evaluated as nominal predication. This is especially indicated
by the possible connection of the 'to be'-verb with tense markers, cf. section
5.1. In verbal predication, it is characteristic of the out-of-focus part of an
utterance to be realized by means of relative-like structures, an operation
observed very frequently across languages.

5.1. Encoding of identity statements

The encoding of identity statements in Byali makes it quite obvious that the
focus marker lè, in specific cases reduced to è, and its negative form PYC
have to be traced back to a verb 'to be', or 'not to be' respectively, which has
an identifying meaning and that is used in both types of identifying struc-
tures, first, the presentational type, and second, the identifying nominal
predication.

5.1.1. The presentational type

The presentational construction makes the identity of a referent known to

the hearer, i.e. “presentation statements provide an object or a class of ob-
jects with a 'name', which may from now on be used by the hearer in refer-
ring to it” (Stassen 1997). In following Lambrecht’s (2001) approach,
Tröbs (2002) emphasizes that one-argument identificational clauses exhibit
a marked information structure where the scope of assertion stretches over
 Identificational operation as a focus strategy in Byali 233

the entire proposition (in the sense of Sasse’s thetic utterances), i.e. the
construction is to a large extent discourse-pragmatically motivated.
In Byali, this one-argument identificational statement consists of a
nominal or pronominal constituent + the 'to be'-verb; the form lè occurring
after pronouns is very often reduced to è in connection with a noun. In case
of negation, the affirmative (l)è is substituted by PYC The presupposition
of such identificational statements is not explicitly expressed, only situa-
tionally implied. In this way, the following statements are understood.

(28) CNG ‘It is you.’

 WNG ‘It is him/her.’
 V‹NG ‘It is us.’
D‹NG ‘It is them.’
 UC"D³C"G ‘It is the books.’
CNG  ‘It is them.’ (= the books < UC"D³C")
 D¸¸³IG(< D¸¸³I‹G) ‘It is the child.’
 M‹NG ‘It is it.’ (= the child < D¸¸³I‹)

Negation:
(29) PPYC ‘It is not me.’
 WPYC ‘It is not him/her.’
DG"IC"PYC ‘It is not the chief.’
 D¸¸³U¸PYC ‘It is not the children.’

5.1.2. The identifying nominal predication

Identifying nominal predications can be classified as equational; two enti-

ties refer to one and the same object. According to Stassen, both noun
phrases have the same semantic status and both can function as subject or
predicate, but their position evidently depends on discourse conditions. So,
the following sentences differ pragmatically:

(30) CQHHK G U[GN³ ¹ O[CV³ ¹

Coffi be.ID village CL teacher CL
‘Coffi is the teacher of the village.’

(31) U[GN³ ¹ O[CV³  G CQHHK

 village CL teacher CL be.ID Coffi
‘The teacher of the village is Coffi.’
234 Brigitte Reineke

The fact that in both types of identity statements the affirmative (l)è and the
negative PYC can be linked to the subject personal pronoun and can be
preceded by tense markers hints at the verbal features of these morphemes.
The emphatic pronominal forms cannot occur as subjects neither in one-
argument identificational clauses nor in identifying nominal predications.

with past marker ([¹):

(32) a. C [¹ NG
2SG PAST be.ID
‘It was you.’
*CYG[¹NG(CYG= 2.P.SG. emphat.)
(cf.CYGC[¹NG‘You, it was you.’,
where CYGrepresents the topic of the sentence.)
b. W[¹PYC  ‘It was not him.’
c. CQHHK[¹NGU[GN¹O[CV¹  ‘Coffi was the teacher of the
village.’
d. CQHHK[¹PYCU[GN¹O[CV¹  ‘Coffi was not the teacher of
the village.’

Following subject pronouns or tense markers, the phonetically non-reduced

form of the affirmative identifying 'to be'-verb is obligatorily used, as (28),
(32a,b), and (33a) show.

with future marker ([¸): With verbs of stative meaning, the future tense
marker [¸is always followed by a nasal.

(33) a. W[¸P"FG(< NG) ‘It will be him.’

b. W[¸P"FGU[G"N¹O[CV¹ ‘He will be the teacher of the
village.’
c. CQHHK["P"PYCU[G"N¹O[CV¹ ‘Coffi won’t be the teacher of
the village.’

In this connection, notice that in predicational relations characterized by a

subject and predicate structure, i.e. in predications in which the subject is
not identical to the entity specified in the predicate nominal, the classifying
copula PYCO occurs. It is also linked to personal and anaphoric subject
pronouns, not to emphatic ones, and can co-occur with tense markers. In
these sentences, the predicate nominal is marked for focus.
 Identificational operation as a focus strategy in Byali 235

(34) a. WPYCOO[CVG ‘He is a TEACHER.’

b. W[¹PYCOO[CVG ‘He was a TEACHER.’
c. ¹PYCOb[C"N‹DG(< b[C"N‹D‹G)‘You are BYALEBE,
 ¹R‹PYCOPCV[GOD‹PYC you are not NATEMBA.’

5.2. Focus constructions

The same phenomenon, i.e. the possible combination of (l)è with tense
markers, occurs when (l)è functions as focus marker, as the following ex-
ample (35) containing the past morpheme [¹shows:

(35) D¸¸³ I‹ [¹ NG W P" [¹ [CtOD‹" M‹
 child CL PAST FM CL.SUBJ REL PAST see.PFV CL.REL
‘It was the CHILD that (s)he had seen.’

However, the variant without the tense marker in the in-focus part is the
preferred one; the temporal reference of the whole predication is designated
by the obligatory occurrence of the tense marker before the verb:

(36) D¸³¸  I(‹)³G W P" [¹ [CtOD‹" M‹

 EJKNFCL FM CL.SUBJ REL PAST see.PFV CL.REL
‘It was the CHILD that (s)he had seen.’

It is interesting to note that the combination of the focus marker with tense
morphemes is not restricted to the fronted constituent; its occurrence is also
possible, even if infrequent and unusual, with constituents focused in-situ:

(37) W [¹ [CtOD‹" D¸¸³ I‹ [¹ NG

CL.SUBJ PAST see.PFV child CL PAST FM
‘(S)he had seen the CHILD.’
or: W[¹[CtOD‹"D¸¸³I(‹)³G

This applies also to instances of verbal focus:

(38) W [¹ F¹ [¹ NG EQ³ W

CL.SUBJ PAST eat.PFV PAST FM porridge CL
‘(S)he had EATEN the porridge.’
or: W[¹F¹¹GEQW
236 Brigitte Reineke

The possible occurrence of the focus marker with tense morphemes in ex-
situ as well as in in-situ position of the focused constituents indicates very
clearly the original verbal feature of NGand supports the interpretation of
the sequence N - (tense marker) - focus marker as an original nominal
predication. Another observation supporting the hypothesis of the verbal
origin of the focus marker is the use of personal pronouns and anaphoric
pronominal forms in subject function when they are focused. The use of
their free or emphatic variants respectively is ungrammatical:

(39) C NG P" F W EQ³ W

2SG FM REL eat.PFV CL.REL porridge CL
‘It was YOU that has eaten the porridge.’
*CYG(l)GP"F¹¹WEQW(CYG= 2. SG. emphat.)

It should be remembered that emphatic pronouns are also illicit in one-

argument identificational clauses and in identifying nominal predications.
The examples in (37), (38), and (39) show that the focus marker always
appears in its non-reduced form NG when immediately following tense
markers and subject pronouns.

5.3. In-situ constructions: More grammaticalized?

What about in-situ constructions? How could they be interpreted? Whereas

the ex-situ constructions are characterized by an imbrication of two propo-
sitions within a single utterance, such an imbrication cannot be established
for in-situ focus. One might be able to consider sentences in which objects
are focus marked in-situ (40), or in which adjuncts take a sentence-final
position (41), to consist of two subsequent predications as well:

(40) W P7PF‹" DCPCPC G

 CL.SUBJ buy.PFV banana FM
‘(S)he has bought BANANAS.’
= ?? (S)he has bought // it is bananas

(41) Coffi [CtOD‹" D¸¸³ I‹ MCT³ º [CtJ‹ G

 Coffi see.PFV child CL market CL inside FM
‘Coffi has seen the child ON THE MARKET.’
= ?? Coffi has seen the child // it is on the market.
 Identificational operation as a focus strategy in Byali 237

Such an interpretation would be somewhat more difficult for sentences with

focused verbs taking one or two arguments; there, the verb is separated
from its argument(s) by (l)è, cf. the following sentence:

(42) Coffi F¹¹ G EQ¸³ W

Coffi eat.PFV FM porridge CL
‘Coffi has EATEN the porridge.’
 = ?? Coffi has eaten // it is // the porridge.

A similar problem arises with adjuncts that can also take the position be-
tween the verb and its argument in addition to sentence-final position, so
that apart from SVOA the sequence SVAO is possible as well. Cf. (43) vs.
(41), where the adjunct MCT¹[CtJ‹G ‘(it is) on the market’ occurred in
sentence-final position: 9

(43) Coffi [CtOD‹" MCT³ ¹ [CtJ‹ G D¸¸³ I‹

Coffi see.PFV market CL inside FM child CL
‘Coffi has seen the child ON THE MARKET.’
= ?? Coffi has seen // it is on the market // the child.

In none of the cases can in-situ constructions be evaluated as an imbrication

of two propositions within a single utterance. In general, one might inter-
pret the current status of these constructions as more grammaticalized.

6. Summary

First, in tracing back the (affirmative) focus marker (l)è to a verb 'to be'
with identifying meaning, the identificational strategy is considered deci-
sive for explaining ex-situ focus constructions in Byali. The identifying 'to
be'-verb exhibits by itself discourse-pragmatical features; the constructions
in which it occurs are to a large extent pragmatically motivated.
Second, the ex-situ focus construction contains two interrelated predica-
tions, a presupposed verbal process (or state) representing the background
and a constituent followed by the 'to be'-verb with identifying meaning and
placed at the left periphery for which the verbal predication holds true. The
operation of interrelating these two predications results in a focus construc-
tion with biclausal character. The identifying 'to be'-verb functions in this
238 Brigitte Reineke

construction as focus marker while still possessing verbal features, as evi-

denced by its co-occurrence with subject pronouns and tense markers.
The out-of-focus part of ex-situ constructions is marked as pragmati-
cally dependent in two ways: First, the non-focal part exhibits relative
structure; second, there is a clearly pragmatically conditioned verbal mor-
phology, restricted to the perfective aspect.
Third, whereas an imbrication of two propositions within a single utter-
ance could be shown for the ex-situ constructions, no such overlap can be
established for the in-situ focus. The in-situ constructions are considered as
more grammaticalized.

Abbreviations
ASS assertive suffix
CL class marker
DEP dependent marker
EMPH emphatic
FM focus marker
IPFV imperfective
NEG negation particle
OBJ object
PFV perfective
POSS possessive pronoun
REL relative marker
SG singular
SUBJ subject

Notes

1. Usually the phonologically reduced variant Gis employed. If the focused

entity ends with a nasal, the focus marker (l)G) is realized with its allo-
morph variant nè.
2. If both categories, topic and focus, appear in a single sentence, the se-
quence topic – focus corresponds to the universal fact that, contrary to fo-
cus, topic is external to the sentence.
3. In the same way, the non-occurrence of the assertive verbal suffix in rela-
tive clauses, temporal and conditional clauses, sequential clauses as well
as in negative sentences is due to their pragmatics, cf. Frajzyngier (2004)
 Identificational operation as a focus strategy in Byali 239

who emphasizes the pragmatical dependency of these clauses in languages

of the Niger-Congo as well as the Chadic families.
4. Cf. e.g. E. Kiss (1998).
5. In ex-situ focus constructions, the negative focus marker follows the ne-
gated entity placed in sentence initial position as well.
6. Cf. Sasse 1987, 1995.
7. This fact conflicts with the situation in the Ditammari language belonging
to the same subgroup. In Ditammari, nominal constituents and verbs are
focused by different markers.
8. If it is not the negated verb that stands in the scope of focus but the entire
utterance meaning something like “It’s true that they didn’t steal it.”, then
the negation morpheme R‹ and the affirmative focus marker (l)Gcan co-
occur in a sentence. A combination of [R‹+ Verb] [(l)G] in a sentence
represents the statement of the truth value of a negated proposition
whereas the combination [R‹+ Verb] [PYC] represents the negation of a
focused verb or term, e.g. OR‹N‹UºPF‹"G‘(It is true), I HAVEN’T
TURNED IT ON.’ cf. Reineke (2006).
9. A pragmatic difference between the two positional variants was negated
by the informant. Still, further research will have to resolve whether or not
the immediate post-verbal position is related to stronger emphasis.

References

Bearth, Thomas
1993 Satztyp und Situation in einigen Sprachen Westafrikas. In Beiträge
zur afrikanischen Sprach- und Literaturwissenschaft, Wilhelm J.G.
Möhlig (ed.), 91–104. Köln: Rüdiger Köppe.
Caron, Bernard
2000 Assertion et préconstruit: topicalisation et focalisation dans les
langues africaines. In Topilisation et focalisation dans les langues
africaines, Bernard Caron (ed.), 7–42. Louvain-Paris: Peeters.
É. Kiss, Katalin
1998 Identificational focus versus information focus. Language 74 (2):
245–273.
Frajzyngier, Zygmunt
2004 Tense and aspect as coding means for information structure: a poten-
tial areal feature. Journal of West African Languages 30: 53–67.
Hopper, Paul
1979 Some observations on the typology of focus and aspect in narrative
language. Studies in Language 3: 37–64.
240 Brigitte Reineke

Reineke, Brigitte
2006 Verb- und Prädikationsfokus im Ditammari und Byali (to appear). In
“Mama Miehe”. Festschrift für Gudrun Miehe zum 65. Geburtstag,
Kerstin Winkelmann and Dymitr Ibriszimow (eds.), 163–180. Köln:
Rüdiger Köppe.
Sasse, Hans-Jürgen
1987 The thetic/categorical distinction revisited. Linguistics 25 (3): 511–
580.
1995 ‘Theticity’ and VS order: A case study. In Verb-subject order and
theticity in European languages (Sprachtypologie und Universalien-
forschung, 48: 1/2), Yaron Matras and Hans-Jürgen Sasse (eds.), 3–
31. Berlin: Akademie-Verlag.
Stassen, Leon
1997 Intransitive predication: Oxford studies in typology and linguistic
theory. Oxford: Clarendon Press.
Tröbs, Holger
2002 Some notes on one-argument-identificational clauses in Manding
(Western Mande, Niger-Congo). Afrika und Übersee, 85: 129–144.
Exhaustivity marking in Hausa: A reanalysis of the
particle nee/cee

Katharina Hartmann and Malte Zimmermann

Abstract

The article presents a reanalysis of the Hausa particle nee/cee, which is frequently
analyzed as a focus marker (Tuller 1986, Green 1997). The syntactic distribution of
nee/cee is shown to be not typical of genuine focus markers: nee/cee is optional. It
is not restricted to a particular position as it can occur with ex-situ and in-situ focus
(with different frequencies). And, nee/cee can associate with focus at a distance.
Taking up observations by Green (1997), it is shown that nee/cee is not optional
from a semantic point of view, as its presence triggers an exhaustive interpretation
of focus. Nee/cee is therefore analyzed as a focus-sensitive marker introducing a
conventional implicature, which is responsible for the exhaustive interpretation.
This reanalysis of nee/cee also accounts for its infelicity in non-exhaustive con-
texts.

1. Introduction

Hausa focus constituents are often accompanied by the particle nee or its
feminine counterpart cee. The literature usually describes this particle as a
focus marker. Green (1997:29) mentions that the particle leads to an exaus-
tive interpretation of the focus but does not further develop this idea. In this
article, we take up Green’s observation and show that nee/cee always indi-
cates exhaustivity. Our reanalysis of the particle is based on several obser-
vations showing that nee/cee does not share most of the typical properties
of grammatical focus markers. It is similar to focus markers in being focus-
related. It differs from them in at least three respects: First, it is optional,
even if focus is not marked by other strategies. Second, it can associate
with the focus at a distance, an untypical property of focus markers. Third,
if present, the particle nee/cee has a semantic impact in form of a conven-
tional implicature: it causes an exhaustive interpretation of the focus. It is
therefore excluded in non-exhaustive environments such as mention-some
242 Katharina Hartmann and Malte Zimmermann

contexts, or in contexts where a property is known to hold of more than the

focused entity.

2. Focus in Hausa

Hausa 1 is a tone language with three lexical tones: a high tone, which is not
marked in the examples, a low tone (`), and a falling tone (^). Its basic word
order is SVO. The uninflected verb is preceded by a separate morpheme
that encodes temporal, aspectual, and agreement specifications (the auxil-
iary). In the perfective and continuous aspects, the auxiliary has two differ-
ent morphological forms, depending on whether some constituent of the
sentence is fronted (Tuller 1986). We follow the traditional terminology
and call the auxiliary in clauses without fronting the absolute auxiliary. In
clauses with fronting, we refer to it as the relative auxiliary.
Hausa has two strategies to express focus. A focus constituent can either
be fronted (ex-situ focus) or it can remain in its base-position (in-situ fo-
cus). Ex-situ focus (cf. Tuller 1986, Green 1997, Newman 2000, Jaggar
2001, Hartmann and Zimmermann 2007) is repeatedly marked: It is syntac-
tically marked through fronting. In addition, it is morphologically marked
since syntactic fronting triggers the relative auxiliary. The fronted focus
constituent can be followed by the particle nee (or its feminine form cee).
Finally, focus fronting is prosodically marked by an intonational phrase
boundary between the ex-situ constituent and the rest of the clause (cf. Le-
ben, Inkelas and Cobler 1989). An example of ex-situ focus is given in (1),
with the focus printed in bold face. 2 Here, as in most other examples, focus
is pragmatically controlled for by means of question-answer pairs.

(1) Q: Mèe sukà kaamàa ?

what 3PL.REL.PERF catch
‘What did they catch?’
A: Kiifii (nèe) sukà kaamàa.
fish PRT 3PL.REL.PERF catch
‘They caught FISH.’

In-situ focus, on the other hand, is not marked morpho-syntactically (Jaggar

2001, 2004, Green and Jaggar 2003, Hartmann and Zimmermann 2007).
Due to the absence of focus movement, the auxiliary appears in its absolute
form. In-situ focus is also not marked prosodically. Furthermore, it is only
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 243

rarely accompanied by the particle nee, which is often shortened to ne in

sentence-final position. (2) gives a typical question-answer pair. Since wh-
phrases are almost always fronted, the relative auxiliary occurs in the ques-
tion. In the answer, however, the focus (kiifii, ‘fish’) is not moved, hence
the auxiliary is absolute. There is no formal indication of the focus in the
answer at all, which can therefore only be determined pragmatically.

(2) Q: Mèe sukà kaamàa ?

what 3PL.REL.PERF catch
‘What did they catch?’
A: Sun kaamà kiifii.
3PL.ABS.PERF catch fish
‘They caught FISH.’

While ex-situ focus is restricted to maximal projections, in-situ focus is

possible for any constituent (heads and maximal projections alike), with the
exception of subjects. If a subject is focused, the auxiliary must be relative,
indicating ex-situ focus. This is illustrated in (3).

(3) Q: Wàa yakèe kirà-ntà ?

who 3SG.REL.CONT call-her
‘Who is calling her?’
A: [NP Dauda] (nèe) yakèe / *yanàa kirà-ntà.
D. PRT 3SG.REL.CONT/ 3SG.ABS.CONT call-her
‘DAUDA is calling her.’

Our analysis of the distribution and meaning of the particle nee/cee di-
verges from the opinion held in the recent literature on focus in Hausa
where it is proposed that nee/cee is a focus marker (Green 1997, 2004,
Newman 2000). Since nee/cee always appears together with focus, this
analysis seems to be plausible at first glance. There are three arguments
against this position, however. In a nutshell: Assuming that focus markers
are required to mark focus, the optionality of nee/cee is unexpected (section
3). Second, the frequency of the particle considerably differs between ex-
situ and in-situ focus: nee/cee freely associates with ex-situ focus, but is
quite restricted with in-situ focus. Given that, at least with question-answer-
focus, the in-situ strategy is the prominent strategy (for a quantitative study
of in-situ and ex-situ focus, cf. Hartmann and Zimmermann 2007), the rare
occurrence of nee/cee in these cases would be unaccounted for if it was
244 Katharina Hartmann and Malte Zimmermann

indeed a focus marker. Finally, the semantic behavior of nee/cee is untyp-

ical of focus markers: it always appears in exhaustive environments (sec-
tion 4). From these considerations, we conclude that nee/cee is not a focus
marker, but a focus-sensitive exhaustivity marker.

3. Syntactic properties of nee/cee

In this section, we further investigate the syntactic distribution of nee/cee.

We first discuss the particle after ex-situ focus and then turn to its more
restricted occurrence with in-situ focus.

3.1. Ex-situ focus

It is well-known from the literature on Hausa that the particle nee/cee is

focus-related, i.e. that it only appears if a constituent is focused (Parsons
1963, Schachter 1966, Tuller 1986, Green 1997, Newman 2000, Jaggar
2001, Green and Jaggar 2003, Hartmann and Zimmermann 2007). The
particle can occur after a fronted focus. The following examples illustrate
subject focus (4ab), object focus (4c), PP focus (4d), and (nominalized) VP
focus (4e) (examples (4ade) are from Newman 2000: 187ff.).

