Word Order Change (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated] 9780198747307, 0198747306
This volume explores word order change within the framework of diachronic generative syntax. Word order is at the core o
140 105 4MB
English Pages 352 [345] Year 2018
Cover
Word Order Change
Copyright
Contents
Series preface
List of figures and tables
List of abbreviations
Notes on contributors
1: Word order change from a diachronic generative syntax perspective
1.1 Introducing the topic
1.2 Clause structure, syntactic movement, and word order change
1.3 Book structure and chapter outline
1.3.1 Part I—targets for movement
1.3.2 Part II—triggers for movement
1.3.3 Verb movement into the left peripheries
1.3.4 Types of movement and constraints on it
1.4 Closing up
Acknowledgements
Part I: Targets for movement: Changes in the functional architecture of the clause
2: Configurational change in Indo-European coordinate constructions
2.1 Introduction
2.2 From coordination to junction
2.2.1 Background assumption
2.2.2 An enriched structure
2.2.2.1 Den Dikken’s J(unction)
2.2.2.2 Lexicalised J: Avar
2.3 Indo-European
2.3.1 Alternation in linear placement
2.3.2 Morphemicity
2.4 Deriving peninitial placement
2.5 Summary and outlook
Acknowledgements
List of textual sources
3: Discontinuous noun phrases and remnant-internal relativization in the diachrony of Portuguese
3.1 Introduction
3.2 Discontinuous noun phrases in Latin and ESP
3.3 Remnant-internal relativization in ESP
3.3.1 The definition of remnant-internal relativization
3.3.2 Core properties of remnant-internal relativization
3.3.3 The interpretation of remnant-internal relativization
3.4 The syntax of remnant-internal relativization: competing analyses
3.4.1 The syntax of RRCs
3.4.2 The syntax of discontinuous noun phrases
3.4.3 Towards an analysis of remnant-internal relativization
3.5 The proposal
3.5.1 Focus and prosody
3.5.2 The copy theory of movement
3.5.2.1 Bošković and Nunes ()
3.5.2.2 Stjepanović ()
3.5.3 The derivation of remnant-internal relativization
3.5.3.1 Pattern I
3.5.3.2 Pattern II
3.5.3.3 Pattern III
3.6 ESP and CEP in contrast
3.7 Conclusion
Primary Sources
Corpora
Editions
4: The relative cycle in Hungarian declaratives
4.1 Introduction
4.2 The data
4.3 Grammaticalization and the relative cycle
4.4 Simplex complementizers
4.5 Multiple complementizers
4.6 Complex complementizers
4.7 The position of hogy
4.8 Relative clauses
4.9 Further combinations
4.10 Functions of hogy
4.11 Conclusion
Acknowledgements
Textual sources
5: Word order change at the left periphery of the Hungarian noun phrase
5.1 Introduction
5.2 The noun phrase in Modern Hungarian
5.2.1 Demonstratives
5.2.2 Possessives
5.3 The noun phrase in Old Hungarian
5.3.1 Demonstratives and articles
5.3.2 Possessives
5.4 Middle Hungarian phenomena
5.4.1 Demonstratives
5.4.2 Possessives
5.5 Summary
Acknowledgements
Sources
Part II: Triggers for movement: Changes in nature or stability
6: Particle-verb order in Old Hungarian and complex predicates
6.1 Introduction
6.2 On Modern Hungarian word order
6.3 Word order change and consequences
6.3.1 SOV vs SVO in Old Hungarian
6.3.2 The Germanic parallels
6.4 Particle-verb constructions in Old Hungarian
6.4.1 The class of particles
6.4.2 The distribution of particles
6.4.2.1 Neutral sentences
6.4.2.2 Non-neutral sentences
6.5 The syntactic properties of Old Hungarian PRT-V constructions
6.5.1 The syntax of particles
6.5.2 Complex predicates
6.6 Extending predicate movement: other predicative elements
6.7 Conclusions
Acknowledgements
Old Hungarian sources
7: An effect of residual T-to-C movement in varieties of English
7.1 Introduction
7.2 Discontinuous subjects in Older Scots and Appalachian English
7.2.1 Older Scots discontinuous subjects as transitive expletives
7.2.2 Discontinuous subjects in Appalachian English
7.3 Verb second in Older Scots
7.4 Negative auxiliary inversion in Appalachian English
7.5 Conclusion
Acknowledgements
Primary Sources
8: Word order and information structure in the Würzburg Glosses
8.1 Introduction
8.2 The Bergin kerfuffle
8.3 Methodology
8.4 The data
8.4.1 Intí-sentences
8.4.2 Invisible clefts
8.4.3 Fronting + cleft
8.5 Summary and syntax
8.5.1 Clefts and focus
8.5.2 Final ordering
8.6 Diachronic perspectives
8.7 Conclusions
Acknowledgements
Sources
Part III: Verb movement into the left peripheries
9: Subject inversion in transitive sentences from Classical to Modern European Portuguese A corpus-based study
9.1 Introduction
9.2 The quantitative change from ClP to EP
9.3 The qualitative change
9.4 The nature of the change
9.4.1 There is a high position available for post-verbal subjects in ClP
9.4.1.1 Two different positions for post-verbal subjects
9.4.1.2 Other pieces of evidence
9.4.1.3 The derivation of VOS
9.4.2 Evidence that the position of the subject, and consequently of the verb, is diachronically reanalysed
9.4.2.1 The sharp increase in the frequency of SVO
9.4.2.2 The emergence of the XSV-cl order
9.4.2.3 The high frequency of pronouns in VSO sentences
9.4.2.4 The loss of the matrix-subordinate asymmetry
9.5 Concluding remarks
Acknowledgements
Source
10: Analyticization and the syntax of the synthetic residue
10.1 Introduction
10.2 The formal syntax of analyticization
10.3 Deriving verb movement options
10.4 Verb movement and A(rgument)-syntax in a highly analytic system
10.4.1 Tense-rich analyticity and the cartography of the Coptic clause
10.4.2 Verb movement of synthetically inflected infinitives and statives
10.5 Analyticization as an emergent property in Brazilian Portuguese
10.5.1 The temporal system of Brazilian Portuguese and the Richness of Tense question
10.5.2 Tense-to-Aspect reanalysis and the loss of Verb-to-Tense1 movement
10.6 A ‘hidden’ word order change: on the comparative syntax of the adverbial sempre in European and Brazilian Portuguese
10.6.1 The aspectual and modal uses of sempre in European Portuguese
10.6.2 The exclusively aspectual use of sempre in Brazilian Portuguese
10.7 Concluding remarks
Sources
11: Loss of laten-support in embedded infinitivals in fifteenth-century Low Saxon
11.1 Introduction
11.2 Infinitival verb fronting
11.3 Infinitival fronting in fifteenth-century Low Saxon
11.4 Laten-support
11.5 IV2 as a subjunctive construction
11.6 Decline of the finite subjunctive
11.7 The ende+infinitive as T-to-C
11.8 The loss of a separate class of auxiliaries
11.9 T-to-C in the history of Dutch
11.10 Conclusions
12: The distribution of quantifiers in Old and Modern Italian Everything or nothing
12.1 Introduction
12.2 Universal quantifiers in Old Italian
12.3 From Old to Modern Italian
12.4 The dichotomy between bare and non-bare universal quantifiers
12.5 Negative quantifiers
12.6 Conclusive remarks
Sources
Part IV: Types of movement and its constraints: Word order change in Latin
13: The decline of Latin VOAux: Neg-incorporation and syntactic reanalysis
13.1 Introduction
13.1.1 Two directionality alternations in the Latin clause
13.1.2 A constraint on word order variation: the ‘Final-Over-Final Constraint’
13.1.3 A note on the empirical scope of the study
13.2 Classical Latin clause structure and the availability of VOAux
13.2.1 OV: strictly local ‘roll-up movement’
13.2.2 VPAux: non-local A-movement
13.3 The diachrony of VOAux
13.3.1 A corpus study
13.3.2 Discussion
13.3.3 A hypothesis
13.4 The diachrony of Latin negation
13.4.1 Jespersen’s cycle
13.4.2 Neg-incorporation, and its consequences
13.5 Syntactic reanalysis, and the demise of VOAux
13.6 Conclusion
Acknowledgements
14: On the decline of edge-fronting from Latin to Romance
14.1 Introduction
14.2 Deconstructing non-configurationality
14.2.1 Deconstructing grammatically free word order: roll-up
14.2.2 Deconstructing pragmatically free word order: edge-fronting
14.3 The decline of edge-fronting
14.3.1 Previous accounts
14.3.2 Proposal
14.3.2.1 Reanalysis: roll-up ) edge-fronting
14.4 Conclusion
References
Index of names
Index of languages
Index of subjects
Recommend Papers
![Clause Structure and Word Order in the History of German (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated]
9780198813545, 0198813546](https://ebin.pub/img/200x200/clause-structure-and-word-order-in-the-history-of-german-oxford-studies-in-diachronic-and-historical-linguistics-illustrated-9780198813545-0198813546.jpg)
![Parameter Theory and Linguistic Change (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated]
9780199659203, 0199659206](https://ebin.pub/img/200x200/parameter-theory-and-linguistic-change-oxford-studies-in-diachronic-and-historical-linguistics-illustrated-9780199659203-0199659206.jpg)
![Word Order and Parameter Change in Romanian: A Comparative Romance Perspective (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated]
9780198807360, 0198807368](https://ebin.pub/img/200x200/word-order-and-parameter-change-in-romanian-a-comparative-romance-perspective-oxford-studies-in-diachronic-and-historical-linguistics-illustrated-9780198807360-0198807368.jpg)
![Micro-change and Macro-change in Diachronic Syntax (Oxford Studies in Diachronic and Historical Linguistics) [1 ed.]
9780198747840, 0198747845](https://ebin.pub/img/200x200/micro-change-and-macro-change-in-diachronic-syntax-oxford-studies-in-diachronic-and-historical-linguistics-1nbsped-9780198747840-0198747845.jpg)
![Syllable and Segment in Latin (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated]
9780199660186, 0199660182](https://ebin.pub/img/200x200/syllable-and-segment-in-latin-oxford-studies-in-diachronic-and-historical-linguistics-illustrated-9780199660186-0199660182.jpg)
![Verb Second in Medieval Romance (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated]
9780198804673, 0198804679](https://ebin.pub/img/200x200/verb-second-in-medieval-romance-oxford-studies-in-diachronic-and-historical-linguistics-illustrated-9780198804673-0198804679.jpg)
![Word Order Change (Oxford Studies in Diachronic and Historical Linguistics) [Illustrated]
9780198747307, 0198747306](https://ebin.pub/img/200x200/word-order-change-oxford-studies-in-diachronic-and-historical-linguistics-illustrated-9780198747307-0198747306.jpg)
- Author / Uploaded
- Ana Maria Martins (editor)
- Adriana Cardoso (editor)
File loading please wait...