(4) a. [NP Dèelu] cèe takèe sôn àgoogo.

D. PRT 3SG.F.REL.CONT want watch
‘DEELU wants a watch.’
b. [NP Kànde da Hàwwa] nee sukà zoo.
K. and H. PRT 3PL.F.REL.PERF come
‘KANDE AND HAWWA came.’
c. [NP Àgoogo] nèe Dèelu takèe sô.
watch PRT D. 3SG.F.REL.CONT want
‘Deelu wants A WATCH.’
d. [PP Dà wuÝaa] nèe ya sòokee shì.
with knife PRT 3SG.REL.PERF stab him
‘He stabbed him WITH A KNIFE.’
e. [VP Biyà-n hàrÄaajì-n] nee Tankò ya yi.
paying-GEN taxes-DEF PRT T. 3SG.REL.PERF do
‘It was PAYING THE TAXES that Tanko did.’
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 245

The particle has a tonal peculiarity in that it always carries polar tone, i.e. a
tone opposite to the preceding tone (cf. Parsons 1963:166). We further as-
sume that the particle nee/cee is formally unspecified: it neither carries a
tense specification nor is it specified for agreement features, with the ex-
ception of gender. It is only with feminine singular noun phrases that gen-
der is specified and cee is used instead of nee (4a). In all other cases, e.g.
with masculine NPs (see (1), (3), and (4c)), plurals (including coordinated
feminine NPs (4b)), PPs (4d), and VPs (4e), nee must occur (cf. Parsons
1963).
Hausa scholars usually analyze nee/cee after ex-situ focus as an em-
phatic marker (Schachter 1966) or as a focus marker (cf. Tuller 1986,
Green 1997, Green 2004, Newman 2000, Jaggar 2001, Green and Jaggar
2003). Given the existence of focus markers in a large variety of other Af-
rican languages (see Bearth 1999 for an overview), this assumption is not
far-fetched. Green (1997) represents the most elaborate analysis of focus in
Hausa. In her account, nee/cee is the head of a focus phrase (FP). Provided
with focus features, the particle attracts the focus phrase to its specifier:

(5) [FP [NP BintàF] [F’ [F cee [S tsubj takèe biyà teelà]]]]
B. PRT 3SG.F.REL.CONT pay tailor
‘BINTA paid the tailor.’

All analyses acknowledge that the particle is optional. The examples in (4)
are equally grammatical in the absence of nee/cee. In other words, the pres-
ence of nee/cee is not obligatory for focus marking. It is still optional if
there is no word order variation, e.g. with focused subjects, which appear in
the same linear position as unfocused subjects. The only indication of sub-
ject focus in (6) is the relative auxiliary.

(6) Tankò (nee) ya biyà hàrÄaajì-n.

T. PRT 3SG.REL.PERF pay taxes-DEF
‘TANKO paid the taxes.’

Recall from section 2 that the absolute-relative distinction within the auxil-
iary paradigm is only attested in the perfective and continuous aspect. In
the future and habitual aspect, the auxiliary has the same form independent
of focus fronting. If nee/cee was a focus marker, one might expect it to be
obligatory when focus is not marked by other morpho-syntactic means
(word order, relative auxiliary), such as subject focus in the subjunctive,
246 Katharina Hartmann and Malte Zimmermann

future, and habitual aspect. However, nee/cee may be absent even then. (7)
illustrates subject focus in the future aspect. 3

(7) Q: Wàanee nèe 4 zâi tàfi Jamùs ?

Who PRT FUT.3SG go Germany
‘Who will go to Germany?’
A: Audù zâi tàfi Jamùs.
Audu FUT.3SG go Germany
‘AUDU will go to Germany.’

In light of these data, the hypothesis that the particle nee/cee is a focus
marker appears to be unwarranted. Its possible absence in sentences with
no other morpho-syntactic signs of focus marking suggests that the primary
function of the particle is not that of a focus marker. In section 4, we pre-
sent an alternative analysis showing that the presence of nee/cee adds a
conventional implicature which leads to an exhaustive interpretation of the
focus. We will argue that the particle is a focus-sensitive exhaustivity
marker, rather than a syntactic focus marker.

3.2. In-situ focus

In section 2, we pointed out that focus constituents do not have to be

fronted but may remain in their base position. In-situ focus is quite fre-
quent, it even represents the predominant focus-strategy for new-
information focus (cf. Hartmann and Zimmermann 2007). In addition, we
argued that in-situ focus need not be marked at all.
As concerns the particle nee/cee, it is optional with in-situ foci as well
although it seems to occur much less frequently with these. If nee/cee ap-
pears, it generally follows the in-situ focus (Jaggar 2001:497). There seem
to be two positions for nee/cee with in-situ focus. It can either appear in the
sentence-final position (8-A1), or “at the end of the core sentence but be-
fore adverbial adjuncts or complements”, cf. (8-A2) (Newman 2000:546).

(8) Q: Mèenee nèe Tankò ya sàyaa à kàasuwaa ?

what PRT T. 3SG.REL.PERF buy at market
‘What did Tanko buy at the market?’
A1: Tankò yaa sàyi [NP kàazaa][PP à kàasuwaa] nè.
T. 3SG.ABS.PERF buy chicken at market PRT
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 247

A2: Tankò yaa sàyi [NP kàazaa] nèe [PP à kàasuwaa]. 5

‘Tanko bought CHICKEN at the market.’

Notice that the particle does not have to be adjacent to the in-situ focus, see
(8-A1) where the particle follows the locative adverbial, which belongs to
the informational background.
In the following, some further examples are presented that provide more
evidence for the two particle positions with in-situ focus. First, when the
right edge of the focus extends to the right periphery of the clause, nee/cee
has to appear in clause-final position. This is shown for in-situ object focus
(9), locative focus (10), predicate focus (11), and sentence focus (12).

(9) Q: Mèenee nèe Audù ya sàyaa ?

What PRT A. 3SG.REL.PERF buy
‘What did Audu buy?’
A: Audù yaa sàyi [NP zoobèe] ne.
A. 3SG.ABS.PERF buy ring PRT
‘Audu bought a RING.’

(10) Q: (A) ìnaa nèe Tankò ya sàyi kiifíi ?

where PRT T. 3SG.REL.PERF buy fish
‘Where did Tanko buy fish?
A: Tankò yaa sàyi kiifii [PP à kàasuwaa] nè.
T. 3SG.ABS.PERF buy fish at market PRT
‘Tanko bought fish AT THE MARKET.’

(11) Q: Mèe Hàwwa ta yi ?

what H. 3SG.F.REL.PERF do
‘What did Hawwa do?
A: Hàwwa taa [VP yankà naamàa] ne.
H. 3SG.F.ABS.PERF cut meat PRT
‘Hawwa CUT THE MEAT.’

(12) Q: Mèe ya fàaru ?

what 3SG.REL.PERF happen
‘What happened?’
A: [IP Muusaa yaa yi minì màganàa] ne.
M. 3SG.ABS.PERF do me speech PRT
‘MUSA TALKED TO ME.’
248 Katharina Hartmann and Malte Zimmermann

The particle can also follow an in-situ focus in non-final position, cf. (8-
A2), as well as (13):

(13) Q: Mèenee nèe màkaanikèe ya gyaaràa à gaareÄ ejì ?

what PRT mechanic 3SG.REL.PERF repair at garage
‘What did the mechanic repair at the garage?’
A: Màkaanikè yaa gyaarà [NP mootàr] nee à g.
mechanic 3SG.ABS.PERF repair car PRT at g.
‘The mechanic repaired the CAR at the garage.’

In addition, nee/cee can associate with the focus at a distance, as already

observed in connection with (8-A1). A further example is given in (14).

(14) Q: Wàacee cèe ka ganii à makarÄantar ?

who.fem PRT 2SG.REL.PERF see at school
‘Whom did you see at school?’
A: Naa gaa [NP Dèelu] à makarÄantar nè.
1SG.ABS.PERF see D. at school PRT
‘I saw DELU at school.’

The fact that the particle does not have to follow the focus immediately
corroborates our conclusion from section 3.1 that nee/cee is not a typical
focus marker. Grammatical markers are usually adjacent to the constituent
they mark.
We propose that the position of the particle is not primarily determined
syntactically (as in Green 1997), but follows from prosodic requirements
instead: Nee/cee always occurs before a prosodic phrase boundary. 6 In
Hausa, there are obligatory phrase boundaries between an ex-situ focus
constituent and the rest of the clause, and between the direct object and
subsequent embedded clauses and/or adverbials (cf. Leben, Inkelas and
Cobler 1989). 7 As it happens, these are exactly the positions where nee/cee
appears. It goes without saying that the end of a sentence also demarcates a
prosodic boundary, hence, the occurrence of clause-final nee/cee is pre-
dicted here, too. That the particle is sensitive to its prosodic environment
receives further support from the fact that it is sensitive to another phono-
logical property of the preceding material, i.e. its tone. Recall that nee/cee
has polar tone, a tone with opposite pitch to the preceding one.
Note that there is no prosodic phrase boundary between the verb and the
object NP in transitive sentences. It is therefore not surprising that nee/cee
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 249

is banned from this position. This restriction holds even if the verb is nar-
rowly focused. Such cases are illustrated in (15) and (16). If the particle is
present, it must appear after the direct object. 8

(15) Q: Mèenee nèe màkaanikèe ya yi

what PRT mechanic 3SG.REL.PERF do
wà mootàr à gaarÄeejì ?
with car at garage
‘What did the mechanic do with the car at the garage?’
A: Màkaanikèe ya [V gyaarà] (*nee) [NP mootàr] (nee) à gaarÄeejì.
‘The mechanic REPAIRED the car at the garage.’

(16) Q: Mèe Tanko ya yi wà hàraajì-n ?

what T. 3SG.REL.PERF do to taxes-DEF
‘What did Tanko do with the taxes?’
A: Tanko yaa [V biya] (*nèe) hàraajìn (ne).
Tanko PAID the taxes.’

The examples in (15) and (16) suggest a close structural relationship be-
tween the verb and the object. It might seem unexpected that verb focus
does not lead to a restructuring of the prosodic structure, as it happens for
instance in some Bantu languages (Kanerva 1990). But recall from section
2 that in-situ focus is absolutely unmarked, even prosodically. Hence, in-
situ focus has no repercussion on the prosodic structure in Hausa, and the
tight connection between verb and object remains even under verb focus.
Nee/cee may also occur between the indirect and the direct object in
double object constructions, cf. (17). On a prosodic account, this is ex-
pected given that “there is typically a phrase boundary between the two
objects of double object constructions” (Inkelas & Leben 1990:19).

(17) Q: Wàacee cèe Ìbrahìm ya bai wà kud’ii ?

whom.fem PRT I. 3SG.REL.PERF give to money
‘To whom did Ibrahim give the money?’
A: Ìbrahìm yaa bai wà tsoohuwarsà nee kud’ii.
I. 3SG.ABS.PERF give to mother PRT money
‘Ibrahim gave the money to his MOTHER.’
250 Katharina Hartmann and Malte Zimmermann

We conclude this section with a further observation. A clause-final particle

is incompatible with ex-situ focus. Such examples are consistently judged
ungrammatical. This is shown for ex-situ subject and object focus:

(18) Q: Wàanee nèe ya zoo ?

who PRT 3SG.REL.PERF come
‘Who came?’
A:*Audù ya zoo nè.
A. 3SG.REL.PERF come PRT
‘AUDU came.’

(19) Q: Mèenee nèe Harúuna ya kaawoo dàgà Jamùs ?

what PRT H. 3sg.rel.perf bring from Germany
‘What did Haruna bring from Germany?
A:*ěeediyòo Hàruunà ya kaawoo (dàgà J.) ne.
radio H. 3SG.REL.PERF bring from G. PRT
‘Haruna brought a RADIO from Germany.’

Our language consultants unanimously agreed that (18A) and (19A) are
only grammatical as yes-no questions where the final particle functions as a
question tag. 9 A declarative reading of these sentences is not available. At
present, the source of this additional restriction is mysterious to us. The
data in (18) and (19) appear to fall neatly under the syntactic account pro-
posed by Green (1997) in (5): As nee/cee heads the FP, it must be right-
adjacent to the fronted focus constituent in Spec,FP. On the other hand,
Green’s analysis does not easily account for the sentence-internal occur-
rences of nee with in-situ focus in (13) and (15). We will leave this matter
open for further research.

To summarize, Hausa has a particle nee/cee which optionally appears after

the focus constituent, whether in-situ or ex-situ. The properties of the parti-
cle described in the present section lead us to assume that it does not be-
have like a typical focus marker. Typical focus markers, as employed in
many other African languages, are obligatory. They consistently mark the
focus in a sentence. The particle nee/cee, on the other hand, is optional
even if focus is not marked by any other means. Moreover, it can associate
with the focus at a distance. This property is typical of focus-sensitive par-
ticles but not of grammatical focus markers. We conclude that focus in
Hausa does not imply the presence of nee/cee. Rather, the reverse holds: if
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 251

nee/cee occurs, a focus must occur to its left. Since such a dependency on
focus is typical of focus-sensitive particles, we conclude that nee/cee is a
focus-sensitive particle, rather than a focus marker.

4. Nee/cee as a focus-sensitive exhaustivity marker

As we concluded in the last section, the distribution of nee/cee is not pri-

marily determined by structural factors. Instead, we will argue that its oc-
currence is motivated by semantic considerations only. More precisely, we
show that the presence of nee/cee introduces a conventional implicature
triggering an exhaustive focus interpretation.

4.1. Green’s (1997) observation

In her dissertation, Green (1997) observes a semantic distinction between

cases of focus fronting where nee/cee is present and cases where it is ab-
sent. Usually, a sentence containing a focus may be followed by an after-
clause that introduces an alternative to the focus constituent. This is shown
for Hausa in (20) (from Green 1997:29). The fronted focus à kân teebùr ‘on
the table’ is extended in the afterclause by another PP. Such an extension is
illicit if the fronted focus is followed by nee. If nee is present, the focus
receives an exhaustive interpretation: A focus constituent is interpreted
exhaustively if the property denoted by the backgrounded part of the clause
holds of the entity denoted by the focus constituent, and only of this entity.
With respect to (20), this means that the books are put on the table and
nowhere else.

(20) À kân teebùr (*nee) sukà sâ lìttàttàfai,

upon table PRT 3PL.REL.PERF put books
dà kuma cikin àkwàatì.
and also inside box
‘They put the books on the table, and also inside the box.’

Green (1997) accounts for this observation by treating nee/cee as an ex-

haustive focus marker. In the following sections, we provide new data that
corroborate Green’s claim that nee/cee adds exhaustivity to the semantic
interpretation. As pointed out above, though, we analyze nee/cee as a focus
252 Katharina Hartmann and Malte Zimmermann

sensitive exhaustivity marker, rather than as a proper focus marker. In what

follows, we therefore gloss the particle as EXH for exhaustivity marker.

4.2. *Nee/cee in non-exhaustive contexts

The data discussed in this section have in common that the focused entity is
not the only one satisfying the property denoted by the background, to the
effect that an exhaustive interpretation of the focus becomes impossible.
This is achieved by adding an afterclause in which the same backgrounded
property is predicated of an alternative value. In all such contexts, the parti-
cle nee/cee is illicit.
Notice first that we were able to reproduce the facts discussed in (20). If
the focus in the main clause is followed by nee/cee, extension by an also-
phrase is excluded. The presence of nee/cee excludes all focus alternatives
except for the focused entity itself. In (21A), nee forces the interpretation
that nobody else apart from Musa returned from Kano. Similarly in (22), no
additional individuals may be added to the denotation of the predicate satis-
fying the focused object, if this is followed by cee.

(21) Q: Wàa ya daawoo dàgà Kano ?

who 3SG.REL.PERF return from Kano
‘Who returned from Kano?’
A:#Musa nèe ya daawoo dàgà Kano
M. EXH 3SG.REL.PERF return from Kano
dà kuma Hàliimà cee ya daawoo dàgà Kano.
and also H. EXH 3SG.REL.PERF return from Kano
‘MUSA returned from K. and HALIMA, too, returned from K.’

(22) Hàwwa (#cèe) mukà ganii.

H. EXH 1PL.REL.PERF see
Kuma mun ga Hàliimà dà Dèelu.
also 1PL.PERF see H. and D.
‘We saw HAWWA, also we saw Halima and Deelu.’

The examples in (20) to (22) show that the meaning component introduced
by nee/cee cannot be cancelled. This suggests that nee/cee introduces a
conventional implicature in the sense of Karttunen & Peters (1979).
Second, nee/cee is illicit when the focus denotes an entity in a domain
that is explicitly introduced as containing more than the focused entity, as
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 253

illustrated in the following examples. If, as in (23a), a pluralic group is

introduced (mutàanee dà yawàa ‘many people’), a focus with nee cannot
pick a unique individual from this group (23b). In the absence of nee, the
focused entity can be one among others in the denotation of the predicate.
This is emphasized by the possibility of the additive particle maa ‘also’.

(23) a. Naa san mutàane dà yawàa

1SG.PERF know people many
dà sukà sayar dà àyàbà à kàasuwaa.
that 3PL.REL.PERF sell banana at market
‘I know many people that sold bananas at the market.’

b. Maalàm Shehù #née / (maa) ya sayar dà àyàbà.

Mr. S. EXH also 3SG.REL.PERF sell bananas
‘MR. SHEHU nee / (also) sold bananas.’

Assuming that nee/cee is an exhaustivity marker, the infelicity of (23b)

with nee follows directly: The presence of nee in this sentence indicates
that the property under discussion, i.e. the selling of bananas, only holds of
a unique individual. This is in contradiction with the plural group intro-
duced in (23a).
Third, nee/cee is also illicit in mention-some environments. Consider
the following context and the subsequent question-answer pair.

(24) Context: Musa knows that many students have passed last year’s
exam. In order to prepare for this year’s exam, Musa wants to talk
to one of them beforehand. (He has no time to talk to all of them).
Unfortunately, Musa does not know who passed the exam, but he
does know that his friend Amadu knows everybody who passed.
Therefore Musa addresses Amadu in the following way:

M: Kaa san wad’àndà sukà ci jarràbâwaa ?

you know who.PL 3PL.REL.PERF eat exam
‘Do you know who passed the exam?’
A: Î, dàgà ciki Ùmarù #nee/ maa ya ci j.
yes from among U. EXH also 3SG.REL.PERF eat exam
‘Yes, among them UMARU passed the exam.’
254 Katharina Hartmann and Malte Zimmermann

Amadu mentions to Musa one of the students that passed the exam last
year. In the answer, he cannot use the exhaustivity particle nee after the
focused subject since this would entail that only Umaru and nobody else
passed. This would contradict the contextual condition that both, Musa and
Amadu, know that many students were successful in the exam.
The infelicity of nee/cee in mention-some contexts can be mended by
means of accommodation: the property under discussion is specified in
such a way that it applies to a unique individual, in congruence with the
exhaustivity requirement. Reconsidering Amadu’s answer, the per se infe-
licitous presence of nee can trigger an accommodation such that the prop-
erty under discussion is not only that of passing the exam but that of pass-
ing it in a special way, e.g. with the highest or lowest score. This property
can now apply to the unique individual Umaru, as shown in (24’):

(24’) M: Kà fàÍaa minì: Wàa ya ci jarràbâwaa ?

2SG.SUBJ tell me who 3SG.REL.PERF eat exam
‘Tell me: Who passed the exam?’
A: Ùmarù nee ya ci jarràbâwaa
Umar EXH 3SG.REL.PERF eat exam
‘UMAR passed the exam (with the highest/lowest score etc.).’

A similar observation holds with respect to example (23). Below is a

slightly extended context, which is followed by a question-answer pair.

(25) Context: Maalam Haruna wants to buy bananas at the market. He

knows that there are many people selling bananas but not who
exactly. He does not have much time and only wants to get the
name of one of them. Therefore he asks his friend Maalam Shehu:

H: Kaa san waÍàndà sukèe sayar dà àyàbà ?

2sg.masc know who.PL 3PL.REL.CONT sell bananas
‘Do you know who sells bananas?’
S: Îi, dàgà ciki Hamiidù nee yakèe sayar-waa.
yes from among Hamidu EXH 3SG.REL.CONT sell-NMLZ
‘Yes, among them HAMIDU always/certainly sells bananas.’

Again, nee may follow the focused subject in Mr. Shehu’s answer, even
though an exhaustive focus interpretation contradicts the mutual knowledge
of Haruna and Shehu that many people sell bananas at the market. And
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 255

again, the presence of nee can be licensed by accommodation, leading to an

inherent quantification over times. It is understood that, among all the ba-
nana-sellers at the market, Hamidu always sells bananas. Thus, we are
faced with a methodological problem to be reckoned with: due to the possi-
bility of accommodation, native speakers will often judge nee/cee in men-
tion-some contexts as acceptable.

4.3. Inferences based on (strong) exhaustivity

The following example is a variation of (24). Recall that the context given
required a non-exhaustive interpretation of the focus. Accordingly, nee/cee
was illicit (without accommodation).

(26) Context: A student D’ (as in Íaalìbii ‘student’) who is anxious

that he might have failed a test approaches teacher M (as in
maalàamii ‘teacher’) and asks: ‘Can you tell me whether I have
passed or not?’ Unfortunately, teachers are by law forbidden to
tell a student directly about his or her result. However, there is no
law forbidding them to talk about other students’ performances.

D’: (Koo) naa ci jarràbâwaa ?

Q 1SG.PERF eat exam
’Have I passed the exam?’
M: Bà zâ-n gayà makà ba
NEG FUT-1SG tell you NEG
àmmaa Musà (nee) bà-i ci jarràbâwaa ba.
but M. PRT NEG-3SG eat exam NEG
’I will not tell you, but MUSA (nee) has not passed the test.’
Æ with nee: D’can assume that he has passed.
Æ without nee: D’ cannot find out anything about himself.

The context in (26) allows for nee in the answer in principle. However, the
amount of information differs depending on whether or not the teacher
decides to use the particle. If the particle is absent after the focus constitu-
ent Musa, the student learns about Musa’s result, but he cannot draw any
conclusions concerning his own score. If the particle is present, the student
can deduce that he passed the exam in the following way: Since the particle
marks the focus as exhaustive, Musa must be the only student who did not
pass. The student D’ can therefore infer that he must have passed the test,
256 Katharina Hartmann and Malte Zimmermann

although this is not explicitly asserted. If nee/cee were an optional focus

marker, there should be no asymmetry in interpretation between the two
variations. More precisely, the presence of nee/cee should not allow for any
inference which is based on exhaustivity.

4.4. Nee/cee and other exhaustivity markers

With adverbial exhaustivity markers, such as kawài ‘just, only, merely,

simply’, kaÍai ‘only, alone’, or sai ‘only, just, except’, nee/cee is typically
or often (Newmann 2000:190, Jaggar 2001:511) omitted.

(27) a. Sai Garba mukà ganii.

only Garba 1PL.REL.PERF see
‘It’s only Garba we saw.’
b. Hakà kawài zaa-kà yi.
this only FUT-2SG do
‘That is just what you have to do.’

The analysis of nee/cee as an exhaustivity marker predicts the typical omis-

sion of nee/cee with adverbial exhaustivity markers on grounds of redun-
dancy. When present, nee/cee can serve to reinforce kawái or kaÍai (Jaggar
2001:511). In contrast, an analysis of nee/cee as a plain focus marker leaves
the highly restricted occurrence of nee/cee with other exhaustivity markers
unexplained.

Even though the adverbial exhaustivity markers kawái or kaÍai ‘only’ and
nee/cee have similar semantic effects, the two kinds of expressions are not
identical in meaning. It shows that nee/cee is semantically weaker than the
adverbial exhaustivity markers. Compare (26-M) above, with nee present,
to (28-M), with nee replaced by kawài. According to our consultant’s
judgments, the difference between the two answers is the following: (28-M)
asserts that only Musa has not passed the exam, so that the student knows
for sure that he has passed, while (26-M) (with nee) makes the student only
assume that he must have passed.

(28) D’: (Koo) naa ci jarràbâwaa ?

Q 1SG.PERF eat exam
’Have I passed the exam?’
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 257

M: Bà zâ-n gayà makà ba àmmaa

NEG FUT-1SG tell you.m NEG but
Musà kawài bà-i ci jarràbâwaa ba.
M. only NEG-3SG eat exam NEG
’I will not tell you, but only MUSA has not passed the test.’
Æ Student knows for sure that he has passed.

The difference in interpretation between the mininal pair (26-M) and (28-
M) shows that the adverbials kawài and kaÍai introduce exhaustivity into
the assertion as part of their truth conditions. The exhaustivity marker
nee/cee, on the other hand, is weaker in that it does not add exhaustivity to
the assertion. Nee/cee only adds a conventional implicature to this effect. It
therefore does not translate as ‘only’. (Often, it does not translate at all,
which might also have led to the erroneous impression that it is a gram-
matical focus marker.) The presence or absence of nee/cee does not change
the truth-conditions of clauses. However, if nee/cee is dropped, the exhaus-
tivity effect disappears. This shows that the semantic effect is detachable.
That the semantic import of a lexical item is not cancelable, but detachable
is a typical property of conventional implicatures. We therefore conclude
that nee/cee triggers a conventional implicature.
Putting the results of this section together, we assume the following
meaning of nee/cee (where S stands for the clause containing nee/cee):

(29) [[nee/cee S]] = [[S]] = p defined iff

i. [[S]] z { [[S]] }
f 0
(Æ focus-sensitivity)
ii. p’  [[S]]f : Âp’ Æ p’ = [[S]]0 (Æ exhaustivity)

Nee/cee is a propositional operator that denotes a partially defined identity

function: When applied to an arbitrary clause denoting the proposition p, it
gives back the value p iff (i.) S has a non-trivial focus value (i.e. it contains
a focus) and (ii.) the only focus alternative that is true is p. The first clause
accounts for the focus-sensitivity of nee/cee, the second for the exhaustivity
effect.
Finally, by comparing the paradigms of focus-sensitive particles in
Hausa and English (or German) we observe that the Hausa paradigm is
more complete. While English only has a truth-conditional focus particle
with universal force (only), Hausa has both truth-conditional particles
(kawài, kaÍai) as well as a non-truth-conditional particle (nee/cee) with
universal force.
258 Katharina Hartmann and Malte Zimmermann

4.5. Summary

In this section, we have presented ample evidence in support of the claim,

originally hinted at in Green (1997), that nee/cee is an exhaustivity marker.
The presence or absence of nee/cee in a clause has semantic effects beyond
the introduction of those presuppositions that are usually associated with
focus: Nee/cee exhibits typical exhaustivity effects. First, it is infelicitous
or highly marked when the context suggests non-exhaustivity of the focus
domain. Second, it is typically left out in the presence of other exhaustivity
markers, such as kawài or kaÍai ‘only’. Nee/cee triggers an exhaustivity
effect by means of a conventional implicature, and, unlike only, not as part
of its truth-conditions. Finally, like exhaustivity markers in other lan-
guages, nee/cee is focus-sensitive, which accounts for its dependence on
focus. Being focus-sensitive, nee/cee can associate with focus constituents
at a distance, accounting for the non-adjacency with in-situ foci (see also
section 3.2). Altogether, these properties make an analysis of nee/cee as a
purely grammatical focus marker highly implausible. 10
Finally, the optionality of the exhaustivity marker nee/cee with ex-situ
foci and in wh-questions may have repercussions for the theory of focus
and questions in general. First, the dependency of an exhaustive interpreta-
tion of ex-situ foci on the presence of nee/cee shows that exhaustivity is not
structurally encoded in Hausa. In this respect, Hausa differs from Hungar-
ian, where, following É. Kiss (1998), exhaustive focus is always fronted.
Second, the optionality of the exhaustivity marker nee/cee in wh-questions
(see n.4 in section 3.1) suggests that wh-questions are not inherently ex-
haustive in Hausa. This is in contradiction to what has been claimed for
questions in other languages by Groenendijk & Stokhof (1984) but in ac-
cordance with claims in Heim (1994) and Beck & Rullmann (1999).

5. Conclusion

In this paper, we have argued that – despite first appearances – the Hausa
particle nee/cee is not a grammatical focus marker, but a focus-sensitive
exhaustivity marker. The particle nee/cee does not exhibit the typical prop-
erties of grammatical focus markers. Rather, its presence or absence is gov-
erned solely by semantic factors while its syntactic distribution seems to
depend on prosodic factors. However, since nee/cee is focus-sensitive, its
presence can serve as an indirect indicator for focus because focus-sensitive
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 259

elements need a focus in order to be licensed. The lesson to be learned is

that not every grammatical formative that frequently co-occurs with focus
constituents is best analyzed as a grammatical focus marker. From a cross-
linguistic perspective, Hausa seems to differ from many other West African
languages in that it does not have a grammatical focus marker. At the same
time, our findings for Hausa should be tested against other instances of
apparently optional focus markers in other African languages in order to
find out whether these elements are genuine grammatical focus markers, or
not just focus-sensitive exhaustivity markers.

Notes

We would like to thank our Hausa consultants Dan Asabe, Balarabe Zu-
lyadaini, Rabiu Shehu, Umar Ibrahim, Sa’adatu Garba, Aisha Mahmud,
and Mu’awiya Jibir for their help and cooperation. Many thanks to Daniel
Büring for his helpful comments and suggestions. This article was written
within project B2 “Focusing in Chadic Languages” as part of the SFB 632
“Information Structure”, funded by the German Science Association
(DFG). We thank the participants of the international workshop “Topic
and Focus: Information Structure and Grammar in African Languages”,
held in Amsterdam in December 2004, for comments on a preliminary
version of this paper, as well as Lutz Marten and Florian Schwarz for
comments on the present version.
1. Hausa is a Chadic language spoken primarily in northern Nigeria. The
Chadic languages belong to the Afro-Asiatic family. With more than 35
million speakers, Hausa is the biggest representative of the Chadic group.
2. We use the following abbreviations: 1,2,3 = person number markers, sg =
singular, pl = plural, perf = perfective, cont = continuous, rel = relative,
abs = absolute, fut = future, subj, = subjunctive, fem = feminine, masc =
masculine, NEG = negation, NMLZ = nominalizer, PRT = particle, EXH =
exhaustivity marker, DEF = definite.
3. Notice that subject focus in the aspects under discussion is marked
prosodically by local High-tone raising, as is the case with all other in-
stances of ex-situ focus (cf. Leben, Inkelas and Cobler 1989). Given this,
it could be argued that nee/cee is absent in (7A) because focus is marked
prosodically after all.
4. The question-pronouns for ‘who’ and ‘what’ can be either morphologi-
cally simple (wàa, mèe), or they can be complex (wàanee nèe, mèenee
nèe). In the latter case, they contain the particle nee/cee. If the particle oc-
curs in the question, its presence in the answer seems to be obligatory.
260 Katharina Hartmann and Malte Zimmermann

5. The judgements of our language consultants, all L1 speakers living mostly

outside of the Hausa heartland, varied considerably concerning the gender
of the particle with in-situ focus. Since kàazaa ‘chicken’ is feminine,
some speaker preferred cee here. While cee is always obligatory with
feminine singular ex-situ focus, the picture is not so clear with in-situ fo-
cus. We cannot offer an account for the gender variation other than con-
tributing it to dialectal differences, cf., e.g., Abubakar (2001).
6. The prosodic unit that Leben, Inkelas and Cobler (1989) call an intona-
tional phrase is referred to as phonological phrase in other approaches
(cf., e.g., Nespor and Vogel 1986). The term accent phrase is also used
sometimes, primarily for accent languages (e.g. Uhmann 1991).
7. The location of prosodic phrase boundaries can be tested by a number of
prosodic processes, the application or blocking of which is sensitive to
their presence, cf. Leben, Inkelas and Cobler (1989:47–49).
8. A similar observation has been made with respect to the Hausa discourse
particle fa. As Zec and Inkelas (1990:369ff) show, fa can only appear at
intonational phrase boundaries and is also excluded after verbs.
9. Note that double occurrences of nee/cee in declarative sentences are
equally ruled out. This is surprising given the possibility of multiple wh-
questions in Hausa (i-Q), see also Green (1997:116), as well as the possi-
bility to combine an ex-situ with an in-situ focus in the corresponding an-
swer (i-A).
(i) Q: Suwàa sukà ganii à ìnaa ?
who.pl 3pl.rel.perf see at where
‘Whom did they see where?’
A: Muusaa (nèe) na ganii à kàasuwaa.
M. PRT 1sg.rel.perf see at market
‘I saw MUSA at the MARKET.’
Double occurrences of nee/cee are expected to be grammatical as long as
one particle follows the in-situ focus. However, the only possible reading
of such sentences is that of a yes/no question where the “in-situ particle”
is interpreted as a question tag, indicating a certain degree of uncertainty
or suspicion:
(ii) Muusaa nèe ya sha ruwaa nè?
M. PRT 3sg.rel.perf drink water Q
‘Musa drank water, (didn’t he)?’
not: ‘MUSA drank WATER.’ (as an answer to ‘Who drank what?’)
A tentative solution to the impossibility of double occurrences of nee/cee
in declaratives would go as follows: Sentences with two instances of focus
are marked and require strong contextual licensing, e.g. in form of multi-
ple wh-questions (cf. i-Q). According to our consultants, the correspond-
ing wh-question for (ii) Who drank what? has the strong presupposition
that there are various people drinking various beverages. In section 4,
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 261

though, we will argue that nee/cee triggers an exhaustivity implicature. As

a result, the implicature of (ii), when interpreted as a declarative, would be
that only Musa drank only water (and nobody else drank anything else).
This implicature is incompatible with the presupposition of the licensing
question, ruling out (ii) as an answer.
10. Note that nee/cee is always obligatory in predicative constructions, where
the particle is usually described as a copula verb (cf. McConvell 1973,
Green 1997, 2004, Newman 2000, Jaggar 2001). (i) and (ii) illustrate this
for adjectival and nominal predicates:
(i) Teebùr ÝanÝanèe *(nee). (ii) Nii Bàtuurìyaa
*(cèe).
table small COP I European.fem COP
‘The table is small.’ ‘I am a European.’
The obligatory occurrence of nee/cee in predicative constructions can be
derived from the fact that predicates necessarily involve focus: in the
standard case, a (new) property is predicated of a (given) entity (see Green
2004). The proposed analysis of nee/cee as an exhaustivity marker pre-
dicts that the property denoted by the predicate is the only property (under
discussion) that holds of the subject. Further research has to show whether
this prediction for predicative constructions holds in general.

References

Abubakar, Abdulhamid
2001 Isoglosses in Hausa Dialectology. In Maiduguri Journal of Linguis-
tic and Literary Studies (MAJOLLS) 3, Mairo Kidda Awak (ed.), 60–
83.
Bearth, Thomas
1999 The contribution of African linguistics towards a general theory of
focus: Update and critical review. Journal of African Languages and
Linguistics 20. 121–156.
Beck, Sigrid, and Hotze Rullmann
1999 A Flexible Approach to Exhaustivity in Questions. Natural Lan-
guage Semantics 7. 249–297.
É. Kiss, Katalin
1998 Identificational Focus Versus Information Focus. Language 74. 245–
273.
Green, Melanie
1997 Focus and Copular Constructions in Hausa. Ph.D. dissertation,
SOAS, London.
262 Katharina Hartmann and Malte Zimmermann

Green, Melanie
2004 Predication, Equation and Information Structure: Evidence from
Hausa Copular Sentences. Ms. University of Sussex.
Green, Melanie, and Philip J. Jaggar
2003 Ex-situ and in-situ focus in Hausa: syntax, semantics and discourse.
In Research in Afroasiatic Grammar 2 (Current Issues in Linguistic
Theory), Jacqueline Lecarme (ed.), Amsterdam: Benjamins. 187–
213.
Groenendijk, Jeroen, and Martin Stokhof
1984 Studies on the Semantics of Questions and the Pragmatics of An-
swers. Ph.D. dissertation, Universiteit van Amsterdam.
Hartmann, Katharina, and Malte Zimmermann
2007 In Place - Out of Place? Focus in Hausa. In On Information Struc-
ture, Meaning and Form, Kerstin Schwabe and Susanne Winkler
(eds.), 365–403. Amsterdam: John Benjamins.
Heim, Irene
1994 ‘Interrogative Semantics and Karttunen’s Semantics for Know’. In
Proceedings of the Ninth Conference of the Israeli Association for
Theoretical Linguistics (IATL 1), Rhonna Buchalla and Anita Mitt-
woch (eds.), 128–144. Jerusalem: Akademon.
Inkelas, Sharon, and William R. Leben
1990 Where phonology and phonetics intersect: the case of Hausa intona-
tion. In Papers in Laboratory Phonology I. Between the grammar
and physics of speech, John Kingston and Mary E. Beckman (eds.),
17–34. Cambridge: Cambridge University Press.
Jaggar, Philip J.
2001 Hausa. Amsterdam: John Benjamins.
2004 More on in situ Wh- and Focus Constructions in Hausa. Ms., SOAS,
London. To appear in: Chadic Linguistics/Linguistique Tchadique/
Tschadistik, vol. 3., Dymitr Ibriszimow, Henry Tourneux, and Ekke-
hard Wolff (eds.), Cologne: Rüdiger Köppe.
Kanerva, Jonni M.
1990 Focus and Phrasing in Chichewa Phonology. New York: Garland.
Karttunen, Lauri, and Stanley Peters
1979 Conventional Implicature. In Syntax and Semantics 11, Choon-Kyu
Oh and David A. Dinneen (eds.), 1–56. New York: Academic Press.
Leben, William R., Sharon Inkelas, and Mark Cobler
1989 Phrases and Phrase Tones in Hausa. In Current Approaches to Afri-
can Linguistics, Paul Newman and Robert D. Botne (eds.), 45–61.
Dordrecht: Foris.
McConvell, Patrick
1973 Cleft Sentences in Hausa? A Syntactic Study of Focus. Ph.D. disser-
tation. SOAS, London.
 Exhaustivity marking in Hausa: A reanalysis of the particle nee/cee 263

Nespor, Marina, and Irene Vogel

1986 Prosodic Phonology. Dordrecht: Foris.
Newman, Paul
2000 The Hausa Language. New Haven & London: Yale University Press.
Parsons, Frederick William
1963 The Operation of Gender in Hausa: Stabilizer, Dependent Nominals
and Qualifiers. African Language Studies IV. SOAS, University of
London: 166–207.
Schachter, Paul
1966 A Generative Account of Hausa Ne/Ce. Journal of African Lan-
guages 5. 34–53.
Tuller, Laurice
1986 Bijective Relations in Universal Grammar and the Syntax of Hausa.
Ph.D. dissertation. UCLA.
Uhmann, Susanne
1991 Fokusphonologie. Tübingen : Niemeyer.
Zec, Draga, and Sharon Inkelas
1990 Prosodically Constrained Syntax. In The Phonology-Syntax Connec-
tion, Sharon Inkelas and Draga Zec (eds.), 365–378. Chicago: CSLI.
Part V

Focus and related constructions

Narrative focus strategies in Gur and Kwa

Anne Schwarz and Ines Fiedler

Abstract

The paper considers certain morphosyntactically marked focus constructions in

five Ghanaian languages which display special structural features in the out-of-
focus part of the sentence compared to the canonical clause. Such a phenomenon is
widespread in West African languages, and these non-canonical structures are
often labeled “relative” because they seem to be typical for relative clauses. How-
ever, while the parallels with relative clauses are never complete in the languages
studied, there is a constant structural isomorphism with clauses encoding sequen-
tial events in narration. Consequently, it is suggested that these focus constructions
should be analyzed as extra-clausal structures in which the non-focal information is
presented as a narrative clause.

1. Introduction

In this paper, we examine a phenomenon that is well known from several

West African languages like Hausa, Fulfulde, and others (cf. Bearth 1993,
Frajzyngier 2004), namely the parallels between morphosyntactically
marked focus constructions on the one hand and relative and narrative
clause morphology (and sometimes the morphology of other clause types)
on the other. In Hausa for example, one can differentiate between two mor-
phosyntactic aspectual codings: a canonical paradigm A and a paradigm B
which is found in focus constructions, relative clauses, and – with respect
to the perfective aspect – in narrative contexts. This second paradigm is
often called the “relative” tense/aspect. Our investigation shows that paral-
lels similar to these can also be found in languages of the Gur and Kwa
group. It also shows that part of the focus constructions can be analyzed as
morphosyntactically marked constructions containing a narrative clause.
Our language sample consists of five Ghanaian languages which we in-
vestigated during our field research in summer 2004. We deal with two Gur
languages: Buli and Dagbani which belong to different subgroups of the
Oti-Volta branch and are spoken in the Northern area of Ghana, as well as
268 Anne Schwarz and Ines Fiedler

with three Kwa languages of different subgroups: inland dialect of Ewe

(Gbe), Asante dialect of Akan (Potou Tano), and Lelemi, which is treated
as a Togo remnant or Ghana-Togo-Mountain language (na-Togo) which
according to Rongier 1997 (cited in Blench 2001) are still under discussion
as to whether they belong to Gur or to Kwa. The following grammatical
features shared by all of the five languages will be relevant for our discus-
sion: the strict SVO structure of canonical sentences and the lack of mor-
phological agreement on the verb phrase in the presence of nominal argu-
ments in simple sentences, with the exception of Lelemi. In this language,
lexical subjects co-occur with verb initial agreement markers.
We restrict the topic of this article to morphosyntactically marked con-
stituent focus structures (ex-situ) in affirmative utterances where a focused
nominal constituent takes sentence initial position. 1 The following findings
concerning the parallels between these constructions and relative as well as
narrative clauses emerge in our language sample: (1) there is no complete
correspondence with (restrictive) relative clauses while (2) the correspon-
dence with narrative clauses is surprisingly obvious. Our concern here is to
give evidence for the latter by presenting first results of ongoing work. The
similarity of ex-situ focus constructions with narrative clauses shows up
primarily in non-subject focus constructions. We will discuss the implica-
tions of these observations for the analysis of ex-situ focus constructions
and provide a diachronic approach to account for these patterns and the
language-specific differences concerning the degree of grammaticalization
of the narrative clause.
The paper is structured as follows: First, we present evidence for the
asymmetry between ex-situ subject und non-subject focus constructions
(henceforth SF resp. NSF) (2.1.). Then we demonstrate the parallelism
between ex-situ non-subject focus constructions (NSF) and narrative
clauses (NAR) (2.2.). Following this, we discuss the narrative pattern found
in ex-situ focus constructions and the various degrees of grammaticalization
of narrative sequential clauses into parts of focus constructions for each of
the sample languages (3.). Finally, we summarize our findings compara-
tively in the conclusion (4.).