Citation preview
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word Order Change
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
OXFORD STUDIES IN DIACHRONIC AND HISTORICAL LINGUISTICS GENERAL EDITORS
Adam Ledgeway and Ian Roberts, University of Cambridge ADVISORY EDITORS
Cynthia Allen, Australian National University; Ricardo Bermúdez-Otero, University of Manchester; Theresa Biberauer, University of Cambridge; Charlotte Galves, University of Campinas; Geoff Horrocks, University of Cambridge; Paul Kiparsky, Stanford University; Anthony Kroch, University of Pennsylvania; David Lightfoot, Georgetown University; Giuseppe Longobardi, University of York; George Walkden, University of Konstanz; David Willis, University of Cambridge RECENTLY PUBLISHED IN THE SERIES
Portuguese Relative Clauses in Synchrony and Diachrony Adriana Cardoso Micro-change and Macro-change in Diachronic Syntax Edited by Éric Mathieu and Robert Truswell The Development of Latin Clause Structure A Study of the Extended Verb Phrase Lieven Danckaert Transitive Nouns and Adjectives Evidence from Early Indo-Aryan John J. Lowe Quantitative Historical Linguistics A Corpus Framework Gard B. Jenset and Barbara McGillivray Gender from Latin to Romance History, Geography, Typology Michele Loporcaro Clause Structure and Word Order in the History of German Edited by Agnes Jäger, Gisella Ferraresi, and Helmut Weiß Word Order Change Edited by Ana Maria Martins and Adriana Cardoso For a complete list of titles published and in preparation for the series, see pp. –
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word Order Change Edited by ANA MARIA MARTINS and ADRIANA CARDOSO
1
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
3
Great Clarendon Street, Oxford, OX DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © editorial matter and organization Ana Maria Martins and Adriana Cardoso © the chapters their several authors The moral rights of the authors have been asserted First Edition published in Impression: All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press Madison Avenue, New York, NY , United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: ISBN –––– Printed and bound by CPI Group (UK) Ltd, Croydon, CR YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Contents Series preface List of figures and tables List of abbreviations Notes on contributors . Word order change from a diachronic generative syntax perspective Ana Maria Martins and Adriana Cardoso
vii viii x xx
Part I. Targets for movement: Changes in the functional architecture of the clause . Configurational change in Indo-European coordinate constructions Moreno Mitrović . Discontinuous noun phrases and remnant-internal relativization in the diachrony of Portuguese Adriana Cardoso
. The relative cycle in Hungarian declaratives Julia Bacskai-Atkari
. Word order change at the left periphery of the Hungarian noun phrase Barbara Egedi
Part II. Triggers for movement: Changes in nature or stability . Particle-verb order in Old Hungarian and complex predicates Veronika Hegedűs
. An effect of residual T-to-C movement in varieties of English Judy B. Bernstein
. Word order and information structure in the Würzburg Glosses Cara M. DiGirolamo
Part III. Verb movement into the left peripheries . Subject inversion in transitive sentences from Classical to Modern European Portuguese: A corpus-based study Charlotte Galves and Alba Gibrail . Analyticization and the syntax of the synthetic residue Chris H. Reintges and Sonia Cyrino
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
vi
Contents
. Loss of laten-support in embedded infinitivals in fifteenth-century Low Saxon Gertjan Postma
. The distribution of quantifiers in Old and Modern Italian: Everything or nothing Jacopo Garzonio and Cecilia Poletto
Part IV. Types of movement and its constraints: Word order change in Latin . The decline of Latin VOAux: Neg-incorporation and syntactic reanalysis Lieven Danckaert
. On the decline of edge-fronting from Latin to Romance Adam Ledgeway
References Index of names Index of languages Index of subjects
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Series preface Modern diachronic linguistics has important contacts with other subdisciplines, notably first-language acquisition, learnability theory, computational linguistics, sociolinguistics and the traditional philological study of texts. It is now recognized in the wider field that diachronic linguistics can make a novel contribution to linguistic theory, to historical linguistics and arguably to cognitive science more widely. This series provides a forum for work in both diachronic and historical linguistics, including work on change in grammar, sound, and meaning within and across languages; synchronic studies of languages in the past; and descriptive histories of one or more languages. It is intended to reflect and encourage the links between these subjects and fields such as those mentioned above. The goal of the series is to publish high-quality monographs and collections of papers in diachronic linguistics generally, i.e. studies focussing on change in linguistic structure, and/or change in grammars, which are also intended to make a contribution to linguistic theory, by developing and adopting a current theoretical model, by raising wider questions concerning the nature of language change or by developing theoretical connections with other areas of linguistics and cognitive science as listed above. There is no bias towards a particular language or language family, or towards a particular theoretical framework; work in all theoretical frameworks, and work based on the descriptive tradition of language typology, as well as quantitatively based work using theoretical ideas, also feature in the series. Adam Ledgeway and Ian Roberts University of Cambridge
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of figures and tables Figures . The loss of the double system of coordination in Indic
. Grammar of conjunction in Latin: et, que, and atque from first century BC to fourth century CE
. Grammar of conjunction in Greek: καὶ and τε from eighth century BC to fifteenth century AD
. A diachronic sketch of syntactic development of coordination in Indo-European
. The relative VS/SV frequency in matrix declarative contexts, by author (Galves and Paixão de Sousa )
. The evolution of Subject-Verb-Object order in matrix declarative transitive sentences (by author)
. The evolution of Subject-Verb-Object order in matrix declarative transitive sentences (mean value by century)
. SVO, VSO, and VOS in matrix and embedded clauses
. Occurrence of IV in fifteenth-century Drenthe per , words
. Ratio of laten/all verbs in fifteenth-century Drenthe
. Relative frequency of laten in fifteenth-century Drenthe in tokens/, words
. Three fits to observed data on laten-support in fifteenth-century Drenthe
. Occurrence rate of present tense subjunctive forms in Low Saxon until
. Rise of sich/sick in reflexive contexts in fifteenth-century Drenthe (taken from Postma )
. Dialects with du ‘thou’; Double AGR dialects
. Frequency of VO in VPAux-clauses with a modal auxiliary, c. BC– AD
Tables . Categorial distribution of sub/clausal conjuncts for Ṛgvedic pen/initial coordinators (N = , ; numbers from Klein a,b)
. Development and loss of the double system of coordination in Indic
. Grammatical change from first century BC to fourth century CE in the conjunction system of Latin: et, que, and atque
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of figures and tables
ix
. Grammatical change in the Greek conjunction system of Greek from eighth century BC to fifteenth century AD
. Dunkel’s () reconstruction of two coordinator series in IE
. Clitic combinatorics as strategy for development of orthotone coordinators
. Complementizer word orders
. Some changes affecting hogy
. The distribution of multiple complementizers
. The distribution of complex complementizers
. Structures standing for finite subclauses
. Historical language stages of Hungarian and types of sources
. The proportion of definite determiners (a/az) in five Old Hungarian codices
. Positions of dative-marked possessors in Jókai Codex
. The spread of the constructions ‘ez/az az N’
. Particles and other predicative elements
. The typology of null subject and verb raising languages
. The inventory of the most common grammatical particles in Coptic
. Alternating lexical verb stems (√kt ‘to build’)
. The temporal system of Brazilian Portuguese
. Breakdown of occurrence of IV in fifteenth-century Drenthe
. Correlation table of IV and laten
. Absolute and relative use of laten as periphrastic auxiliary
. Occurrence of present subjunctive forms per , words
. Neg-doubling by a verbal clitic in fifteenth-century Drenthe in function of the finite verb’s position (C/Vfinal) and its auxiliaryhood (Aux/Lex): – + – (Postma and Bennis )
. Neg-doubling by a verbal clitic in fifteenth-century Drenthe in function of the finite verb’s position (C/Vfinal) and its auxiliaryhood (Aux/Lex): – (Postma and Bennis )
. Description of a Latin prose corpus
. Word order in Latin clauses with a modal auxiliary (absolute figures)
. Word order in Latin clauses with a modal auxiliary (percentages)
. Observed vs. estimated frequency of VOAux (VP-internal objects only: extraposed (VAuxO) and scrambled (OVAux) objects excluded)
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of abbreviations 0
a head-level category
0
an intermediate-level category (within X-bar theory)
^
a movement triggering feature
*
an asterisk marks an ungrammatical sentence
&
Coordination (head)
&P
Coordination Phrase
μ
coordinating functional head
√
Root
√P
Root Phrase
[ε]
a [EPP]-like feature, which, unlike [EPP], attracts and induces movement of the closest and the smallest syntactic object, a terminal/ head
[ε]
a notation that refers to whether a particle is a Wackernagel element, requiring second-position ([+ε]), or not ([ε])
[F]/+F/F
Focus feature; diacritic used to mark focused elements
φ-features
person-number features
[+N]
nominal feature
Xmin
a minimal syntactic category
0
X
π
XπP
a phasal head a phasal complement
∅
empty / empty category
0
a notation for phonological silence
first person
second person
third person
BoTU
Die Boǧazköy Texte im Umschrift (see end of Chapter )
A-movement
argument(al) movement
0
A -movement
non-argument(al) movement
ABL
Ablative
ABS
Absolutive
ACC/Acc
Accusative
AcI
Accusativus cum Infinitivus
ACT
Active
AD
Anno Domino
ADE
Adessive
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of abbreviations Adv
Adverb
Aen.
Aeneid (see end of Chapter )
AGR
Agreement
AGRC
Agreement in complementizer
AgrOP
Object Agreement Phrase
AgrSP
Subject Agreement Phrase
AGRT
Agreement in tense
Alb
Albanian
ALL
Allative
Anecd.
Anecdota from Irish Manuscripts (see end of Chapter )
AOR
Aorist
ARG
Argument
Asp
Aspect
AspP
Aspect/Aspectual Phrase
Aux
Auxiliary
AuxNegSOV
Auxiliary-Negation-Subject-Object-Verb order
AuxSO-Neg-V
Auxiliary-Subject-Object-Negation-Verb order
AuxSOV
Auxiliary-Subject-Object-Verb order
AuxVP
Auxiliary-Verbal Phrase order
b.
born
BC
Before Christ
BCE
Before the Common/Current Era
BHR
Biberauer, Holmberg, and Roberts ()
BP
Bhāgavatapurāṇa (see end of Chapter )
c.
century
C
Complementizer
c-command
constituent command
C-NSR
Nuclear Stress Rule (sensitive to asymmetric c-command)
c-selection
category selection
ca.
circa
CA.
Codex Argenteus (see end of Chapter )
Jn.
John
Mk.
Mark
Mt.
Matthew
Caes. B. C. Gal.
Caesar De Bello Ciuili De Bello Gallico
Cal.
Calabrian (dialects of extreme south of Italy)
Cat.
Catalan
xi
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
xii
List of abbreviations
CCC
CompertCon Culainn (see end of Chapter )
CE
Common/Current Era
Celtib.
Celtiberian
CEP
Contemporary European Portuguese
Cic.
Cicero
Att.
Epistulae ad Atticum
Cat.
In Catilinam
De Or.
De oratore
Diu.
De diuinatione
Fam.
Epistulae ad familiares
Nat. D.
De natura deorum
Phil.
Orationes Philippicae
Off.
De officiis
Or.
De oratore
prou. cons.
De prouinciis consularibus
Rab. Perd.
Pro Rabirio Perduellionis Reo
S. Rosc.
Pro Sexto Roscio Amerino
Sen.
De senectute
CIL
Corpus Inscriptionum Latinarum
CIL2
Corpus Inscriptionum Latinarum, nd edition
CL
Clitic
Class
Classifier
ClassP
Classifier Phrase
ClP
Classical Portuguese
CLuw
Cuneiform Luwian
CM.
Codex Marianus (see end of Chapter )
Jn.
John
Mt.