2. Recurring features of ex-situ focus constructions

In the following, we will demonstrate two observations concerning the

structure of morphosyntactically marked focus constructions. 2
 Narrative focus strategies in Gur and Kwa 269

2.1. SF vs. NSF asymmetry

There is a constant structural asymmetry between SF and NSF construc-

tions which shows up in several ways and to different degrees in the se-
lected languages. We will start with weaker developed cases and then move
on to languages showing stronger asymmetries.

2.1.1. Ewe

The characteristics of ex-situ focus constructions in Ewe are as follows:

First, the focused element can be marked by the focus marker (henceforth
FM) ([)G"(cf. Ameka 1992) which is obligatory for subject focus and op-
tional for non-subject focus with the exact constraints for the optionality
not being clear yet. Second, there are special subject pronouns for 2nd and
3rd person singular which are used only in NSF constructions (2) while in
SF no pronominals are preposed to the verb (1).

SF
(1) P"VUW"³C"³G" VUn"³G
man-DEF-YE take-O:3sg
‘The MAN took it.’(not the woman)
NSF
(2) GF\(³G") YQ- ÇW
top(-YE) DEP:3sg eat
‘He WON.’ (i.e., He was on TOP.)

cf. the canonical sentence

 G" ÇW F\¸
3sg eat top
‘He WON.’ (i.e., He was on TOP. ~ He WAS ON TOP.)

2.1.2. Akan

In Akan, the constructions for both SF and NSF are characterized by two
features. They obligatorily make use of the FM nà (cf. inter alia Boadi
1974, Bearth 2002) and in the perfective aspect they display tonal changes
at the verb in the out-of-focus part: compare the falling tone at the verb in
the focus construction in (3) with the low verb tone in the following ca-
270 Anne Schwarz and Ines Fiedler

nonical sentence (cf. also the so-called “link tone” in Bearth 2002; cf.
Schachter 1973 as well).
In SF (3), an invariable expletive subject pronoun ('³), an anaphoric
subject pronoun, is used alternatively. 3 In NSF (4), on the other hand, there
is no general syntactic filler in the postverbal position. Instead, an ana-
phoric object pronoun has to be selected due to semantic criteria. While
animates require a pronoun (cf. 4a), especially when human, inanimates do
not allow one. In the absence of a pronominal object, the perfective verb in
sentence-final position carries the suffix -[' (cf. 4b).

SF
(3) '³[' CDTGYC" PQ" PC '³F¸º CFWC" PQ
EXPL-COP old.woman DEF NA EXPL-eat beans DEF
‘It is the OLD WOMAN who ate the beans.’

cf. the canonical sentence

CDTGYC" PQ" Fºº CFWC" PQ
old.woman DEF eat beans DEF
‘The old woman ate the BEANS.’ ~ ‘... woman ATE THE BEANS.’

NSF
(4) a. PG MTC"OC"P PC RCRC" PQ UW"C PQ
POSS:3sg dog NA man DEF carry O:3sg
‘The man carried his DOG.’
b. PG DC"IG PC n³UW"C³['"
POSS:3sg bag NA 3sg-carry-YE
‘He carried his BAG.’

2.1.3. Lelemi

In Lelemi, the difference between subject and non-subject focus construc-

tions lies above all in the verbal morphology. Lelemi has two sets of TMA-
markers: one used in simple tenses, the other in so-called “relative” tenses
(Allan 1973). Not every simple tense has a counterpart in the relative
tenses.
The “relative” tenses (relative past, relative present, relative future, and
relative present for verbs of state) appear in SF constructions. They display
only one invariant form prefixed to the verb in each tense for all persons.
 Narrative focus strategies in Gur and Kwa 271

The simple tenses, on the other hand, are used in NSF. They always indi-
cate tense and person (incl. gender agreement) in form of verbal prefixes.
The morphological coding device for NSF consists of the morpheme nà
following the focused constituent (cf. the homophone FM in Akan). Yet,
while some informants treated it as obligatory, others claimed that the syn-
tactic marking, i.e. the sentence-initial position, already makes clear that
this constituent is in focus.

SF not simple tense form:

(5) a. nPCCDº WOYº R'" On-F¸ MWVW" *nn-Fº
boy one only REL.PRS(dyn)-eat orange
‘Only ONE boy is eating an orange.’
b. WNGMW" n"On PC"³Fº *n"³F
woman DEF REL.PAST-eat
‘THE WOMAN ate them.’

NSF
(6) CMC"D" C"YnF" (PC) WNGMW n"On n"³F
beans raw (NA) woman DEF 3sg.PAST-eat
‘The woman ate RAW BEANS.’
cf. the canonical sentence
WNGMW" n"On n"³F CMC"D"
woman DEF 3sg.PAST-eat beans
‘The woman ate BEANS.’ ~ ‘The woman ATE BEANS.’

2.1.4. Buli

Buli has a preposed affirmative FM ká (negative suppletive FCC) in SF as

well as in NSF. The affirmative morpheme seems to be optional in both
constructions. In SF, the focused constituent is always followed by the con-
junction NG while in NSF we typically find the conjunction VG. 4 In NSF, the
occurrence of NGis less common but not totally excluded.
Verb tone deviates in both focus constructions from that in simple sen-
tences and depends on the respective conjunction. Buli has three paradigms
distinguished by grammatical tone on dynamic perfective verbs in clause-
final position: a canonical paradigm A in simple sentences, a paradigm B
after conjunction NG(SF), and a paradigm C that shows up in clauses with
the conjunction VG (NSF). 5 In the out-of-focus part of both marked con-
272 Anne Schwarz and Ines Fiedler

structions, there are no pronouns coreferent or agreeing with the focused

constituents.

SF
(7) (MC") O+¸ NG EJG0 Paradigm B, not: *VG
(FM) 1sg:DISJ CNJ go
‘I went. ~ It is ME who went.’ (not you)

cf. the canonical sentence

O¸ EJG"0³MC" UC"PFGO Paradigm A
1sg:DISJ go-FM 6 Sandema
‘I went to SANDEMA.’ ~ ‘I WENT TO SANDEMA.’
NSF
(8) a. (MC") UC"PFGO VG O¹ EJG0 Paradigm C, rare: NG
(FM) Sandema CNJ 1sg:DISJ go
‘It is SANDEMA where I went.’ (not Navrongo)
b. LQJP VG O+¸ HnD
John CNJ 1sg:DISJ slap
‘It is JOHN whom I slapped.’

2.1.5. Dagbani

Dagbani provides a strong structural asymmetry between SF and NSF. The

common SF construction contains a postponed syllabic nasal called “em-
phatic” by Olawsky (1999: 59f.) 7 . In NSF, the so-called FM MC (Olawsky
1999: 63) appears at the beginning of the out-of-focus part.
Similar to Buli, both focus construction types are distinguished from
each other by grammatical verb tone and deviate from the verb paradigm in
simple sentences. In general, neither focused verb arguments nor other
focused constituents are additionally expressed in the out-of-focus part.
SF
(9) RC"¢C" OC"xC"O Dn"N³n Paradigm B
woman DEF N call-O:3sg
‘The WOMAN called him.’
cf. the canonical sentence
Q DnN³NC" IGQTIG Paradigm A
8
sg call-FM George
‘She called GEORGE.’ ~ ‘She CALLED GEORGE.’
 Narrative focus strategies in Gur and Kwa 273

NSF
(10) IGQTIG MQ"³xQ" DnN+º Paradigm C
George KA-3sg call
‘She called GEORGE.’ ~ ‘It is GEORGE whom she called.’

2.2. NSF-NAR parallelism

Our second observation concerns the fact that there is a recurring formal
parallelism between the out-of-focus part of ex-situ non-subject-focus con-
structions (NSF) and narrative clauses (NAR). In some of our sample lan-
guages, the narrative structure extends to SF as well. We use the term NAR
for clauses that encode the succession of events in realis mood and that
serve to continuatively develop the main story line. Labov regards this
function as follows: “Each clause … describes an event that is understood
to shift reference time, i.e. it follows the event immediately preceding it,
and precedes the event immediately following it.” (1972, cited in Schiffrin
1994: 284). Syntactically, this succession of events is encoded iconically by
coordinated clauses which have special, language-specific structural fea-
tures: clausal conjunctions “and (then)”, particles, special pronouns, special
verb forms, etc.
In the following, we illustrate that the formal parallels show up in sev-
eral ways in the selected languages.

2.2.1. Ewe

In Ewe, one indication of the structural similarities between NSF and NAR
consists in the use of special subject pronouns (restricted to 2nd and 3rd per-
son singular) 9 in both constructions, as can be seen in ex. (11) (former (2))
and (12). Westermann (1930: 61) mentions that these pronouns are used “in
the continuation of a sentence, or closely to connect one sentence with a
preceding one”.
Additionally, the FM ([)G"resembles a clause coordinating conjunction
(G")[G" 10 ‘and (then)’ which is found in narrative contexts and which is often
reduced to [G".

NSF
(11) GF\(³G") YQ³ ÇW (= ex. 2)
top(-YE) DEP:3sg eat
‘He WON.’ (i.e., He was on TOP.)
274 Anne Schwarz and Ines Fiedler

NAR
(12) ÇGX"C" NnP G"[G" YQ³ F\G On"
child.DEF agree CNJ DEP:3sg arrive path
‘The child agreed and started the journey.’

2.2.2. Akan

Akan has a clausal sequential conjunction (é)nà with the meaning ‘and
(then)’ (Bearth 2002) which is identical with the FM – with the restriction
that the occurrence of the generally rare initial vowel has not been attested
so far within focus constructions. In the perfective aspect, the verbal mor-
phology including the verb tone changes in both clauses (cf. -F¸º³ in (13)–
(14) vs. -Fºº³ in the canonical simple sentence), although this remains a
matter of further research. 11

NSF
(13) CFWC" PC n³F¸º³['"
beans NA 3sg-eat-YE
‘He ate BEANS.’

NAR
(14) a. OC"COG" PQC"C CFWC" PC P³CFC"OHWQ F¸º³['"
Maame cook beans CNJ POSS:3sg-friend eat-YE
‘Maame cooked beans and her friend ate them.’
b. RCRC" PQ" HCC PG DC" PC n³UW"C PQ"
man DEF take POSS:3sg child CNJ 3sg-carry O:3sg
‘The man took his child and carried it.’

2.2.3. Lelemi

In Lelemi, NSF and NAR clauses show identical features: In both, the sim-
ple tense is used. Furthermore, the FM is homophone with the narrative
conjunction ‘and (then)’ which coordinates two clauses, and we suppose
that it is the same morpheme. It is segmentally identical with the “relative
past” tense morpheme as well (cf. ex. 5c).

NSF
(15) CMC"D" C"YnF" (PC) WNGMW n"On n"³F (= ex. 6)
beans raw (NA) woman DEF 3sg.PAST-eat
‘The woman ate RAW BEANS.’
 Narrative focus strategies in Gur and Kwa 275

NAR
(16) ‘The youngest child went …’
 PCW"³V¹ WNW I'y'y n"On
CNJ 3sg.PAST-take road right DEF
‘and he took the right road.’

2.2.4. Buli

Buli, too, displays a striking parallel between NSF and NAR. First, in both
structures, the same clausal conjunction VG ‘and’ is used. Second, the iden-
tity pattern stretches onto the verb in both: it bears the same grammatical
tone (paradigm C) after VG and differs thus from (i) the canonical paradigm
A and (ii) the paradigm B that is found in combination with the marker NG
(i.e. mainly with SF, cf. 2.1.4.). It should be noted, however, that beyond
the perfective aspect some structural peculiarities can be observed (cf. 3.4.).

NSF
(17) (MC") UC"PFGO VG YC EJG0
(FM) Sandema CNJ 3sg go
‘It is SANDEMA where he went.’

NAR
(18) ‘... and his mother was happy with him’
VG DC FºI LGPVC0C...
CNJ 3pl cook soup.DEF
‘and they cooked the soup ...’

2.2.5. Dagbani

Finally, Dagbani also affirms the parallel pattern between NSF and NAR
convincingly. The morpheme kà which follows the focused constituent in
NSF constructions has a clausal conjunction counterpart kà ‘and’ in narra-
tive contexts. Furthermore, the grammatical tone of dynamic verbs in such
clauses differs in the same way from the subject congruent verb tone in
simple clauses, irrespective whether we deal with a real narrative context or
a focus construction.
The coding of the second clause in ex. (20b) demonstrates Olawsky’s
(1999: 44) observation that if the subject of the clause introduced by kà is
coreferent with the subject of the preceding clause, it has to be elided.
276 Anne Schwarz and Ines Fiedler

NSF
(19) [¸N‹" OC"xC" Pº MQ"³xQ" Fº
house DEF in KA-3sg eat
‘In the HOUSE she ate.’

NAR
(20) a. ‘... and the mother sent the youngest child’
MC D¸¸ OC"xC" EJC0
CNJ child DEF go
‘and the child went ...’
b. RC"¢C" OC"xC" FC"C"³xNC" RGVGT MC" x0OG"³Q
woman DEF push-FM Peter CNJ hit-O:3sg
‘The woman pushed and hit Peter.’ not: MC*Q0OG³Q

2.3. Overview

Summarizing our observations concerning affirmative ex-situ focus con-

structions and narrative clauses, the ex-situ focus construction minus the
focus constituent itself (F) resembles more or less a narrative clause. All of
the five Ghanaian Gur and Kwa languages considered here show this phe-
nomenon and most of them display the parallelism especially between NSF
and NAR.
Table 1
Ewe Akan Lelemi Buli Dagbani
SF F (y)é ('[') F nà F (ká) F (à)lG FN
expl. ' relative tone B tone B
tone change tense
NSF F (y)é ('[') F nà F nà (ká) F (à)tè F kà
dep. pron. tone change simple tense tone C tone C
NAR ... (é)yé ... (é)nà ... nà ... (à)tè ... kà
dep. pron. tone change simple tense tone C tone C

Thus, we conclude that the parallelism between the out-of-focus part of

morphosyntactically marked (N)SF and narrative clauses (NAR) is no coin-
cidence but is due to a systematic narrative basis of the respective focus
constructions.
 Narrative focus strategies in Gur and Kwa 277

3. Narrative hypothesis

From the structural analyses above, it is evident that the parallelism be-
tween (N)SF and NAR is a systematic pattern. We propose that in fact a
narrative clause constitutes the non-focal part of such ex-situ focus con-
structions and suggest the following structure, here exemplified for the non-
subject focus construction:

(21) Extra-clausal structure of NSF

focal part: DP / clause non-focal part: narrative clause

[Non-Subject+/-focus marker] [CNJ 12 Subject Predicate]

In a simple transitive sentence with only one complement, the predicate in

the non-focal part contains the verb and only in certain cases also pronomi-
nal objects. The conditions of the pronouns’ occurrence are the same as in
real sequential narrative contexts and they are language-specifically deter-
mined by the animacy hierarchy.
We therefore do not follow the movement hypothesis as suggested, for
example, by Aboh (2004) and Green (1997) for Gbe and Hausa respec-
tively. Like in the cleft hypothesis assumed for various languages
(Schachter 1973, Heine/Reh 1984, Givón 1990, Lambrecht 2001, Drubig
2003), our proposal considers the constructions as involving a clausal
boundary, but unlike in the cleft hypothesis the non-focal part does not
represent a relative clause. Even if there are close structural correspon-
dences to restrictive relative clauses in most of these languages, some es-
sential elements like the obligatory relative pronoun resp. particle and the
common clause-final determiner are missing. For a more detailed discus-
sion of the incomplete relation between relative clauses and the non-focal
clause of ex-situ focus constructions compare Fiedler & Schwarz (2005:
120–126).
While the non-focal part of the ex-situ construction undoubtedly repre-
sents a clause of its own, the focal part has in most occurrences no clausal
status. For some languages, it is analyzed as a reduced clause with a zero
copula in affirmation (cf. Drubig 2003: 32f).
278 Anne Schwarz and Ines Fiedler

3.1. Ewe

The synchronic FM yé differs from the conjunction (G")[é ‘and (then)’ only
in the prefixed vowel. Thus, we assume – also based on the evidence in the
other languages of the sample – that the FM has developed out of the con-
junction leading to a functional split between conjunction and focus
marker. In normal speech, the latter is usually further eroded to vowel é and
suffixed to the preceding DP, with phonological reduction being a common
feature of grammaticalization processes. Additionally, Ewe displays a
homophone morpheme yé occurring in nominal predications like ‘It is a
pen.’ – RGP[G". Here, its function is comparable to a copula verb but not
identical with the identifying/classifying copula in Ewe 13 . This structure
can therefore be seen as an elliptical focus construction (cf. Ameka 1992)
which the background information is dispensed with.
The narrative analysis is supported by the requirement for special sub-
ject pronouns (in some persons) in NSF. Note that the correspondence be-
comes less close in SF where no coreferent subject pronoun is allowed in
the non-focal part even though it is required in narrative contexts.

3.2. Akan

An argument for the narrative hypothesis in Akan is the identity of FM and

narrative conjunction, at least if we disregard the conjunction’s rare initial
vowel. Our informants perceived nà in focus constructions still as conjunc-
tion so that if there is a certain degree of grammaticalization at all, as sug-
gested by its description as FM by some authors, this could only be due to a
functional split in the very inceptive stage.
The first part of the extra-clausal focus construction is often only repre-
sented by a DPAlternatively, the initial part starts with '[', i.e. an exple-
tive pronoun plus copula verb ‘to be’ (cf. ex. 3). ̓In the non-focal clause, the
subject position must always be filled although variation between the ana-
phoric form and an expletive pronominal form is possible in SF as well as
in narratives.

3.3. Lelemi

In Lelemi NSF, the non-focal clause is formally identical with the narrative
clause, i.e. any grammaticalization of the narrative clause is restricted to the
 Narrative focus strategies in Gur and Kwa 279

functional level and has no effects on the structural level. Accordingly, we

consider the morpheme nà in both functions a conjunction.
In SF, the conjunction is missing and “relative” tense forms are used in-
stead of the simple tense forms. Nevertheless, we can assume it to be pre-
sent in one of the “relative” tenses/aspects. The TMA morpheme for the
“relative past” tense is high toned ná. We analyze it as a conglomeration of
the conjunction nà (with inherent Low tone) plus a High tone which is
borne by the subject pronoun in the simple past.
Like in Akan, the slot for the subject has to be filled even though this is
only achieved by using an invariant form.

3.4. Buli

In Buli, the narrative hypothesis is valid for the prototypical NSF construc-
tion which is formed with conjunction VG and tone paradigm C. Since these
features are shared by sequential clauses in narrative contexts as well, the
non-focal part in NSF can be regarded to be represented by a narrative
clause.
Some structural peculiarities support the proposed narrative clause type.
While non-perfective verb forms are naturally excluded from sequential
events in real narrative contexts, they are subject to certain accommodation
in NSF constructions: in the imperfective aspect, dynamic verbs prefer a
subjunctive encoding while stative verbs appear in the same tonal form as
in canonical simple sentences.
The SF construction, on the other hand, requires the conjunction NG
which cannot be related to the narrative conjunction as such but is segmen-
tally identical with the NP coordinating conjunction NG‘and, with’. This
structural similarity among the two le-junctors might be an indication for a
semantico-syntactically closer conceptualization of the SF construction as
one single information structural unit compared to the evidently extra-
clausal NSF organization with VG.

3.5. Dagbani

Like in Buli, the extra-clausal NAR pattern is restricted to NSF in Dagbani.

There are no constitutional structural differences between the non-focal part
in NSF construction and the basic NAR clause. The so-called “focus
marker” MC (Olawsky 1999: 63) which requires verb tone paradigm C (for
280 Anne Schwarz and Ines Fiedler

dynamic verbs) is in both contexts analyzed as conjunction by us though it

has some potential for grammaticalization into a FM.
Interesting is the parallel to Buli as the “emphatic” (Olawsky 1999: 59f.)
marker N in SF resembles the DP coordinating conjunction (mì)nì~P ‘with,
and’. It seems that Dagbani has a similar tendency towards a more intra-
clausal organization of SF compared to NSF and hence does not make use
of the coordinative narrative pattern in SF.

4. Comparative summary

As we have shown, there are striking similarities on the morphosyntactic

level between focus constructions and narrative clauses. In all the lan-
guages we are concerned with here, this parallelism can be accounted for
by acknowledging that the conjunction of the narrative clause is an appro-
priate device to link together focus constituent and non-focal part in a non-
hierarchical way. In some of the languages, the inceptive stages of gram-
maticalization processes of the clause-initial conjunction into FM can be
perceived, a grammaticalization chain that may even stretch further into a
copula-like predicative morpheme, as noticed by Stassen (1997: 85). The
actual stage of such grammaticalization chains in our sample languages is
shown in table 2:
Table 2
CNJ o FM o COP~PRD
Ewe (é)yé yé, -é yé
Akan (é)nà (nà) --
Lelemi nà (nà) --
Buli (à)tè -- --
Dagbani kà (kà) --

Three of the languages, namely Akan, Lelemi 14 , and Dagbani, display the
same pattern insofar as they have a conjunction which has been interpreted
by some authors as a right-adjacent FM (Boadi 1974, Ameka 1992,
Olawsky 1999). According to our research, the respective morphemes do
have the potential for such a functional split but apparently that stage has
not yet been reached since we could not notice any relevant categorial or
structural changes of the conjunction towards a FM.
 Narrative focus strategies in Gur and Kwa 281

As for the clausal Buli conjunction, there are no indications at all that it
might take the grammaticalization path into a focus and then also into a
predicative marker in the near future. Responsible for that is firstly its re-
striction to NSF, a fact that the Buli conjunction shares with the respective
Dagbani conjunction. Secondly, the Buli clause conjunction is in affirma-
tive focal contexts relatively often accompanied by the FM ká left to the
focus constituent while such a frequently used affirmative counterpart is
missing in Dagbani. If the focused constituent is negated, all five languages
need, however, to use negative copula forms. We conclude that the rarer the
copula forms in affirmation are the higher are the chances for reanalysis of
the clausal conjunction as FM.
Contrary to the rather inceptive stage of grammaticalization – if existent
at all in most of the languages – the development in Ewe seems to have
been longer. The original conjunction already shows signs of erosion when
functioning as FM and often being suffixed to the constituent in focus.
As noted in 3.3., in Lelemi, the conjunction nà has taken a special direc-
tion in grammaticalization. Together with the High tone borne by the sub-
ject prefix in other syntagmata, it has become a “relative past” tense marker
in SF. Such development from a conjunction denoting the accomplishment
of actions to a past marker was also shown by Hopper (1979) for Malay, an
Austronesian language.
Table 3
CNJ o “Relative Past”
Lelemi nà ná (m nà + ´ )

An interesting picture in our small language sample is displayed by the

distribution of the narrative structures in ex-situ focus constructions, as
shown in table 4.
Table 4

Akan, Ewe Lelemi Buli, Dagbani

SF SF SF

NSF NSF NSF

282 Anne Schwarz and Ines Fiedler

In the Kwa languages studied, including Lelemi, the narrative pattern is

more or less overtly extending into SF constructions. On the other hand, in
the two Gur languages studied here, SF constructions do not participate in
the narrative pattern. Schwarz (in preparation) shows that in languages of
this group SF rather tends to be represented by a syntactically more hierar-
chical intraclausal construction and that the distribution of the two ex-situ
focus constructions can be accounted for on discourse organizational
grounds taking into account the notion of topic.
Our aim here was to propagate the existence of a constant narrative pat-
tern in ex-situ focus constructions as an alternative to cleft and movement
approaches. We showed that not only the in-focus part but also of the out-
of-focus part is important for an adequate analysis of ex-situ focus con-
structions. In this respect the functional load verb morphology, including
tone, should not be underestimated in African languages.

Abbreviations

CNJ conjunction
COP copula
DEF definite marker
DEM demonstrative (pronoun)
DEP dependent (pronoun)
DISJ disjunctive (pronoun)
DYN dynamic (verb)
EXPL expletive pronoun
F focus constituent
FM focus marker
NAR narrative (clause)
NSF non-subject focus (construction)
O object
PAST past tense
POSS possessive (pronoun)
PRD predicator
PRS present (tense)
PROG progressive marker
REL relative (tense)
SF subject focus (construction)
STAT stative (verb)
TMA tense-mood-aspect
 Narrative focus strategies in Gur and Kwa 283

Notes

This article was written within project B1 “Focus in Gur and Kwa Lan-
guages” as part of the SFB 632 “Information Structure” funded by the
German Science Association (DFG). We thank our colleagues who com-
mented the preliminary version of this paper which was presented at the
international workshop “Topic and Focus: Information Structure and
Grammar in African Languages”, December 2004, Amsterdam, and an
anonymous reviewer for critical comments.
1. For practical reasons, we will use the term focus constituent here also in
those cases in which only part of the clause constituent is focal.
2. Please note that the data is transcribed with surface tones and that versalia
in the English translation indicate the respective focal part of the utter-
ance. The following diacritics are used for tone marking: high tone ( ´ ),
mid tone (  ), low tone ( ` ), and downstepped high tone ( !´ ).
3. Bearth et al. (2002) describe the change of the subject pronouns as re-
stricted to human referents. This distribution is not supported by our data.
4. Both conjunctions are sometimes provided with an initial vowel
(CNGCVG). This vowel occurs with other clause-inital conjunctions as well
as with clause-initial serialized verbs and is always correlated with a pro-
sodic break before the clause.
5. For clause-final verbs in the indicative perfective, paradigm B is charac-
terized by an “instable rising tone” (Schwarz 2004: 38) and paradigm C
by an invariable low tone.
6. Please note that in the canonical indicative sentence in which either the
complement or the whole VP is focal FM MC"is encliticized to the verb. In
ex-situ focus constructions, any postverbal occurrence of this morpheme is
completely excluded.
7. There is another SF construction formed with post-subjectival NG"G" which
is almost restricted to questions and not considered here.
8. In case of complement or VP focus, suffixed FM ³NC"occurs in Dagbani
indicative sentences. Like Buli FM MC" this suffix is excluded from the
postverbal position in ex-situ focus constructions.
9. This has already been noted by Duthie (1996: 53) and Ameka (2004: 17).
10. Concerning the tonal behavior of the conjunction, we follow Clements
(1977: 172) who analyzes it as having overall high tones. The first high is
raised to extra-high in the beginning of a clause, therefore giving it a ex-
tra-high – high contour which is often interpreted as high – mid contour
(cf. Westermann 1954).
11. Bearth (2002) postulates the existence of a “link tone” on the verb as well
as the existence of a so-called “dependent” morpheme [' in ex-situ focus
284 Anne Schwarz and Ines Fiedler

constructions while our own data exhibit these two features in other con-
texts, too, including sequential events with clause-initial conjunction nà
12. In some languages (for example, Knnni, a Gur language closely related to
Buli), the most common sequential narrative clause type as well as the
out-of-focus clause occur without clause-initial conjunction (cf. Schwarz,
in preparation).
13. Cf. Ndayiragije 1992 who gives evidence for the non-verbal status of the
cognate FM w'in the closely related Gbe language Fon.
14. As can be seen in the table, Akan and Lelemi exhibit homophone mor-
phemes which could be a result of borrowing from Akan to Lelemi since
loans from Akan are common in the Togo mountain area.

References

Aboh, Enoch Oladé

2004 The morphosyntax of complement-head sequences. Clause structure
and word order patterns in Kwa: Oxford studies in comparative syn-
tax. Oxford, New York: Oxford University Press.
Allan, Edward Jay
1973 A grammar of Buem: The Lelemi language, School of Oriental and
African Studies, University of London: PhD Thesis.
Ameka, Felix
1992 Focus constructions in Ewe und Akan: A comparative perspective. In
Proceedings of the Kwa comparative syntax workshop M.I.T. 1992,
Chris Collins and Victor Manfredi (eds.), 1–15. Cambridge: MIT.
2004 Grammar and cultural practices: The grammaticalization of triadic
communication in West African languages. Journal of West African
Languages 30: 5–28.
Bearth, Thomas, et al.
2002 Ali Akan. Version 56. CD-ROM.
Bearth, Thomas
1993 Satztyp und Situation in einigen Sprachen Westafrikas. In Beiträge
zur afrikanischen Sprach- und Literaturwissenschaft, Wilhelm J. G.
Möhlig (ed.), 91–104. Köln: Köppe.
2002 Fokus im Akan - Outline. Handout of a presentation at the Linguistic
Colloquium of the Seminar of African Studies at the Humboldt-
University, 28.05.2002.
Blench, Roger
2001 Comparative Central Togo: what have we learnt since Heine? Paper
presented at 32nd Annual Conference on African Linguistics, Cam-
bridge.
 Narrative focus strategies in Gur and Kwa 285

Boadi, Lawrence A.
1974 Focus-marking in Akan. Linguistics 140: 5–57.
Clements, George N.
1977 Four tones from three: The extra-high tone in Anlo Ewe. In Lan-
guage and linguistic problems in Africa. Proceedings of the VII. con-
ference of African linguistics, Paul F. A. Kotey and Haig Der-
Houssikian (eds.), 168–181. Columbia: Hornbeam Press.
Drubig, Hans Bernhard
2003 Toward a typology of focus and focus constructions. Linguistics 41:
1–50.
Duthie, Alan S.
1996 Introducing Ewe linguistic patterns. Accra: Ghana University Press.
Fiedler, Ines, and Anne Schwarz
2005 Out-of-focus encoding in Gur and Kwa. Interdisciplinary Studies on
Information Structure 3, Shinichiro Ishihara, Michaela Schmitz, and
Anne Schwarz (eds.), 111–142. Potsdam: Potsdam University.
Frajzyngier, Zygmunt
2004 Tense and aspect as coding means for information structure: a poten-
tial areal feature. Journal of West African Languages 30: 53–67.
Givón, Talmy
1990 Syntax. A functional-typological introduction. vol. 2. Amsterdam:
John Benjamins.
Green, Melanie
1997 Focus and copula constructions in Hausa, School of Oriental and
African Studies, University of London: Ph.D thesis.
Heine, Bernd, and Mechthild Reh
1984 Grammaticalization and reanalysis in African languages. Hamburg:
Helmut Buske.
Hopper, Paul J.
1979 Some observations on the typology of focus and aspect in narrative
language. Studies in Language 3: 37–64.
Hopper, Paul J., and Elizabeth Closs Traugott
1993 Grammaticalization. Cambridge: Cambridge University Press.
Labov, Wilhelm
1972 The transformation of experience in narrative syntax. In Language in
the Inner City, 354–96. Philadelphia: University of Philadelphia
Press.
Lambrecht, Knud
2001 A framework for the analysis of cleft constructions. Linguistics 39:
463–516.
Ndayiragije, Juvénal
1992 Structure syntaxique des clivées en Fon. Journal of West African
Languages 22: 3–56.
286 Anne Schwarz and Ines Fiedler

Olawsky, Knut J.
1999 Aspects of Dagbani grammar. With special emphasis on phonology
and morphology: LINCOM studies in African linguistics; 41.
München, Newcastle: Lincom.
Rongier, Jacques
1997 Langues autonomes du Togo, entre Gur et Kwa?, Ms., n.p. (cited in
Blench 2001).
Schachter, Paul
1973 Focus and relativization. Language 49: 19–46.
Schiffrin, Deborah
1994 Approaches to discourse. Oxford, Cambridge: Blackwell.
Schwarz, Anne
2004 Aspekte der Morphosyntax und Tonologie im Buli. Mit Schwerpunkt
auf dem Buli von Wiaga. Dissertation, Humboldt-Universität zu Ber-
lin.
in prep. Topic/Focus-Intersection in discourse: Distribution of copulative and
narrative patterns in Gur.
Stassen, Leon
1997 Intransitive predication: Oxford studies in typology and linguistic
theory. Oxford: Clarendon Press.
Westermann, Diedrich
1930 A study of the Ewe language. London: Oxford University Press.
Focused versus non-focused wh-phrases

Enoch Oladé Aboh

Abstract

This paper discusses the relation between focused versus non-focused wh-phrases
and seeks to identify how they frame the information structure in wh-question-
answer pairs. It is argued that the well-known complementary distribution between
focused constituents and wh-phrases masks another reality: Wh-phrases need not
be focused. Cross-linguistic evidence indicates that, in languages where wh-
phrases lack focus morphology, these do not interact with focused constituents,
unlike focused wh-phrases which appear sensitive to focused constituents. Accord-
ingly, it is suggested that focused wh-phrases and non-focused wh-phrases require
different formal licensing: They target different structural positions in syntax and
consequently require different information structuring.

1. Introduction

This paper discusses the relation between focused versus non-focused wh-
phrases and seeks to identify how they frame the information structure in
wh-question-answer pairs. I informally define focused wh-phrases as those
wh-phrases that are displaced to a designated focus position (where they
may co-occur with a focus marker). Non-focused wh-phrases, on the other
hand, represent wh-phrases that occur in other positions than the focus po-
sition.
It is commonly assumed that focused constituents and wh-phrases are
closely related because they interact in question-answer pairs and appear to
exclude each other in many languages. In terms of Hovarth (1986) and
much related work, such interaction is regarded as piece of evidence that
wh-expressions are inherently focused. In a framework that assumes the
existence of a focus feature [F] in syntax, this amounts to saying that wh-
phrases and focus phrases bear the feature [F] and therefore fall under the
same formal licensing constraints (e.g. Haegeman 1995, Rizzi 1996, 1997,
2004, Aboh 2004a among others).
288 Enoch Oladé Aboh

This paper shows that such characterization might be misleading. I sug-

gest instead that the apparent complementary distribution of wh-phrases
and focused expressions masks the fact that wh-phrases need not be fo-
cused. Cross-linguistic evidence indicates that in languages where wh-
phrases do not embed any focus features these do not interact with focused
constituents. On the other hand, only focused wh-phrases appear sensitive
to focused constituents. With regard to syntax, these facts suggest that fo-
cused wh-phrases and non-focused wh-phrases require different formal
licensing and therefore target different structural positions. I further pro-
pose that wh-phrases of different types require different information struc-
tures. In question-answer pairs, for instance, focused wh-phrases require an
answer containing a focus marked constituent, unlike non-focused wh-
phrases which don’t seem to impose such a restriction on the answer. In-
stead, non-focused wh-phrases only convey non-presupposed information
for which the comment holds (É. Kiss 1998), but the constituent expressing
such information cannot be formally focus marked. In languages with a
focus marker, for instance, it appears that such responses do not typically
involve a constituent associated with a focus marker.
The paper is structured as follows. Section 2 discusses the interaction
between focus phrases and wh-phrases and concludes that there are two
focus projections within the clause – at the clause periphery and at the VP-
periphery – where these elements can occur. Section 3 distinguishes be-
tween focused and non-focused wh-phrases and shows that focused wh-
phrases, on a par with focused constituents, may occur within the clause
left periphery or at the VP-periphery where they realize the specifiers of
focus projections. Non-focused wh-phrases, on the other hand, occur in
various IP-internal positions. With regard to information structure of ques-
tion-answer pairs, it is proposed in section 4 that the focus or non-focus
status of the relevant constituent in the answer depends on the focus or non-
focus status of the wh-phrase in the question. This view implies a direct
mapping of information structure onto clause structure, as already proposed
in the literature (Drubig 2003). Section 5 concludes the paper.

2. The interaction between focused phrases and wh-phrases

The relation between focused items and wh-phrases has always been an
intricate one. The following sections discuss their interaction with regard to
clause structure and information structure.
 Focused versus non-focused wh-phrases 289

2.1. Clause structure, focused phrases, and wh-phrases

In the following Gungbe examples (1a–b) and (1a’–b’), the focused con-
stituents and the wh-phrases are parallel because they occur in a designated
position to the left of the focus marker w(a. 1 These Gungbe examples could
be taken as representative of languages where focused expressions and wh-
phrases appear to follow the same syntactic path.

(1) a. S(sínú w(a dà Àsíàbá [Gungbe]

Sessinou Foc marry Asiaba
‘SESSINOU married Asiaba.’
a’. M(nù w(a dà Àsíàbá ?
who Foc marry Asiaba
‘WHO married Asiaba?’
b. Àsíàbá w(a S(sínú dà
Asiaba Foc Sesinou marry
‘Sessinou married ASIABA.’
b’. M(nù w(a S(sínú dà ?
who Foc Sessinou marry
‘WHO did Sessinou marry?’

The ungrammatical sentences in (2) further show that this position is

unique in Gungbe (and most Kwa languages) because focused constituents
and wh-phrases are mutually exclusive in it. 2

(2) a.* Àsíàbá w(a m(nù w(a dà ? [Gungbe]

Asiaba Foc who Foc marry
b.* M(nù w(a Àsíàbá w(a dà ?
who Foc Asiaba Foc marry

Similar facts are reported in typologically different languages. Rizzi (1997,

2001), for instance, indicates that focused expressions and wh-phrases are
in complementary distribution in Italian main clauses (3). As already illus-
trated by the Gungbe examples in (2), this incompatibility persists in Italian
too, irrespective of whether the focused constituent precedes or follows the
wh-phrase.

(3) a.* A chi questo hanno detto (non qualcos’altro)? [Italian]

To whom THIS they said (not something else)?
290 Enoch Oladé Aboh

b.* Questo a chi hanno detto (non qualcos’altro)?

THIS, to whom they said (not something else)?

In many languages where there seems to be no focus movement, the coun-

terparts of the Gungbe sentences under (1) quite often translate as clefts.
Such a systematic correspondence between focus movement constructions
and cleft constructions is commonly taken as evidence for a close relation
between clefts, focus sentences, and questions. In certain French varieties,
for instance, questions involve clefting, as shown in (4a–b). Observe, how-
ever, from the ungrammatical examples (4c–d) that such questions cannot
include other clefted constituents.

(4) a. C’est qui qui est venu ? [French]

It.is who who be come
‘WHO came?’
b. Qu’est ce que tu as fait ?
What.is it that you have done
‘WHAT did you do?’
c. * C’est quoi, c’est Jean qui a vu ?
It.is what it.is John who has seen
d. * Qu’est-ce que c’est Jean qui a vu ?
What.is it that it.is John who has seen

These data indicate that clefted constituents and clefted questions exclude
each other in a way similar to wh-phrases and focused constituents in
Gungbe and Italian.
If clefting is related to focusing to some extent, we may further con-
clude that focused constituents and wh-phrases are mutually exclusive.
Similar comparisons abound in the literature and give credibility to the idea
that focused expressions and wh-phrases are in competition for the same
position in syntax. In a framework that assumes a single specifier position,
the ungrammatical sentences in (2) and (3) result from the impossibility of
merging two distinct elements (or phrases) in the same position. The same
reasoning extends to the cleft constructions in (4), where ungrammatical
sentences (4c-d) can be accounted for by saying that cleft constituents and
wh-phrases target the same position (e.g. in French).
In terms of Rizzi (1997), focused constituents and wh-expressions occur
in the specifier of a focus phrase whose head may be overtly realized in
some languages. This analysis clearly accounts for data from Gungbe-type
 Focused versus non-focused wh-phrases 291

languages where there is a focus particle that is always adjacent to the fo-
cused phrase or the wh-phrase, as shown in (5a–b) for the Gungbe sen-
tences (1a–a’) respectively, see Aboh (2004a) and references cited there for
discussion.

(5) a. [CP [FocP S(sínú [Foc w(a [TP tS(sínú dà Àsíàbá]]]]

b. [CP [FocP M(nù [Foc w(a [TP tm(nù dà Àsíàbá]]]]

The idea that wh-phrases and focused phrases belong to the same class of
(quantificational) elements led to formalisms such as Hovarth’s (1986: 118)
focus constraint on wh-operators which suggests that “A non-echo question
interpretation can be derived only if the Wh-Q operator bears the feature
focus”. It appears from Hovarth’s definition that wh-phrases triggering
questions must be focused (and quantificational). However, close scrutiny
reveals that this generalization does not hold for all the relevant cases be-
cause wh-phrases are not always focused cross-linguistically.

2.2. Information structure, focused phrases, and wh-phrases

In terms of information structure, the interaction of focused constituents

and wh-phrases seems to correlate with the fact that wh-questions trigger
focused answers in question-answer pairs, as represented in (6). Here, the
Gungbe sentences in (6b) and (6c) represent felicitous replies to the ques-
tion in (6a). In example (6b), the in-situ focused constituent l(sì acts as new
information focus. In the sentence (6c), however, the same constituent en-
codes identificational focus and has been fronted. It is worth mentioning
that, provided the appropriate context, the choice of answer (6c) over (6b)
implies a contrast or requires a continuation of the sentence (e.g. he ate rice
and became sick).