Matthew
COM
Comitative
COMP
Comparative / Complementizer
Comp
Complementizer
COND
Conditional
CONJ
Conjunction
COP
Copula
CP
Complementizer Phrase / the set of functional projections of the CP field
CRH
Constant Rate Hypothesis
(CS)OVAux
Complementizer-Subject-Object-Verb-Auxiliary order
D
Determiner
DAT
Dative
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of abbreviations DEF
Definite
DEL
Delative
DEM/Dem
Demonstrative
DemP
Demonstrative Phrase
DEON
Deontic
DEP
Dependent
DET/Det
Determiner
DIM
Diminutive
DIR
Directional
Dist
Distributive quantifier
DistP
Distributive-quantifier phrase
DL
Dual
DP
Determiner Phrase
DSL
Dictionary of the Scots Language
E/CSR
Emphatic/Contrastive Stress Rule
E-language
External language
EA
External Argument
ECM
Exceptional Case Marking
EL
Elative
EMP
Emphatic Particle
EP
Extended Projection / European Portuguese
EPP
Extended Projection Principle
ESP
Earlier stages of Portuguese
EXT
External
F
Feminine / unspecified functional head / Focus feature
Fin
Finiteness
FinP
Finiteness Phrase
Foc
Focus
FocP
Focus Phrase
FOFC
Final-over-Final-Constraint
ForceP
Force Phrase
FP
unspecified functional projection
FPR
Focus Prominence Rule
Fr.
French
FR
Fiore di rettorica (see end of Chapter )
FUT
Future
Gai.
Gaius
Inst. Gaul.
Institutiones Gaulish
xiii
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
xiv
List of abbreviations
GEN
Genitive
Goth.
Gothic
Gr.
Greek
Hist. Aug.
Historia Augusta
Hit.
Hittite
HMC
Head Movement Constraint
Hom.
Homer
Il.
Iliad
HPP
Head Preference Principle
I-language
Internal-language
IA
Internal Argument
IE
Indo-European
IIr.
Indo-Iranian
Il.
Iliad (see end of Chapter )
ILL
Illative
IMP
Imperative
IMPERF
Imperfect
IMPERS
Impersonal clitic se
IMPF
Imperfect
IND
Indicative
INDEF
Indefinite
INE
Inessive
INF/inf
Infinitive
INST/INSTR
Instrumental
INT/Int
Interrogative
int. al.
inter alia
INT.ARG
Internal Argument
IO
Indirect Object
IP
Inflection Phrase / the set of functional projections of the IP field
IpI
Imperativus pro Infinitivo
IS
Information Structure
It.
Italian
IV
Infinitival verb-second
IX
ninth class
J
Junction (head)
JP
Junction Phrase
Kal.
Kalauz (see end of Chapter )
KUB
Keilschrifturkunden aus Boǧazköy (see end of Chapter )
Lat.
Latin
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of abbreviations LBC
Left Branch Condition
LCA
Linear Correspondence Axiom
Lex
Lexical
lh
The Irish prefaces from the Liber Hymnorum (in POMIC)
LMP
Late Merge Principle
LOC
Locative
M
Masculine
MAND
Morphological Atlas of the Dutch Dialects (MAND)
Mart.
Martial (Epigrammata)
Mbh
Mahābhārata (see end of Chapter )
MID
Middle
Ml
Milan Glosses
Mod
Modal
ModP
Modal Phrase
MS.
Manuscript
Myc.
Mycenaean
N
Noun / Neuter
n-word
negative word
NAI
Negative Auxiliary Inversion
NEG/Neg
Negation / Negative Particle
Neg-SOVAux
Negation-Subject-Object-Verb-Auxiliary order
NegP
Negation Phrase
NOM
Nominative
Nov.
Il Novellino (see end of Chapter )
NP
Noun Phrase
NREL
Negative Relative Particle
NSR
Nuclear Stress Rule
NumP
Number Phrase
O/OBJ
Object
OBL
Oblique
OCS
Old Church Slavonic
OI
Old Italian
OIr.
Old Irish
OP
Old Portuguese
Op
Operator
Or.
De Oratore (see end of Chapter )
OSV
Object-Subject-Verb order
OV
Object-Verb order
Ov.
Ovid
xv
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
xvi
List of abbreviations
Met.
Metamorphoses
OVI
Opera del Vocabolario Italiano
OVNegAux
Object-Verb-Negation-Auxiliary order
OVS
Object-Verb-Subject order
OVSAux
Object-Verb-Subject-Auxiliary order
P
Phrase / Preposition
p-movement
prosodically motivated movement
PART
Participle
PASS
Passive
PassP
Passive Past Participle
PastP
Past Participle
PCL
Particle
PERF
Perfect
PERL
Perlative
PF
Phonological Form
PIC
Phase Impenetrability Condition
PIE
Proto-Indo-European
PJ
Puṇyavanta-Jātaka (see end of Chapter )
PL/pl
Plural
PLD
Primary Linguistic Data
POA
Porto Alegre (Brazil)
Pol
Polarity
POMIC
Parsed Old and Middle Irish Corpus
POSS
Possessive / Possessee/Possessum (suffix on the head noun that encodes its being possessed)
PossP
Possession Phrase
PP
Prepositional Phrase / Adpositional Phrase
pP
particle Phrase
PPRF
Pluperfect
Pred
Predicative
PredP
Predicative Phrase
PREP
Preposition
PRES
Present
PRET
Preterite
PRF
Perfect
pro/PRO
Null Pronoun
PROGR
Progressive
PRS
Present
PRT
Particle / Preterite
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of abbreviations PrtP
Particle Phrase
P.S.
Post Scriptum corpus (see end of Chapter )
PST
Past
Pt.
Portuguese
PTCL
Particle
PTCP
Participle
PV
Preverb
Q
Quantifier / Interrogative / Question Particle
Q. Curt.
Quintus Curtius
Hist.
Historiae Alexandri Magni
QP
Quantifier Phrase
QUOT
Quotative
R
reference time
REFL
Reflexive
REL/rel
Relative
RRC
Restrictive Relative Clause
ṚV
Ṛgveda saṃhitā (see end of Chapter )
S
Subject
S-NSR
Nuclear Stress Rule (sensitive to selectional ordering)
s-selection
semantic selection
S-structure
Surface structure
SBJ
Subject
SBJV
Subjunctive
SC
Small Clause
Sen.