(6) a. Ét( w(a Kòfí ˜ù ? [Gungbe]

what Foc Kofi eat
‘WHAT did Kofi eat?’
b. É ˜ù l(sì
3sg eat rice
‘He ate rice.’
c. L(sì w(a é ˜ù (bò b( àwútù)
rice Foc 3sg eat and start sickness
‘He ate RICE (and became sick).’
292 Enoch Oladé Aboh

Within the generative framework, studies on the interaction between infor-

mation structure (or discourse semantics) and clause structure (e.g. É. Kiss
1998) regard the data in (6) as partial evidence that in-situ focused con-
stituents (6b) and ex-situ focused constituents (6c) target two focus posi-
tions in the clause. The post-verbal position is specialized in encoding new
information focus, but the peripheral position expresses exhaus-
tive/identificational focus. I will return to this asymmetry in (6) in section 4
(see also Zubizarreta 1998, Drubig 2003, Belletti 2002, and references cited
there for discussion).
At this stage of the discussion, however, it is worth mentioning that un-
der such a view of clause peripheral and clause internal focus positions the
traditional ex-situ versus in-situ distinction appears to be misleading. In-
deed, the idea that clause structure involves two focus positions (i.e. clause
periphery vs. VP-periphery) where focused elements occur implies that ex-
situ strategies include focused elements that appear in the clause periphery
(i.e. within CP) or within the VP-periphery. This suggests that focused
elements necessarily move to a designated position. On the other hand,
genuine in-situ strategies would involve elements that are internally fo-
cused (e.g. a DP-internal focus, see Aboh 2004b) and which occur in posi-
tions (including first merge positions) other than designated focus posi-
tions. This in turn implies that the term in-situ should not be understood to
hold for first merge positions only. In this paper, I argue that syntactically
focused elements appear in the VP-peripheral or clause peripheral focus
positions, but non-focused constituents are excluded from such positions
and must realize some other positions within the clause.
Going back to (6) and the assumed informational difference between the
VP-peripheral focus position and the clause peripheral one, a look at
Gungbe subject wh-phrases suggests that things might not be so clear-cut.
Indeed, Gungbe subject wh-phrases do not allow for an answer where the
target subject remains IP-internal. Given the question in (7a), for instance,
the ungrammatical sentence (7b) indicates that a reply without the focus
marker is infelicitous, unlike (7c), which includes the focus marker.

(7) a M(nù w(a ˜ù l(sì ? [Gungbe]

what Foc eat rice
‘WHO ate rice?’
b.* Kòfí ˜ù l(sì
Kofi eat rice
‘KOFI ate rice.’
 Focused versus non-focused wh-phrases 293

c. Kòfí w(a ˜ù l(sì

Kofi Foc eat rice
‘KOFI ate rice.’

Put together, these Gungbe facts lead us to conclude that there seem to be
two focus positions involved in object wh-question-answer pairs (clause
peripheral vs. VP-peripheral) with possible differences in interpretation
(e.g. contrast, D-linking). On the other hand, subject wh-question-answer
pairs appear to exploit just one (i.e. clause peripheral) position.
Some facts from Italian wh-questions provide us with complementary
data on this issue. Based on the distribution of new information focused
constituents in Italian question-answer pairs, Belletti (2002) argues that
Italian post-verbal focused subjects occur in a low focus position immedi-
ately dominating the VP, where they express new information focus only.
On the other hand, sentence-initial focus is used to realize contrastive fo-
cus. Accordingly, sentence (8b) is a felicitous answer to the subject wh-
question in (8a), unlike example (8c).

(8) a. Chi ha parlato? [Italian]

Who has spoken
‘Who spoke?’
b. Ha parlato GIANNI
has spoken Gianni
‘Gianni spoke.’
c.* Gianni ha parlato

In the same context, however, sentence (9), pronounced with primary stress
(or emphasis) on Gianni, can only be interpreted as contrastive focus.

(9) GIANNI ha parlato (non tutta la classe). [Italian]

‘Gianni has spoken (not the whole class).’
‘GIANNI spoke (not the whole class).’

Belletti (2002) and much related work therefore concludes that the clause
structure involves two focus positions that correlate with informationally
different focus expressions: new information focus versus exhaustive focus
(É. Kiss 1998).
Brunetti’s (2003) discussion of object wh-questions suggests a different
picture. According to the author, sentence (10b), with the object in-situ, is
294 Enoch Oladé Aboh

an appropriate answer to the question (10a) and expresses new information.

On the other hand, sentence (10c), even though marginally felicitous, does
not imply a contrast (Brunetti 2003).

(10) a. Che cosa ti ha regalato Gianni? [Italian]

‘What did Gianni give you as a present?’
b. Gianni mi ha regalato [un libro]
‘Gianni me has given a book.’
c. ??[Il libro]F mi ha regalato, non il cd.
the book me (he) has given not the cd
‘He gave me the book, not the cd.’

Brunetti (2003) further shows that the contrast in (10b–c) disappears in

non-interrogative contexts where both post-verbal and clause peripheral
focus positions are used without pragmatic differences. Starting with the
statement in (11a), the reactions in (11b–c) indicate that both positions can
encode contrastive focus.

(11) a. La felpa, l’ha vinta Gianni. [Italian]

‘As for the sweatshirt, Gianni won it.’
b. No, Gianni ha vinto LA MAGLIETTA.
‘No, Gianni won the T-shirt.’
c. No, LA MAGLIETTA ha vinto Gianni.
‘No, it is the T-shirt that Gianni won.’

Brunetti (2003) concludes from this that focus, whether realized at the
clausal periphery or at the VP-periphery, has the same informative value at
any of these interfaces.
Put together, the Italian and Gungbe facts therefore support the view
that there are two linear focus positions within the clause, one at the clausal
left periphery and one at the VP-periphery (i.e. post-verbal position). How-
ever, these data are inconclusive as to whether these two positions map
onto informative differences such that the low focus position is restricted to
information focus while the higher position is devoted to contrastive focus.
In recent years, the expression of clause peripheral focus has been
shown to be a property of a focus projection (FocP) within the complemen-
tizer system (Rizzi 1997). The specifier of this projection hosts the focused
constituent while its head may be realized by a focus marker, as is the case
in certain African languages (e.g. Kwa; Aboh 2004a).
 Focused versus non-focused wh-phrases 295

Not much is known, however, about the structural properties of the VP-
peripheral focus position described in previous paragraphs, and there is still
some doubt as to whether a focus phrase could project in the lower portion
of the clause where the post-verbal focus is licensed. The following section
provides additional empirical evidence supporting the existence of such a
focus projection. The discussion shows that an asymmetry between certain
Kwa (e.g. Gungbe), Romance (e.g. Italian), and Bantu languages (e.g.
Aghem) with regard to expression of focus could reduce to the question
whether they solely use the clause peripheral or VP-peripheral focus pro-
jection, or both.

2.3. Focus at the VP-periphery

This section discusses data from Aghem (Watters 1979, Hyman 1979,
2005), but the reader is referred to Biloa (1997), Sabel & Zeller (2002),
Downing, Mtenje and Pompino-Marschall (2004), Good (2005), and refer-
ences cited there for discussions of post-verbal focus in other Bantu lan-
guages (e.g. Chitumbuka, Zulu). Aghem is an SVO Grassfield Bantu lan-
guage spoken in Cameroon that displays the basic structure exemplified in
(12a) and illustrated in (12b). 3

(12) a. T,-bvÊ tËa-bËaghà m2¶ zËa kË-b( n( [Aghem]

dogs two Past eat fufu today
‘The two dogs ate fufu today.’ [Hyman 2005: 1]
b. S... Aux… Verb… (Focus)… Object… Adjunct

Starting from representation (12b), the examples in (13) and (14) indicate
that a contrastively focused constituent as well as a wh-phrase must occur
in the position immediately after the verb. The sentence in (13a) is an in-
stance of object focusing. Here, the focused object is right adjacent to the
verb while the subject occurs in the canonical position (i.e. pre-verbally). In
addition, the locative adjunct phrase án ‘sóm ‘in the farm’ is displaced to a
pre-verbal position. This example contrasts, however, with the subject fo-
cus construction in (13b) where an expletive realizes the canonical subject
position and the associate DP-subject occurs post-verbally. Finally, the
adjunct focus in (13c) indicates that the focused adjunct must occur imme-
diately to the right of the verb.
296 Enoch Oladé Aboh

(13) a. Éná2 m2a án ‘sóm zi [b(-k2] [Aghem]

Inah Past in farm eat fufu
‘Inah ate FUFU at the farm.’
b. Á m2a zi [éná2] b(-k2 án ‘sóm
Expl Past eat Inah fufu in farm
‘INAH ate fufu at the farm.’
c. T,-bvÊ tËa-bËaghà m2¶ zËa n( b(-k2
dogs two Past eat today fufu
‘The two dogs ate fufu TODAY.’ [Hyman 2005: 1]

Similarly, Aghem wh-phrases occur in the same position as focused con-

stituents, that is, immediately after the verb, as shown in (14) (Hyman
2005). The sentence in (14a) illustrates adjunct wh-phrases, while the ex-
ample in (14b) represents subject wh-phrases. Note from sentence (14c)
that in multiple wh-questions one wh-phrase immediately follows the verb
(presumably in the focus position) while all the others follow successively
(see Biloa 1997, Hyman 2005).

(14) a. Fil a-m2¶ zËa zËn b(-k2 ? [Aghem]

friends SM.Past eat when fufu
‘WHEN did the friends eat fufu?’
b. À m2a zËa ndúgh2 b(-k2 ?
Expl Past eat who fufu
‘WHO ate fufu today?’
c. À m2a zËa ndúgh2 kw2ak2a zËn ?
Expl Past eat who what when
‘WHO ate what when?’

The distribution of focused constituents and wh-phrases in Aghem as well

as the correlation between a focused subject and the presence of an exple-
tive in the sentence (e.g. 12a vs. 13b) suggest that this language does have a
fixed focus position immediately after the verb. This description is further
supported by the fact that focusing on the truth value of the proposition
requires the presence of the Aghem focus marker nó in the position imme-
diately after the verb, as illustrated in (15a). The example in (15b) indicates
that a contrastively focused object may left adjoin to the focus particle nó.
This example shows that the verb and the focus particle do not form a con-
stituent. Instead, it seems as if the focus particle nó scopes over the con-
stituent immediately left adjacent to it.
 Focused versus non-focused wh-phrases 297

(15) a. T,-bvÊ tËa-bËaghà m2¶ zËa nó b(-k2 [Aghem]

dogs two Past eat Foc fufu
‘The two dogs ATE fufu.’ [Hyman 2005: 1]
b. ZËa b(-k2 nó
eat fufu Foc
‘eat FUFU’

If we take the Gungbe facts discussed previously, where – in order to be

somewhat comparable to the Aghem data – the focus marker w(a occupies a
focus head position within the clausal left periphery and attracts a focused
element to its left (5), we are lead to conclude that the Aghem functional
category expressed by the particle nó plays a function similar to the
Gungbe focus category realized by w(a. The two languages differ, however,
because the Aghem focus category occurs within the VP-periphery only.
More precisely, I suggest that the Aghem particle nó occupies a focus head
position immediately to the right of the surface position of the verb.
Even though Gungbe, Italian, and Aghem differ in many respects as to
how they realize the category focus, I regard the facts discussed thus far as
representative of the distribution of structural focus positions cross-
linguistically. These typologically different languages therefore suggest the
following description, where languages only differ with respect to whether
they realize the clause peripheral or VP-peripheral focus positions or both. 4

Gungbe [Focus[subject/object, adjunct]- w(a ...V…Focus[object/ adjunct]…..]

Italian [Focus[subject/object, adjunct]…… V…Focus[subject]…..]
Aghem […………………………….V…Focus-[subject/object/ adjunct]-nó] 5

Clause structure [Focus….[subject…verb…Focus…]]

This description is compatible with Belletti’s (2002) representation of the

clause structure in (16), where there is a topic/focus articulation both at the
clausal left periphery and at the VP-periphery. In languages of the Gungbe-
or Aghem-type, focus markers head these projections (Aboh 2004a).

(16) [ForceP [TopP [FocP [FinP …[ TP ....[TopP [FocP [VP…]]]]]]]]

Under this approach, the Italian post-verbal focus occurs in the VP-
peripheral focus position. In a sentence like (8b), the subject follows the
verb because the latter raises to some higher position in the clause to be
298 Enoch Oladé Aboh

licensed for tense and aspect as schematized in (17a). The same reasoning
applies to the Aghem example (15b). Following Aboh’s (2004a) proposal
of V-to-Asp movement as determining VO versus OV structures in Kwa as
well as Sabel and Zeller’s (2002) analysis of Nguni (Zulu) as involving V-
to-T movement, I propose (by analogy) that the verb in Aghem must move
to an aspect position in overt syntax in order to be licensed for aspect. 6 As a
result of this movement, the verb necessarily precedes constituents or wh-
phrases that move to the VP-peripheral focus position, as illustrated in
(17b). This analysis extends to wh-phrases in Aghem.

(17) a. AspP b. AspP

2 2
spec Asp' spec Asp'
2 2
Asp° FocP Asp° FocP
parlatoi 2 zËa 2
spec Foc' spec Foc'
Gianni 2 b(-k2 2
Foc° VP Foc° VP
6 (nó) 6
tGianni tparlato tzi tb(-k2

With regard to information structure, we have already seen from the discus-
sion on focus constituents in Gungbe and Italian (section 2.2) that there
seems to be no systematic mapping of É. Kiss’ (1998) identificational focus
onto clause peripheral focus and new information focus onto VP-peripheral
focus.
Nevertheless, given the interaction between focused constituents and
wh-phrases, one could still ask whether such a dichotomy may arise with
wh-questions instead. More precisely, one wonders whether wh-phrases are
sensitive to the contrastive versus new information distinction such that, in
question-answer pairs, the focus value of the answer would match that of
the wh-phrase in the question. Under such a view, a contrastive focus wh-
question would require a contrastive focus constituent in the clausal left
periphery of the answer while a new information focus wh-question would
require a constituent carrying new information to occur within the VP-
periphery of the answer (É. Kiss 1998, Belletti 2002). The next section
deals with these issues.
 Focused versus non-focused wh-phrases 299

3. Focused wh-phrases versus non-focused wh-phrases

The discussion in section 2 suggests that wh-phrases and focus phrases

exclude each other because they target the same position. This leads to the
conclusion that wh-phrases have something focal about them and must be
focused cross-linguistically (Hovarth 1986). However, close scrutiny indi-
cates that wh-phrases do not always occur in a focus position cross-
linguistically. Instead, it appears that some languages distinguish between
focused and non-focused wh-phrases.
In Italian matrix clauses (cf. 3), for example, the incompatibility be-
tween focused constituents and wh-phrases that serves as empirical evi-
dence for regarding wh-phrases as inherently focused elements appears less
conclusive when it comes to embedded clauses. Consider the following
sentences.

(18) a. */? Mi domando [a chi] QUESTO abbiano detto

(non qual cos’altro).
‘I wonder to whom THIS they have said
(not something else).’
b. */? Mi domando QUESTO [a chi] abbiano detto
(non qualcos’ altro).
‘I wonder THIS to whom they have said
(not something else).’
c. Mi domando A GIANNI [che cosa] abbiano detto (non a Piero).
‘I wonder TO GIANNI what they have said (not to Piero).’

As already discussed in Rizzi (2001) and much related work, these sen-
tences show that when the focused element is the direct object and the wh-
phrase is a prepositional phrase both orders are degraded as in (18a–b). On
the other hand, when the wh-phrase is the direct object and the focused
phrase involves a prepositional phrase, as in (18c), the order focus preced-
ing wh-phrase becomes acceptable (or marginal). Rizzi (2001) concludes
that when wh-phrases are not forced to move to the focus position (as in the
Italian matrix clauses), they will not move there. In embedded clauses,
therefore, wh-phrases move to a lower position than the focus position, as
indicated by sentence (18c), where the focus element necessarily precedes
the wh-phrase. These data suggest that wh-phrases must be focused in ma-
trix clauses but need not be in embedded clauses. Put another way, wh-
phrases do not always occur in a focus position because they are not always
300 Enoch Oladé Aboh

focused, contra Hovarth (1986), see Frascarelli & Puglielli (this volume for
similar conclusion). 7
Yiman (1988: 370) reports similar facts in Oromo (a Cushitic SOV lan-
guage) that displays both non-focused and focused wh-phrases. According
to the author, the question in (19a), based on the presupposition that some-
one came, requires the answer in (19b).

(19) a. EeĖĖu GCuf-e ? [Oromo]

who come-3sg-Past
‘Who came?’
b. Túlluu (GCuf-e)
Tulluu come-3sg-Past
‘Tulluu (came).’

In traditional approaches to question-answer pairs, the wh-phrase in (19a)

and the DP-subject in the response (19b) are regarded as focused elements.
Yet, it is important for our discussion to observe that these elements are not
formally marked for focus in Oromo. Instead, they seem to occur in some
other derived position and therefore fall in the class of non-focused catego-
ries, as described here. A piece of evidence supporting this description is
that the wh-phrase in (19a) and the DP-subject in the response are bare and
therefore lack nominative case morphology (Yiman 1988: 371).
In contrast to the examples (19a–b), the sentences in (20) involve a
case-marked DP-subject in (20a) while its correlated reactions involve a
focus marked wh-phrase in the question and a focused DP-subject in the
answer (20b-c). These sentences imply a context of controversy where Túl-
luu’s arrival is denied, as in (20a). This statement triggers the response in
(20b), which in turn leads to the reaction in (20c). According to Yiman
(1988: 371), such a context requires the expression of contrastive focus,
which is why the wh-phrase and the target constituent in the response are
focus marked with the particle tu. In addition, the absence of case morphol-
ogy on these focused elements could be analyzed as evidence that the fo-
cused constituent does not transit through the canonical subject position on
its way to the peripheral focus position (see section 4 for discussion).

(20) a. Túlluu-n hin-GCuf-n-e [Oromo]

Tullu-Nom Neg-come-neg-Past
‘Tulluu did not come.’
 Focused versus non-focused wh-phrases 301

b. EeĖĖu-tu GCuf-e
who-Foc come-3sg-Past
‘Who is it that came?’
c. Fayyisaa-tu GCuf-e
Fayyisaa -Foc come-3sg-Past
‘It is Fayyisaa who came?’

The examples in (19) and (20) and the distinct replies they are associated
with therefore indicate that Oromo non-focused subject wh-questions occur
in a derived position different from the position in which focused subject
wh-questions occur. Put another way, Oromo subject wh-questions involve
non-focused and focused wh-phrases with each strategy being associated
with a distinct information structure. 8 In wh-questions with non-focused
wh-phrases, the target constituent in the response is not focused (19). In
wh-questions containing a focused wh-phrase (20b), however, the latter co-
occurs with the focus marker tu, which necessitates an answer including a
focused constituent marked by the same marker (20c). I conclude from this
that the two strategies in (19) and (20) relate to two distinct discourse prop-
erties where, in a question-answer pair, a focused wh-phrase requires a
focused constituent in the answer while a non-focused wh-phrase requires a
non-focused constituent.
I further propose that whether such a focused element occurs in the
clausal left periphery or within the VP-periphery is subject to parametric
variation. Recall from section 2 that Aghem resorts to the VP-peripheral
focus position for all types of focus (e.g. new information vs. identifica-
tional). Gungbe, on the other hand, mainly uses the clause left peripheral
area for focus purposes while Italian seems to use both the clause periph-
eral and the VP-peripheral focus positions. Following Brunetti (2003), I
therefore suggest that left peripheral focus and VP-peripheral focus are
identical at the discourse-syntax interface. 9
The following section discusses data on subject-object asymmetry in
question-answer pairs that further support the proposed characterization.

4. Question-answer pairs and information structure

This section analyzes subject-object asymmetry in question-answer pairs

which further supports the idea that the information-structural status of the
302 Enoch Oladé Aboh

wh-phrase in the question (focused vs. non-focused) determines the status

of the target constituent in the response. I start with data from Lele.

4.1. On subject versus object asymmetry in wh-questions

Lele forms wh-questions by means of a wh-phrase and a sentence-final

question marker gà. In subject questions, the wh-phrase must be fronted to
the left of the focus marker ba, as in (21a). The ungrammatical example
(21b) indicates that the wh-phrase cannot remain IP-internal (Frajzyngier
2001: 282).

(21) a. Wéy ba é gà ? [Lele]

who FM go Inter
‘Who went away?’
b.*Wéy é gà ?
who go Inter

Object wh-phrases, however, display both IP-internal and clause peripheral

strategies. The sentence in (22a) illustrates the IP-internal context where the
wh-phrase follows the verb, and the sentence involves the sentence-final
question marker. In the clause peripheral context, however, the object wh-
phrase is fronted to the clause periphery, where it occurs to the left of a
focus marker (22b), see (Frajzyngier 2001: 284/86). 10

(22) a. Mè ày wéy gà ? [Lele]

2sg[F] marry who Inter
‘Who did you marry?’
b. Me ba gol dí gà ?
What FM see 3sg[M] Inter
‘What did he see?’

At this stage of the discussion, a brief digression is in order. Indeed, it is

worth mentioning that the presence of the question particle in Lele suggests
that wh-phrases are not inherently interrogative or focused by definition
(see Aboh & Pfau 2006 for discussion). This would mean that in languages
where there is no question particle of the Lele-type and where wh-phrases
must be fronted, such fronting might as well derive from the necessity of
the wh-phrase to be under the scope of some functional element. In the case
of Lele, the obligatory occurrence of the focus particle in example (22b) is
 Focused versus non-focused wh-phrases 303

strong evidence that fronting in this case is contingent to a clause peripheral

focus requirement. I therefore conclude that wh-phrases move only if they
are forced to, that is, when there is some informational and syntactic reason
for them to front. Put differently, wh-phrases derive their semantics from
their surface syntactic positions, not (exclusively) from their internal struc-
ture. This conclusion is compatible with Lipták’s (2001) proposal that wh-
phrases better qualify as variables (i.e. with no inherent quantificational
meaning) and derive their various quantificational meanings from their
surface syntactic positions as well as the type of quantificational elements
that bind them. In terms of this approach, the description provided here
implies that focused wh-phrases acquire their quantificational meaning by
virtue of being focused syntactically (i.e. moved to an operator position).
On the other hand, non-focused wh-phrases derive their (quantificational)
interpretation from their binder (e.g. an interrogative operator).
With this in mind, I conjecture that there is no question-focus correla-
tion such that in question-answer pairs a wh-phrase will necessarily trigger
an answer containing a focus marked constituent. I therefore propose the
general description in (23) and (24).

(23) The discourse properties of the wh-phrase determine the dis-

course properties of the answer.

(24) A focused wh-phrase (in the higher or lower focus position) re-
quires a focused constituent in the response, but a non-focused
wh-phrase doesn’t.

Empirical evidence from Amharic supports this view (Drubig & Schaffar
2001, Demeke & Meyer 2003). According to Demeke (2003), Amharic, an
SOV language, displays multiple wh-phrase constructions where the wh-
phrases can occur in-situ or ex-situ (in the classical sense). The examples in
(25) are instances of in-situ wh-questions involving arguments as well as
adjuncts.

(25) a. Kassa lä-man mïn sät’t’ä ? [Amharic]

Kassa to-whom what gave
‘What did Kassa give to whom?’
b. Kassa mäce lä-man mïn sät’t’ä ?
Kassa when to-whom what gave
‘What did Kassa give to whom and when?’
304 Enoch Oladé Aboh

With regard to ex-situ wh-questions, Demeke (2003) claims that Amharic

does not exhibit any syntactic constraints (e.g. superiority) on the ordering
of fronted wh-phrases. Instead, this order is mapped onto the informative
structure that the speaker expects from the hearer. The following question-
answer pairs where the (b) patterns are responses to the (a) patterns illus-
trate this correlation.

(26) a. Mäce man mïn gäza ? [Amharic]

When1 who2 what3 bought
b. Tinant Kassa mäs’hafgäza
‘Yesterday1 Kassa2 a book3 bought’

(27) a. Mäce mïn man gäza ?

When1 what3 who2 bought
b. Tinant mäs’haf Kassa gäza
Yesterday1 book3 Kassa2 bought

(28) a. Mïn mäce man gäza ?

What3 when1 who2 bought
b. Mäs’haf tinant Kassa gäza
Book3 yesterday1 Kassa2 bought

According to Demeke, the leftmost wh-phrase is interpreted as the most

prominent (i.e. the focused wh-phrase in our terms). It is still unclear what
the exact syntax of multiple fronted wh-phrases is in Amharic, but it is
worth noting that the sequencing of the wh-phrases in the question deter-
mines rigid ordering of the constituents in the answer. This makes Amharic
a strict discourse-configurational language where information structure and
clause structure work in tandem.
Similarly, the described facts represent strong empirical evidence that
wh-phrases are not inherently specified for discourse properties such as
focus or interrogation. Instead, I suggest that just like ordinary DPs they
constitute potential candidates for establishing focus-presupposition or
question-answer relations at the interface level.
The idea that wh-phrases mainly derive their semantics from surface
structure is compatible with the systematic asymmetry found between sub-
ject and object wh-questions, which I turn to now. Recall from the Lele
examples in (21) and (22) that the subject wh-phrases must be fronted to
 Focused versus non-focused wh-phrases 305

the left of the focus marker while object wh-phrases need not. The latter
may occur in-situ (22a) or ex-situ when focused (22b).
Similar facts are found in Gungbe. As mentioned previously in the ex-
amples in (6), repeated here as (29), an object wh-question (29a) may trig-
ger two possible answers, as in (29b–c).

(29) a. Ét( w(a Kòfí ˜ù ? [Gungbe]

what Foc Kofi eat
‘WHAT did Kofi eat?’
b. É ˜ù l(sì
3sg eat rice
‘He ate rice.’
c. L(sì w(a é ˜ù (bò b( àwútù)
rice Foc 3sg eat and start sickness
‘He ate RICE (and became sick).’

In addition, certain speakers allow for ‘optional’ realization of the focus

marker in object wh-questions, as illustrated in (30).

(30) M(nù %(w(a) Kòfí dà ? [Gungbe]

who Foc Kofi marry
‘WHO did Kofi marry?’

Sentences like (30) led many authors working on the Gbe languages to
assume that the focus marker is optional in object wh-questions. I turn to
this discussion shortly and show that example (30) actually masks two dif-
ferent derivations (see Aboh 2004a for discussion).
What matters for the present discussion is that no such variation arises
in subject wh-questions, and speakers only accept sequences where the wh-
phrase immediately precedes the focus marker, as shown in (31).

(31) M(nù *(w(a) dà Kòfí ? [Gungbe]

who Foc marry Kofi
‘WHO married Kofi?’

Similarly, in subject wh-question-answer pairs (7), repeated here as (32), a

reply to (32a) which does not involve subject focusing is infelicitous (32b).
Instead, the response requires a structure like (32c) where the subject oc-
306 Enoch Oladé Aboh

curs in the left peripheral focus position and immediately precedes the fo-
cus marker.

(32) a. M(nù w(a ˜ù l(sì ? [Gungbe]

who Foc eat rice
‘Who ate rice?’
b.* Kòfí ˜ù l(sì
Kofi eat rice
‘KOFI ate rice.’
c. Kòfí w(a ˜ù l(sì
Kofi Foc eat rice
‘KOFI ate rice.’

These data indicate that there is a structural constraint in Gungbe that pre-
vents subject wh-phrases from exploiting the VP-peripheral focus position
in question-answer pairs. Similarly, the Gungbe subject wh-phrase cannot
occur in the canonical subject position (i.e. [spec TP]). As a consequence,
subject wh-question in Gungbe necessarily involve a focused wh-phrase,
which in turn requires a focused subject DP in the response, hence the infe-
licitous example (32b).
Sabel & Zeller (2002) discuss data on subject wh-questions in Zulu
(Bantu) that mirror the Gungbe facts. Zulu has both in-situ and ex-situ wh-
questions, with ex-situ wh-questions being realized as clefts. These corre-
spond to focused wh-questions in our terms. The sentences in (33) illustrate
these options for object wh-questions.

(33) a. U-bona-ni ? [Zulu, Sabel & Zeller 2002:1, 2]

2sg-see-what9
‘What did you see?’
b. Y-ini o-yi-bona-yo ?
COP-what9 RC2sg-OC9-see-RS
‘What is it that you saw?’

Interestingly, subject wh-questions do not allow the subject DP to appear in

its canonical position (i.e. [spec TP]), hence the ungrammatical example
(34a). Instead, two strategies are available: one that involves a cleft in par-
allel to the object wh-question, as in (34b), and, most crucially, one that
involves a post-verbal position with the canonical subject position being
filled by an expletive (34c).
 Focused versus non-focused wh-phrases 307

(34) a.* Ubani u-fike ? [Zulu]

who1a SP1a-arrived
‘Who arrived?’
b. Ng-ubani o-fikile
COP-who1a RC1a-arrived
‘Who arrived?’
c. Ku-fike bani ?
Expl-arrived who1a
‘Who arrived?’

In terms of the two focus positions proposed in this study (i.e. clause pe-
ripheral and VP-peripheral), Zulu appears to exploit the clause peripheral
and the VP-peripheral focus positions for both subject and objects, while in
Gungbe subjects are restricted to the clause peripheral focus position only
while objects can occur in the clause peripheral position and a derived posi-
tion within IP.
Further, observe from examples (34b–c) that Zulu subject wh-phrases
cannot occur in the canonical subject position (i.e. [spec TP]) but rather in
post-verbal position. The Gungbe subject wh-phrase is also excluded from
the canonical subject positions but is restricted to the clausal left peripheral
position. Accordingly, Gungbe subject wh-questions involve the focus
marker and require a response where the DP-subject is fronted to the left of
the focus marker w(a (32a–c). These facts clearly indicate that subject wh-
phrases are excluded from [spec TP] in both Gungbe and Zulu, although the
two languages circumvent this constraint differently.

4.2. When the EPP-feature competes with the focus feature

Under Sportiche’s (1988) VP-internal subject hypothesis, the ban on sub-

ject wh-phrases in [spec TP] reduces to the competition between two syn-
tactic features: the EPP-feature and the focus feature [F] (Chomsky 1995,
Rizzi 1997, Rizzi and Shlonsky 2005). Indeed, the exclusion of the Gungbe
subject-DP in the VP-peripheral focus position could be explained if we
assume that the subject must move to [spec TP] to check the EPP-feature
under T in this language (Chomsky 1995). This movement, however, can-
not pass through the lower [spec FocP], a freezing position as shown by
Rizzi (1996, 1997, 2004), Haegeman (1995), Rizzi and Shlonsky (2005),
and much related work. Under the assumption that focus positions are crite-
rial positions (Rizzi 2004) where the satisfaction of the focus criterion ter-
308 Enoch Oladé Aboh

minates the focus chain, this would mean that movement of the subject wh-
phrase to the intermediate [spec FocP] would freeze the subject in place.
Consequently, such derivation crashes in Gungbe because the nominative
subject fails to check the EPP-feature under T. As proposed by Rizzi and
Shlonsky (2006), however, [spec TP] is also a freezing position. Accord-
ingly, the subject wh-phrase cannot move to [spec TP] on its way to [spec
FocP] within the left periphery.
I therefore propose that Gungbe subject wh-phrases obligatorily move to
the clause peripheral focus position, where they control (and identify), un-
der c-command, an empty category (e.g. pro) in [spec TP]. This empty
category checks the EPP under T. I further suggest that obligatory control
of the subject wh-phrase in [spec FocP] over the empty category in [spec
TP] in Gungbe correlates with the paucity of agreement morphology as
well as the lack of DP-expletives in Gbe. Under Agree (Chomsky 1999), it
is conceivable that the presence of agreement morphology and a DP-
expletive in Gbe could have allowed for a derivation where the empty cate-
gory is directly licensed in [spec TP], say under spec-head agreement,
while the subject DP moves to the VP-peripheral focus position. This, how-
ever, is impossible, and the only option Gungbe is left with is to displace
the subject wh-phrase to the peripheral focus position, where it controls the
element in [spec TP], as illustrated in (35). Projections irrelevant to this
discussion are ignored.

(35) [FocP M(nù [Foc w(a [.....[TP prom(nù [AspP [FocP [ Foc [VP tm(nù˜ù
l(sì]]]]]]]]

This analysis also holds for the response where the answer of a focused wh-
phrase (i.e. a focused subject DP) cannot remain VP-internal and cannot
move to the intermediate [spec FocP] on its way to [spec TP], where it
checks the EPP under T. As a consequence, the focused DP-subject has to
move to [spec FocP] at the clausal periphery from where it controls the
empty category in [spec TP], as sketched in (36).

(36) [FocP Kòfí [Foc w(a [.....[TP proKòfí [AspP [FocP [ Foc [VP tKòfí ˜ù
l(sì]]]]]]]]

The same argument holds for the Zulu facts where the VP-peripheral focus
position is available for both objects and subjects. Under the assumption
that the subject must move to [spec TP] due to the EPP, movement of the
 Focused versus non-focused wh-phrases 309

subject wh-phrase to the VP-peripheral focus position would force a deriva-

tion where the subject will pass through [spec FocP] on its way to [spec
TP]. Given that [spec FocP] is a freezing position, however, subject-
movement to this position freezes the subject there, and the derivation
would crash unless the EPP-feature under T is checked otherwise. In this
regard, it is interesting to see that Zulu, unlike Gungbe, exhibits an exple-
tive that checks the EPP under T, and the derivation converges. Like most
Bantu languages, Zulu also displays a rich agreement morphology, contrary
to Kwa languages in general.
The EPP constraint and the freezing nature of [spec TP], however, do
not affect object wh-question-answer pairs, and languages consequently
show no asymmetry in this respect (e.g. Lele vs. Zulu, Gungbe vs. Zulu,
and Zulu vs. Italian). As suggested in the Gungbe example (29), an object
wh-question may lead to two structurally different responses where the
object DP either stays in-situ or fronts to the clausal left periphery, where it
precedes the focus marker (29b–c). This would mean that object wh-
phrases have access to both the VP-peripheral and the clause-peripheral
focus positions. Considering this, I propose that the optionality in (30) is
only apparent. I suggest that this example masks two different construc-
tions, as illustrated in (37a) and (37c) with their corresponding responses in
(37b) and (37d) respectively.

(37) a. %M(nù Kòfí dà ? [Gungbe]

who Kofi marry
‘Who did Kofi marry?’
b. Kòfí dà Màrí.
Kofi marry Mary
‘Kofi married Mary.’
c. M(nù w(a Kòfí dà ?
who Foc Kofi marry
‘Who is it that Kofi married?’
d. Màrí *(w(a) Kòfí dà.
Mary Foc Kofi marry
‘Kofi married MARY.’

Example (37a) corresponds to a situation where the set of possible girls that
Kòfí can marry is unknown and the speaker is looking for this new informa-
tion as part of the topic of discussion, hence the response in (37b). Sentence
(37c), on the other hand, corresponds to a situation where there is a pre-
310 Enoch Oladé Aboh

existing set of potential girls that Kòfí can marry and the speaker wants to
know who Kòfí marries eventually. The response here must contain a fo-
cused object as in (37d). This suggests that even though the wh-phrase
fronts in (37a) and (37c), it does not target the same positions, and it does
not trigger the same information structure in the responses. This is shown
by the contrast in (37b) and (37d), which leads me to conclude that there is
a projection FP within the clausal left periphery that is lower than FocP but
higher than TP and whose specifier [spec FP] may host non-focused wh-
phrases, as illustrated in (38).

(38) [FocP [Foc [FPM(nù.....[TP Kòfí [AspP [FocP [ Foc [VP tKòfí dà tm(nù]]]]]]]]

This analysis is compatible with the conclusion we reached while discuss-

ing the Italian data in section 3 that wh-phrases do not always move to the
focus position. Instead, it appeared from the discussion that in certain con-
texts wh-phrases occur in a derived position between the focus position and
their first merge position. I further assume that in the response of such
Gungbe wh-questions the target constituent also appears in a derived posi-
tion that is not a designated focus position.
With regard to the focused object wh-phrases, I propose that they neces-
sitate focused object DPs in the response, where the focused object is in the
clause peripheral focus position and precedes the focus marker. This is
illustrated in (39).

(39) a. [FocP M(nù [Foc w(a [FP tm(nù [TP Kòfí [AspP [FocP [ Foc [VP tKòfí
dà tm(nù]]]]]]]]
b. [FocP Màrí [Foc w(a [FP tMàrí [TP Kòfí [AspP [FocP [ Foc [VP tKòfí dà
tMàrí]]]]]]]]

Extending this discussion to the question of wh-movement versus wh-in-

situ languages, the proposed analysis here indicates that, in languages
which resort to the VP-peripheral position only (e.g. Aghem), the so-called
in-situ position does not correspond to a VP-internal argument position.
Instead, what superficially appears to be a first merge position is actually
the designated VP-peripheral focus position where focused constituents
occur in these languages.
Granted this is the right characterization, it seems reasonable to propose
that the constraints on accessing this lower focus position can only be lan-
guage specific. Note, for instance, that while the Gungbe focused subject
 Focused versus non-focused wh-phrases 311

DPs or wh-phrases cannot target this position, no such problem arises in

Aghem, Zulu, or in Italian where the languages involve expletive or pro-
nominal elements that can fill the canonical subject position (i.e. [spec
TP]), allowing the focused subject to move to the intermediate VP-
peripheral focus position.

5. Conclusion

This paper shows that languages distinguish between focused and non-
focused wh-phrases, which target different positions within the clause. The
discussion of the distribution of focused wh-phrases supports recent re-
search within the cartography framework on the existence of two focus
projections within the clause structure, that is, within the clause periphery
and the VP-periphery. However, cross-linguistic data suggest that there is
no semantic (or pragmatic) distinction between these two focus positions.
Instead, it is argued, that in question-answer pairs, the position of the wh-
phrase within the question determines that of the target DP in the response.
This implies that the surface position of the target constituent in the re-
sponse depends on that of the wh-phrase in the question.
In this regard, it has been shown that a focused wh-question requires a
focused constituent in a focus position while a non-focused wh-phrase re-
quires a non-focused constituent. Data from languages with focus markers
show that, in focused wh-phrase questions, the wh-phrase occurs with a
focus marker which also marks the target constituent in the response. In
non-focused wh-questions, however, the wh-phrase occurs in a derived
position that also conditions the surface position of the target constituent in
the response.
I conclude from this that there is no systematic correlation such that in
question-answer pairs a wh-question will necessarily require a response
including a focused constituent.

Notes

1. Small caps in the translations indicate focused constituents.

2. See Aboh (2004a) and references cited there for the discussion on focus
constructions and wh-questions across Kwa.
3. Examples are reproduced as they appear in the cited sources.
312 Enoch Oladé Aboh

4. It is crucial to note that the Gungbe focus marker and the Aghem focus
marker are not isomorphic. For instance the focus marker does not seem to
co-occur with wh-phrases in Aghem, unlike in Gungbe.
5. Other Bantu languages (e.g. Zulu) are also relevant here (see Sabel &
Zeller 2002).
6. See also Manfredi (1997) for further discussions of VO vs. OV in Kwa.
7. See also Lipták (2001) for arguing that wh-phrases are not inherently
quantificational. Similarly, Boskovic’s (2002) work on multiple wh-
fronting languages indicates that wh-phrases undergo movement in these
languages depending on whether they are focused or not, suggesting that
wh-questions in those languages may involve focused versus non-focused
instances.
8. See section 4 for a similar conclusion concerning object wh-questions in
Gungbe. Put together, these data suggest that subject and object wh-
phrases show parallel distribution cross-linguistically even though they
might behave differently language-internally.
9. See Lipták (2001) for similar conclusion for English in-situ focus.
10. Adjunct focusing also displays these two strategies. Whether the Lele data
underscore the VP-peripheral and clause peripheral focus positions is not
clear for the time being and more study is needed before we reach a pre-
cise characterization of these facts.