Seneca
Ben.
De Beneficiis
seq.
sequential
SG/sg
Singular
Skt.
Sanskrit
Sl.
Slavonic
SO-Neg-VAux
Subject-Object-Negation-Auxiliary order
SOV
Subject-Object-Verb order
SOVAux
Subject-Object-Verb-Auxiliary order
Sp.
Spanish
Spec
Specifier
Srd.
Sardinian
STAT
Stative
StBoT
Studien zu den Boǧazköy-Texten (see end of ch- )
SU
Subject
xvii
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
xviii
List of abbreviations
SUB
Sublative
SUBJ/subj
Subjunctive / Subject
SubjP
Subject(-of-Predication) Phrase
SUBS
Substantive
SUP
Superessive
SUPPL
Suppletive base
SV
Subject-Verb order / Sposizione di Vangeli (see end of Chapter )
SVO
Subject-Verb-Object order
SVOAux
Subject-Verb-Object-Auxiliary order
SXO
Subject-unspecified constituent-Object order
SXV
Subject-unspecified constituent-Verb order
T
Tense
t
trace
TA
Tocharian A
Tac.
Tacitus
Hist.
Historiae
TB
Tocharian B
TERM
Terminative
THT
Die Speisung des Bodhisattva vor der Erleuchtung (see end of Chapter )
TAME
Tense-Aspect-Mood-Evidentiality
TN
Tennessee
TochA
Tocharian A
TOP/Top
Topic
TopP
Topic Phrase
TP
Tense Phrase
TYC
Tycho Brahe Parsed Corpus of Historical Portuguese (see end of Chapter )
u
uninterpretable (feature)
v/v
light verb
V
Verb
V
Verb-first
V
Verb-second
VAuxO
Verb-Auxiliary-Object order
Var.
Varro
R.
De re rustica
VC
Vita Constantini (see end of Chapter )
Ven.
Venetic
VeV
Il Libro de’ Viz e delle Virtudi (see end of Chapter )
Vir.
Virgil
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
List of abbreviations Aen.
Aeneid
VM
Verbal Modifier
VNegAux
Verb-Negation-Auxiliary order
VO
Verb-Object order
VOAux
Verb-Object-Auxiliary order
VOC
Vocative
VONegAux
Verb-Object-Neg-Auxiliary order
VOS
Verb-Object-Subject order
vP
Light verb Phrase
VP
Verb/Verbal Phrase
VPAux
Verbal Phrase-Auxiliary order
VS
Verb-Subject order
VSO
Verb-Subject-Object order
VT
Vetus Testamentum Arminiace (see end of Chapter )
Mt.
Matthew
VXS
Verb-unspecified phrase-Subject order
w
word(s)
wh/wh-/WH/WH-
words such as who, which, what, why, when, etc.
X
unspecified head or constituent or feature
X-movement
Head movement
XP
unspecified Phrase
XP-movement
Phrasal movement
XSV
unspecified phrase-Subject-Verb order
YH
Yasna Haptanghāiti (see end of Chapter )
YP
unspecified phrasal complement of a head
ZP
unspecified phrasal specifier
xix
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Notes on contributors JULIA BACSKAI-ATKARI received her PhD in from the University of Potsdam, Germany, where she is currently a postdoctoral research fellow as the principal investigator of the DFGproject, ‘The Syntax of Functional Left Peripheries and Its Relation to Information Structure’. Her dissertation, published in by Universitätsverlag Potsdam, examined comparative constructions in English from a cross-linguistic perspective. In her current project, she investigates the left periphery of West Germanic embedded clauses synchronically and diachronically. While she expertises in Germanic languages, she has written extensively on other European languages as well, including various Slavic languages and Hungarian. JUDY B. BERNSTEIN is professor of linguistics at the William Paterson University of New Jersey. Her research examines syntactic aspects of cross-linguistic variation across Romance languages and varieties of English. She has worked extensively on the structure of the noun phrase. Her research has appeared in several journals (for example, Probus, Lingua, Linguistic Inquiry, Language and Linguistics Compass) and edited volumes. She is currently completing two papers for publication, one (co-authored) on the diachrony of personal articles across varieties of Catalan and another on the development and diversity of proximal and distal demonstratives in Romance languages. ADRIANA CARDOSO is assistant professor at the Higher Education College of Lisbon (ESELx) and researcher at the Linguistics Centre of the University of Lisbon (CLUL). Her main research interests are historical linguistics, comparative syntax, and educational linguistics. She was involved in the project ‘WOChWEL—Word Order and Word Order Change in Western European Languages’ (sponsored by FCT) and has recently published the book Portuguese Relative Clauses in Synchrony and Diachrony (OUP ). SONIA CYRINO is Associate Professor at the Department of Linguistics, University of Campinas, Brazil. Her interests include diachronic and comparative Romance syntax focusing on null objects, verb movement, and DP structure. Recent publications include the chapter ‘Null objects in Romania Nova’ to appear in Studies in Romania Nova (ed. M. Kato and F. Ordoñez, OUP); her collaboration with M. Teresa Espinal, ‘Bare Nominals in Brazilian Portuguese: more on the DP/ NP analysis’, published in Natural Language & Linguistic Theory; the chapter ‘On richness of tense and verb movement in Brazilian Portuguese’, in M. V. Camacho-Taboada et al., Information Structure and Agreement (John Benjamins ). LIEVEN DANCKAERT works as a researcher at the Centre National de Recherche Scientifique (CNRS), and he is affiliated with the University of Lille. His main research interest is Latin syntax, and in particular word order. He is the author of a book on the structure of the clausal left periphery in Latin entitled Latin Embedded Clauses: The Left Periphery (John Benjamins ). His monograph The Development of Latin Clause Structure was published by OUP in . CARA M. DIGIROLAMO is a PhD student in the Department of Linguistics at Cornell University. Her dissertation, Head Movement as Intra-Syntactic Morphology, investigates the viability of the weak pronoun category with perspectives from Middle Welsh. Using a corpus-driven approach,
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Notes on contributors
xxi
she also works on topics such as Japanese deontic modality and Old Irish information structure. Her article, ‘The Fandom Pairing Name: Blends and the Phonology-Orthography Interface’, appeared in Names: The Journal of Onomastics in . BARBARA EGEDI is a research fellow at the Research Institute for Linguistics of the Hungarian Academy of Sciences, currently holding a postdoc position. She has published several papers on the noun phrase structures of Old and Middle Hungarian as well as Ancient Egyptian and Coptic. Her major interests are the change in definiteness marking and in possessive constructions. She has been working on the syntax of adverbials in Modern Hungarian and on complex sentences of Hungarian Sign Language. At present, she also participates in a comparative syntactic research project of Uralic languages and is preparing a descriptive grammar of a Coptic dialect. CHARLOTTE GALVES is Professor of Linguistics at the University of Campinas. Her recent publications include the chapter on ‘Galician and Portuguese’ in The Oxford Guide to the Romance Languages, in co-authorship with Francisco Dubert (OUP ) and the chapter on ‘Main syntactic changes from a principle and parameters view’ in the Handbook of Portuguese Linguistics, in co-authorship with Anthony Kroch (Wiley-Blackwell ) . She is co-editor of Parameter Theory and Linguistic Change (OUP ), and of África-Brasil, Caminhos da Língua Portuguesa (Editora da Unicamp ). She is the founder and coordinator of the Tycho Brahe Annotated Corpus of Historical Portuguese. JACOPO GARZONIO is assistant professor of Linguistics at the University of Padua. His research focuses on the morpho-syntactic variation in Italo-Romance, with a special interest in negation, complementizer systems, wh items, quantifiers and adpositions, both from a synchronic and a diachronic point of view. He is currently working on a book on the history of Italian negation, and is the principal investigator in a project on the grammar of medieval Venetian varieties. ALBA GIBRAIL is a postdoctoral fellow in Linguistics in the Institute of Studies of the Language of the State University of Campinas, Brazil, acting as a collaborating researcher with the thematic project ‘The Portuguese Language in Time and Space: Language Contact, Grammars in Competition and Parametric Change’, funded by FAPESP and directed by Charlotte Marie C. Galves. In her current postdoctoral project, she investigates the prepositional accusative structures in European Portuguese (seventeeth to nineteenth centuries). Her publications focus on the syntax of Classical Portuguese (e.g. topicalization, clitic left dislocation, focalization, and V). VERONIKA HEGEDŰS is research fellow at the Research Institute for Linguistics of the Hungarian Academy of Sciences. She did her PhD in on ‘Non-Verbal Predicates and Predicate Movement in Hungarian’, and she has published papers on the syntax of Hungarian adpositions and verbal particles synchronically and diachronically, on copular clauses and on pseudoclefts. Her research interests include word order, non-verbal predicates, resultatives and motion verbs, the syntax of particles and adpositions, diachronic syntax. ADAM LEDGEWAY is Professor of Italian and Romance Linguistics at the University of Cambridge and since has been Chair of the Faculty of Modern and Medieval Languages. Recent books include The Cambridge History of the Romance Languages. Volume : Structures, Volume II: Contexts, co-edited with Martin Maiden and J.C. Smith; CUP /), From Latin to Romance: Morphosyntactic Typology and Change (OUP ), Diachrony and
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
xxii
Notes on contributors
Dialects. Grammatical Change in the Dialects of Italy (co-edited with Paola Benincà and Nigel Vincent; OUP ), and The Oxford Guide to the Romance Languages (co-edited with Martin Maiden; OUP ). He is currently Principal investigator on two research-funded projects, ‘Reflexive Clitics in Aromanian’ (Marie Curie Foundation) and ‘Fading Voices in Southern Italy: Investigating Language Contact in Magna Graecia’ (Leverhulme Major Research Grant). ANA MARIA MARTINS is Professor at the School of Arts and Humanities of the University of Lisbon. Her research in comparative syntax and historical syntax covers topics such as word order, clitics, negation, emphatic polarity, infinitival structures, passive and impersonal constructions. She has directed projects resulting in parsed corpora for the study of the syntax of European Portuguese dialects (CORDIAL-SIN) and the syntax of Old Portuguese (WOChWEL). She has published articles in journals such as Lingua, Natural Language and Linguistic Theory, Linguistic Inquiry, and chapters in books published by Oxford University Press, Benjamins, and DeGruyter. She is editor of Manual de Linguística Portuguesa, DeGruyter, and co-editor of the journal Estudos de Lingüística Galega. MORENO MITROVIĆ holds a PhD in theoretical and diachronic linguistics from Jesus College, Cambridge and has since worked as lecturer at Saarland University, Germany, and as researcher at Graz University, Austria, the Bled Institute, Slovenia, and the University of Cyprus. During his doctoral research, he was also a visitor at MIT (/) and fellow at Harvard University (/) and the National Institute of Japanese Language & Linguistics (NINJAL) (). His published work focuses on the diachronic and typological syntax/ semantics/pragmatics of logical vocabulary. His thesis, Morphosyntactic Atoms of Propositional Logic: A Philo-Logical Programme (), provides a diachronic syntactic-semantic analysis of quantificational/conjunctive particles in Indo-European (and Japonic). His monograph on the topic is forthcoming with Springer. CECILIA POLETTO holds positions at the Goethe University in Frankfurt and at the University of Padua. She has worked on the syntax and morphology of Italian dialects from a synchronic and diachronic perspective and is part of the Paduan group involved in the ASIt (the online Syntactic Atlas of Italy) http://asit.maldura.unipd.it/. She has more recently worked on Old Italian and has published a monograph, Word Order in Old Italian. She is currently working on two DFG research projects on relative clauses and quantification in Old Italian. GERTJAN POSTMA is senior researcher in variational linguistics at the Meertens Institute in Amsterdam, a research institute of the Netherlands’ Academy of Sciences. He specializes in quantitative and qualitative models in diachronic syntax, especially formal properties, such as negation, pronominalization, and inalienable possession. He built an automatic parser of Middle Dutch, and made an annotated historical corpus (Low Saxon dialect, –). In recent years, he is investigating language change in Dutch and Low Saxon language islands in Brazil, and is currently writing a contrastive grammar of Brazilian Pomeranian, spoken in Espirito Santo. CHRIS H. REINTGES is senior researcher at the Centre National de Recherche Scientifique (CNRS), Paris. He received his PhD in linguistics from Leiden University, The Netherlands, in , with a dissertation entitled ‘Passive Voice in Older Egyptian—A Morphosyntactic Study’. Subsequently he worked extensively on Coptic, which resulted in a comprehensive grammar, entitled Coptic Egyptian (Sahidic Dialect): A Learner’s Grammar (Cologne ). Current research interests include minimalist syntax, parameter theory, historical linguistics and language change. He is preparing The Oxford Guide to Afroasiatic Languages (OUP forthcoming), co-edited with Sabrina Bendjaballah.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