References

Aboh, Enoch Oladé

2004a The morphosyntax of complement-head sequences. Clause structure
and word order patterns in Kwa: Oxford studies in comparative syn-
tax. Oxford, New York: Oxford University Press.
2004b Topic and Focus within D. Linguistics in The Netherlands 21: 1–12.
Aboh, Enoch Oladé, and Roland Pfau
2006 What’s a wh-word got to do with it? Ms. University of Amsterdam.
Belletti, Adriana
2002 Aspects of the low IP area. In The Structure of IP and CP. The Car-
tography of Syntactic Structures 2, Luigi Rizzi (ed.), New York: Ox-
ford University Press.
Biloa, Edmond
1997 Functional Categories and the Syntax of Focus in Tuki. Lincom
Studies in African Linguistics 02. Newcastle: Lincom Europa.
Boškoviü, Željko
2002 On Multiple Wh-fronting. Linguistic Inquiry 33: 351–383.
 Focused versus non-focused wh-questions 313

Brunetti, Lisa
2003 A Unification of Focus. Ph.D. dissertation, Università di Firenze.
Cheng, Lisa Lai-Shen
1991 On the Typology of Wh-questions. Ph.D. dissertation, MIT.
Chomsky, Noam
1995 The minimalist Program. Cambridge, Mass.: MIT Press.
1999 Derivation by phase. MIT Occasional Papers in Linguistics 18. De-
partment of Linguistics, MIT.
Demeke, Girma Awgichew
2003 Wh-questions in Amharic. Ms. University of Tromsoe.
Demeke, Girma Awgichew, and Ronny Meyer
2003 Information Structuring in Amharic. Ms. University of Tromsoe.
Drubig, Hans Bernhard
2003 Toward a Typology of Focus and Focus Constructions. Linguistics
41: 1–50.
Drubig, Hans Bernhard, and Wolfram Schaffar
2001 Focus Constructions. In Language Typology and Language Univer-
sals, Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and
Wolfgang Raible (eds.), Berlin: Mouton.
Downing, Laura, Al Mtenje, and Bernd Pompino-Marschall
2004 Prosody and information structure in Chichewa. ZAS Papers in Lin-
guistics 37: 167–186.
É. Kiss, Katalin
1998 Identificational Focus versus Information Focus. Language 74: 245–
273.
Frajzyngier, Zygmunt
2001 A Grammar of Lele. Stanford Monographs in African Languages,
Stanford, California: CSLI Publications.
Good, Jeff
2005 A fixed-position focus construction in Naki. Paper presented at the
Conference on Focus in African languages, ZAS, Berlin, October
2005.
Haegeman, Liliane
1995 The Syntax of Negation. Cambridge: Cambridge University Press.
Hovarth, Julia
1986 Focus in the Theory of Grammar and the Syntax of Hungarian,
Dordrecht: Foris.
Hyman, Larry M. (ed.)
1979 Aghem grammatical structure. Southern California Occasional Papers
in Linguistics 7. Los Angeles, University Southern California.
314 Enoch Oladé Aboh

Hyman, Larry M.
2005 Focus Marking in Aghem: Syntax or Semantics? Paper presented at
the Conference on Focus in African languages, ZAS, Berlin, October
2005.
Hyman, Larry M., and John R. Watters
1984 Auxiliary Focus. Studies in African Linguistics 15: 233–273.
Lipták, Anikó Klára
2001 On the Syntax of Wh-items in Hungarian. LOT dissertation series. 45
Manfredi, Victor
1997 Aspectual Licensing and Object Shift. In Object Position in Benue-
Kwa, Rose-Marie Déchaine and Victor Manfredi (eds.), 87–122. The
Hague: The Holland Academic Graphics.
Rizzi, Luigi
1996 Residual Verb Second and the wh-criterion. In Parameters and
Functional Heads: Essays in Comparative Syntax, Adriana Belletti
and Luigi Rizzi (eds.), 63–90. New York: Oxford University Press.
1997 The Fine Structure of the Left Periphery. In Elements of Grammar,
Liliane Haegeman (ed.), 281–337. Dordrecht: Kluwer.
2001 On the position ‘Int(errogative)’ in the left periphery of the clause. In
Current Studies in Italian Syntax: Essays Offered to Lorenzo Renzi,
Guglielmo Cinque and Giampaolo Salvi (eds.), 287–296. New York:
Elsevier.
2004 Locality and Left Periphery. Ms. University of Siena.
Rizzi, Luigi, and Ur Shlonsky
2005 Strategies of Subject Extraction. Ms. University of Siena.
Sabel, Joachim, and Jochen Zeller
2002 Wh-question formation in Nguni. To appear in Oxford Working Pa-
pers in Linguistics. Oxford.
Sportiche, Dominique
1988 A Theory of Floating Quantifiers and its Corollaries for Constituent
Structure. Linguistic Inquiry 19: 425–449.
Yiman, Baye
1988 Focus in Oromo. Studies in African Linguistics 19: 365–384.
Watters, John R.
1979 Focus in Aghem: a study of its formal correlates and typology. In
Aghem grammatical structure, Larry Hyman (ed.), Southern Califor-
nia Occasional Papers in Linguistics 7: 137–197.
Zubizarreta, Maria Luisa
1998 Topic, Focus and Word Order. Cambridge, Mass.: MIT Press.
List of contributors

Enoch Oladé Aboh Katharina Hartmann

General Linguistics, Lan- Humboldt Universität zu Ber-
guages and Literature lin
Faculty of Humanities Institut für deutsche Sprache
University of Amsterdam und Linguistik
Spuistraat 210 Unter den Linden 6 / Location:
NL - 1012 VT Amsterdam Dorotheenstr. 24
[email protected] D - 10099 Berlin
[email protected]
Ines Fiedler
Humboldt University of Berlin Victor Manfredi
SFB 632 Information Structu- African Studies Center
re Boston University
Unter den Linden 6 / Location: 270 Bay State Road
Mohrenstr. 40-41 Boston MA 02215
D - 10099 Berlin U.S.A.
[email protected] [email protected]

Mara Frascarelli Lutz Marten

Università degli Studi Roma 3 Department of African Lan-
Dipartimento di Linguistica guages and Cultures
Via Ostiense, 236 School of Oriental and African
I - 00146 Roma Studies
[email protected] Thornhaugh Street
Russell Square
Tom Güldemann UK - London WC1H 0XG
Seminar für Allgemeine [email protected]
Sprachwissenschaft
Universität Zürich Annarita Puglielli
Plattenstrasse 54 Università degli Studi Roma 3
CH - 8032 Zürich Dipartimento di Linguistica
[email protected] Via Ostiense, 236
I - 00146 Roma
[email protected]
316 List of contributors

Brigitte Reineke Florian Schwarz

Eichenring 4a Department of Linguistics
D - 16341 Panketal 226 South College
[email protected] University of Massachusetts
berlin.de Amherst, MA 01003
U.S.A.
Chris Reintges [email protected]
PO box 9515
NL - 2300 RA Leiden Sabine Zerbian
[email protected]. Department of Linguistics
nl University of the Witwaters-
rand
Anne Schwarz Private Bag 3
Humboldt University of Berlin Wits 2050
SFB 632 Information Structu- South Africa
re [email protected]
Unter den Linden 6 / Location:
Mohrenstr. 40-41 Malte Zimmermann
D - 10099 Berlin Universität Potsdam
[email protected] Institut für Linguistik
berlin.de Karl-Liebknecht-Str. 24-25
D - 14476 Golm
[email protected]
Subject index

accent 2; 4; 7; 15–19; 21; 24–32; relative 7; 8; 121; 142; 146;

35–51; 56; 67–68, 74; 76; 79; 150; 152–153; 164–166; 168;
134; 147; 260 174–175; 180–181; 183; 185–
pitch 2; 4; 7; 47; 49–50; 56; 67; 186; 188; 197–198; 203; 212–
74; 147 213; 229; 238; 267–268; 277
*Adjunction 113; 115; 119–120; sequential 238; 268; 279
123–132 small clause (SC) 164
Late *Adjunction 120; 123–132 clause-typing 187; 203
afterthought construction 124; 127 cleft 7; 8; 139–140; 147; 149–158;
Agree 1; 103; 161; 168; 171; 173; 199–200; 262; 277; 282; 285;
175–176; 178; 206; 308 290; 306
agreement 11; 19; 42; 47; 115; 118; clitic 15; 19; 24–31; 35; 38–39; 42;
123; 132; 134–135; 162; 165– 49; 97; 107; 114–115, 123–
167; 180–181; 183–184; 190– 132; 144; 157; 162–164; 167;
191; 193; 204; 207; 209; 213; 169; 178–182–183; 193; 198;
216; 220; 224; 242; 245; 268; 283
271; 308; 309 locative subject clitic 131
anaphoric 115 subject clitic 24–25; 35; 38;
anti-agreement 180–181; 184 123–128; 131–132; 163; 179
wh-agreement 190–191; 204; complementizer 157; 162; 188–
213; 216; 220 200; 206; 209; 212–214; 294
assertion marker 151–152; 155 relative 188; 195; 197; 212–214
asymmetry 79; 126; 167–169; 176; conjoint (form) 31–35; 43; 132
182; 224; 226; 229; 256; 268– conjunction 39; 44; 121; 271–284
269; 272; 292; 295; 301; 302; construction
304; 309 declarative focus 186; 199–200;
subject vs. non-subject 224; 203
229; 268–269; 272 extra–clausal 279
subject vs. object 79; 301–304 focus, see focus
auxiliary periphrasis 84; 99; 102 intra-clausal 280; 282
interrogative focus 202
C-domain 161–162; 169–170; 172; narrative 7; 192; 202; 211; 267–
174–179 268; 273–286
Clause relative 213; 229
declarative 185; 187; 190 wh- 139–141; 150–154; 190;
embedded 142; 150; 152–153; 192; 217
175; 179; 197; 199; 206; 248; copula 36; 139–142; 146; 149;
299 150–151; 155–158; 164; 166;
narrative 267–268; 273; 276– 189–190; 214; 217; 234; 261–
280; 284 262; 277–278; 280–286
neutral 4; 186 defocusing 103–104
318 Subject index

discourse 1; 4–8; 11; 34; 57–61; focus construction 7; 11; 47;

73–76; 97–98; 110; 113–115; 101; 107; 113; 122; 127; 131;
123–124; 127; 130; 134; 156; 139; 143–157; 164; 166–170;
161–162; 165–166; 172–174; 174; 180–188; 199–203; 217–
176–179; 183; 192–193; 201– 218; 223–226; 230; 232; 235;
202; 217; 223; 233; 237; 260; 237; 239; 262; 267–285; 291;
262; 282; 286; 292; 301–304 295; 311; 313
categories 162; 166; 174 focus feature 22; 144; 146; 149;
discourse-pragmatic function 211; 245; 287–288; 307
113–115 focus fronting, see focus
disjoint (form) 29; 31–35 movement
duration 18; 64; 68–70; 73; 76 focus marker 2–8; 33; 35; 55–
Dynamic Syntax 113–117; 122– 56; 74–75; 108; 137; 140; 144;
135 148–151; 156; 159; 163–164;
179; 183; 189–90; 221–259;
EPP-feature 307–309
269; 277–279; 282; 285; 287–
exhaustivity 6; 86; 189; 201; 203;
288; 292; 294; 296–297; 300–
226; 241; 244; 246; 251–261;
314
292–293
affirmative 237; 239
marker 6 ; 244 ; 246 ; 251–253;
negative 239
256–261
focus movement 5; 15; 154;
expletive pronoun 278; 282
159; 195; 242; 245; 251; 290
focus particle 184; 257; 291;
focus
296; 302
assertive 30; 95; 100–101; 104;
focus phrase 139–151; 155–
109
157; 181; 185; 189; 194; 200;
ambiguity 15; 21; 227; 230
208; 212; 245; 287–288; 290;
auxiliary 32; 46; 98; 108; 218;
295; 299
314
focus position 6; 205; 287;
contrastive 6; 95; 100; 129;
292–312
172; 185; 188; 202–203; 293–
focus prominence 20; 37; 166;
296; 298; 300
171; 193
focus copula 189–190
focus sensitivity 18; 25; 37;
exhaustive listing 201; 189;
203; 241; 244; 246; 250–251;
201; 203
257–259
ex situ 7; 15; 139–140; 143–
focused wh-question 7; 175;
155; 229; 232; 237; 239; 242–
287–288; 299–311
245; 248; 250; 259–260; 268–
identificational 11; 114; 122;
269; 276–277; 281–283; 292
124; 126; 158; 226; 239; 261;
focus alternatives 252; 257
291; 292; 298; 313
focus constituent 4; 5; 192; 195;
inherent 98–99; 109
225; 241–242; 246; 248; 250–
251; 255; 258–259; 276; 280–
283; 298
 Subject index 319

in situ 6; 7; 55–56; 60; 71; 74– grammaticalization 77; 83–86; 92;

75; 101; 147; 149; 155; 202; 96; 100; 103–108; 223; 268;
217; 223–224; 236; 238; 241– 278; 280–281; 284–285
250; 260; 262; 291–292; 312 implicature 241; 246; 251–252;
multiple 31; 47; 129; 195 257–262
new information 7; 56; 129; conventional 241; 246; 251–
200–201;226; 252; 257–258; 262
non-focused wh-question 287– inflection 22; 43; 87; 91–92; 96;
288; 299; 301; 303; 310; 311 167; 185; 188–192; 197; 204;
non-subject focus 224–229; 209–211; 214; 217; 219
268–270; 273; 277; 282 special 185; 188–192; 197; 204;
object focus 63; 64; 67; 70; 209–211 ; 214–215; 217; 219
228; 244; 247; 250; 295 inflectional morphology 141;
presentational 114; 122; 126; 185–186; 189–190; 215
131; 201; 214 information structure 1–8; 11–12;
sentence focus 147–149; 157; 15; 35; 51–52; 55; 63; 70–83;
230; 247 92–94; 97–101; 104; 114; 120;
subject focus 23; 38; 39; 149; 122; 127–135; 156; 161–162;
214; 224; 226; 229–230; 244– 172; 176; 183; 213; 226; 232;
246; 259; 269; 295; 305 239; 259; 262; 279; 283; 285;
verbal focus 224; 230; 235 287–288; 291–292; 298; 301;
verb phrase focus 148 304; 310; 313
verum focus 148–149 intonation 2–7; 12; 28; 36–37; 42;
focus-marking 75; 227; 230; 285 45; 47–61; 64; 67; 76–79; 142;
lack/absence of 7; 55–56; 74; 171; 262
129; 202; 242; 246
Lexical Functional Grammar (LFG)
obligatory 37–38
114; 127
non-obligatory 75
lexical information 117–118
optional 256; 259
LINK 113; 115; 119–121; 127–
prosodic 4; 55–56
128; 130
morphological 22
Logical Form (LF) 115–116; 119;
syntactic 8; 75
127; 133; 174; 207; 211; 216–
force 161–162; 166–179; 193;
217; 242
202–203; 208; 297
illocutionary 161; 168; 170; morpho-syntactic flagging 188;
174–175; 178–181 210–212
interrogative 162; 169–170; movement 3; 5;15; 41; 47; 105;
173; 177; 202–203 109–110; 133; 144–145; 149–
Force-Fin system 161–163; 166; 150; 154–155; 157; 159;
169 169178; 182; 185–186; 188–
fronting, see movement 191; 195; 198–200; 204–220;
f-structure 115 242; 245; 251; 277; 282; 290;
fundamental frequency 64; 73–76 298; 302–303; 307– 312
covert 188; 207; 211; 215
Government Phonology 133 long distance 215
320 Subject index

wh-movement 159; 185; 190– Phrasing 3; 12; 18; 20; 25; 31–33;
191; 195; 204–212; 215–216; 35; 37; 39; 52; 55–60; 64; 73;
218; 220; 310 78; 115; 214; 225; 262
Phonological 115
negation 38; 47; 83; 88; 91; 97; 99; Syntactic 20; 25
102; 106–108; 110; 142–143; postverbal position 99; 131; 270;
146; 150; 152; 156; 184; 217; 283
228; 231; 233; 238–239; 259; postverbal subject 126; 132
313 predicate-argument structure 125
pronoun 11; 42; 65; 83; 87–88; 96–
object 98; 103; 114–115; 121–123;
definite 98 127; 132; 134–135; 162–165;
object clitic 24–26; 30–31; 39; 169–170; 172; 186; 189; 191;
97; 114–115; 129–132; 179 194; 197–198; 218; 224; 227;
object shift 84; 92; 109; 314 229; 233–234; 236; 238; 259;
preverbal 83–86; 88–93; 96– 269–273; 277–279; 282–283
102 incorporated 114–115; 132
pronominal 87–88; 91; 104; proposition 94; 115–117; 126–128;
224; 270; 277 131; 133; 161; 166; 201; 223;
operator 20; 25; 33; 36; 92; 98– 233; 236–239; 257; 296
102; 152; 164–166; 169–176; prosody 2; 4; 7; 12–16; 18; 21; 25;
180–181; 185– 186; 188–193; 29; 35; 39–40; 47; 52–54
198; 207–212; 216; 257; 291; prosodic information 115
303 prosodic phrase boundary 3;
operator-variable dependency 248
185; 188; 190; 192; 209–210; prosodic prominence 56
212
out-of-focus part 1; 225; 229; 232; question
238; 269–273; 276; 284–285 constituent question 72; 185–
186; 225
question–answer pair 5; 71;
perception 6; 41–44; 46; 55–56;
126; 242–243; 253–254; 287–
60; 67; 71–74; 78; 219
288; 291; 293; 298; 300–306;
periphery
309; 311
left 51–52; 114; 128; 131; 135;
question marker 170; 179; 302
157; 159; 161–162; 169; 179;
wh-question 7–8; 153–154;
184; 190; 193–195; 198; 204;
171–174; 182; 186–189–190;
206; 220; 232; 237; 288; 294;
200; 212; 216; 227; 258; 260;
297–298; 301; 308–310; 314
287; 291; 293; 296; 298; 301–
right 113–114; 119; 122; 126;
314
128; 133; 247
wh-in situ question 188;
VP-periphery 288; 292; 294–
199200; 206–207; 211–212;
295; 297; 301; 311
214
 Subject index 321

yes/no-question 58; 170–171; topic 1–2; 11; 15; 27; 37; 40–42;
173; 176; 192; 199; 202–203; 48; 50; 55; 75; 79; 91; 101;
213; 250; 260 104; 113–115; 121–134; 145–
146; 152–153; 156–157; 162;
Relevance Theory 133 164; 167; 169–170; 176–184;
scope 27; 29; 31; 38; 44; 47; 164; 188; 193–195; 206; 225; 230;
170–174; 181; 185; 200; 204– 234; 238; 259; 268; 282–283;
214; 227; 232; 239; 236; 302 286–297; 309; 312; 314
semantic composition 116 background 114; 127
semantic representation 113–116; clitic-resumed 169
127; 131; 133 topicalization 146; 153–154; 156;
stress 2–4; 11–56; 64; 70–71; 76; 177–179; 182; 213
134; 293 topic-focus field 162; 206
nuclear 15; 18; 22–23; 28–32;
36; 38; 40; 49 underspecification 41; 123; 135
structural 123
TAM (tense-aspect-mood) 83; 96;
186–214 verb
TAM-marker 186–189; 195; non–finite 89; 90; 98; 102; 103;
197–198; 202; 204; 206; 210; 105
212 serial 44; 84; 88; 107; 109
Relative TAM 187; 188; 195; verbal morphology 22; 225;
106; 197–214 238; 270; 274
templatic morphology 15
tense 20–27; 33; 37–39; 50; 88; wh-extraction, see movement
93–94; 96; 105–106; 118; 132– wh-phrase 139; 172–173; 175; 186;
133; 141; 143; 149; 155–158; 189; 190–195; 200; 204–215;
162; 167; 179; 185; 188; 193; 287–312
196–199; 202–205; 208–214; wh-word 37; 120; 154–155; 204–
218–219; 232–239; 245; 267; 205; 312
270–271; 274; 276; 279; 281– word order 2–5; 8; 18; 29; 31; 44;
282; 285; 298 54; 81–85; 87; 96; 100–110;
relative 185; 187–188; 193; 196; 113– 116; 119; 122; 128; 144;
198; 202–209; 212; 267; 270; 151; 187; 193; 207; 215; 224;
276; 279 242; 245; 284; 312; 314
tone 3–4; 7; 12; 16–21; 24–30; 33– VS order 23; 38; 122–127; 132;
34; 36–38; 40–67; 74; 77–79; 240
129; 132–134; 156; 158; 224; word order alternation 84–85;
226; 242; 245; 248; 259; 262; 100; 102
269–276; 279; 281–285
floating 17; 21
polar 245; 248
tonal change 150; 269
tonal morphology 42
verb tone 269–275; 279
Language index

Afar 161–164; 166; 171–172; 174; Idoma 83; 86; 89; 93; 96; 105
177–182; 184 Igbo 86; 89; 93; 96; 99; 103; 107
Aghem 19; 46; 53; 86; 94; 96–98; (Avu) Igbo 93; 96
110; 295–298; 301; 310–314
Akan 6; 11; 268–271; 274; 276; Konni 284
278–281; 284–285 Kaje 86; 88; 97; 103
Amharic 303; 304–313
Kana (Khana) 86–88; 97; 103; 108
Kikuria (Kuria) 24
Bafut 86; 89; 97; 106; 110 Kikuyu (Gikuyu) 6; 42–43; 139;
Buli 6; 267; 271–276; 279–284;
148; 154; 156; 158–159
286
Kimatuumbi (Matumbi) 4; 16; 20–
Byali 6–7; 223–240
25; 28–31; 37–38; 40; 50–51
Kinande (Nande) 56–57
Chamorro 190; 204; 216
Kinga 24
(Nkhotakota) Chichewa (Nyania)
Kinyarwanda 33–34; 47
3; 11; 42; 56–57; 59; 67–68;
Kirundi (Rundi) 32–35; 49
73; 77–78; 114–115; 126–129;
133–134; 262; 313
Chitumbuka 295 Leggbo (Legbo) 32–35; 49
(Egyptian) Coptic 6–7; 185–188; Lele 302; 304; 309; 312–313
193–219 Lelemi 6; 268; 270; 274; 276; 278;
280–282; 284
Dagbani 6; 267; 272; 275–276; Luhaya 4; 29–31
279–281; 283; 286
Ditammari 224; 240; 339 Makua 24; 38; 42; 52
Mambila 43; 86; 94–98; 100; 110
Ewe 6; 11; 268–269; 273; 276; Mandinka (Manding) 36; 41; 240
278; 280–281; 284–286
Ewondo 86; 88; 97; 103; 110 Naki 86; 94; 97–98; 107; 313
Nateni 224
Fon 284–285 Nen (Njen) 86; 94–98; 100; 109
Fulfulde 267 Northern Sotho 4; 6; 54–64; 67;
73–79
Gungbe 3; 5; 7; 289–298; 301; Nsenga 113; 122–123; 127; 132–
305–310; 312 133
Nupe 86; 89; 92–93; 96; 99; 102–
Hausa 5–7; 12; 74–75; 77; 186– 103; 107-108
193; 202; 204; 210–211; 214;
217–220; 241–263; 267; 277; Oromo 300; 301; 314
285
Ibibio 86–87; 96 Proto-Bantu 86; 97; 103
324 Language index

Safwa 24; 53 Vute 86–88; 97; 110

Setswana (Tswana) 33–35; 39; 42–
43; 47; 53; 77 Waama 224
Somali 6; 7; 161–184
Swahili 18; 47; 113; 123; 134 Xhosa 56–58; 73; 78–79; 115; 134
Swati 130–132
Yoruba 86; 90; 93; 96; 99; 102
Tikar 86; 91; 97; 99; 110
Tumbuka 113; 128–130
Zulu 295; 298; 306–312
Umbundu 16; 51