1 Word order change from a diachronic generative syntax perspective AN A MA R I A M A R TI NS AN D A D R I A NA C A R D OS O
. Introducing the topic Word order is at the core of natural language grammatical systems. It interacts with all their structural components, linking syntax with sound (prosody) and meaning (semantics/pragmatics), manifesting the characteristic ‘displacement property’ (Chomsky ) that lies behind the opposition between base-generated and derived word orders, and feeding variation across the geographic and temporal axes. Deepening our understanding of word order phenomena seems to require being able to articulate the successful but often separate efforts of typologists, historical and theoretical linguists so as to take full advantage of both large-scale cross-linguistic surveys and the more circumscribed and in depth analyses of theory-oriented research. This book aims to be a contribution to that objective, taking linguistic change and generative syntax as its underpinning pillars. It offers thirteen fine-grained studies in word order change set within the framework of diachronic generative syntax. The sample of languages investigated is diverse and displays significant historical depth. Different branches of the Indo-European family are represented both through classical and living languages (namely, a wide range of Early Indo-European languages, Latin, Italian, European Portuguese, Brazilian Portuguese, Irish, Dutch, and English). Besides, three chapters are dedicated to Hungarian and one chapter deals with Egyptian. The wide temporal range covered by the book’s chapters allows for apparently contradictory but actually thought-stirring approaches to complex, enduring issues such as the evolution from synthetic to analytic grammatical systems. While Ledgeway (chapter ) treats the evolution from Classical Latin to Early Romance as a model instance of the synthetic to analytic drift, Reintges and Cyrino (chapter ) characterize present-day Brazilian Portuguese as a language in which analyticization is an emergent grammatical property (due to the ongoing state of the change), distant as it is from Early Romance. This apparent dissonance (between
Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Ana Maria Martins and Adriana Cardoso . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso
authors that share a macroparametric view of the synthetic to analytic drift) only indicates that work in the domain of diachronic generative syntax should contribute to further refining widely used conceptual categorizations through detailed inquiry into the properties of particular linguistic systems, without losing the broad picture. The coherent body of original research on word order change assembled in this book seeks to go precisely in that direction. The chapters in the book use the tools provided by the generative theory of grammar to look into the constrained ways in which older linguistic variants give rise to new ones in the course of time, with the aim of contributing insights into the properties of natural language and the ways those properties constrain language change. Two ingredients of the generative framework make it especially appropriate to deal with word order phenomena, namely movement as a syntactic operation (embedded in the theory of grammar) and a richly articulated clausal architecture composed with lexical but also abstract functional categories. The latter might be thought of as constant across languages or instead partially variable (an unsettled theoretical issue), while the former underlies the distinction between base-generated and derived word orders. If one assumes that clausal architecture may vary across languages (and across constructions), variation in clause structure (i.e. activation/nonactivation, conflation/expansion, structural position and featural make-up of functional categories) interacts with movement operations creating the different options for linearization that result in word order variation. Word order variation emerges along a number of different axes (temporal, spatial, social, cultural, etc.) and according to different media (oral, written, printed, etc.). This book is primarily concerned with the articulation between the temporal and spatial axes, and aims to take full advantage of a comparative inquiry as a tool to uncover non-trivial generalizations and new research problems. While each chapter in the book addresses specific issues and identifies others for future research, the book as a whole incites and allows the reader to consider more general questions motivated by the different contributions it includes. To give a few examples: are the notions of parameter and macroparameter necessary to understand word order variation?; what does word order change tell us about clause structure?; how does it give evidence against the existence of non-configurational languages?; how does it support the view that syntactic systems, like phonological systems, to a certain extent select from a universal set of categories those that belong to a particular grammar?; how might other parallelisms between phonology and syntax be fruitfully drawn on for the purposes of diachronic/historical investigation? (e.g. the notion of chain reaction, drawn from historical phonology, could be applied to syntax and maybe replace the concept of macroparameter to deal with such sets of changes as the drift from synthetic to analytic languages). In the next section we will focus on the relation between movement operations and word order by assembling the pieces of information offered by the book’s authors. We will show that the works here published indicate, when considered together, that word order change might be mainly the effect of the interaction between clause structure and syntactic movement, thus identifying these two
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
components of grammar as the main factors behind word order variation (while, at the same time, significantly lessening the role traditionally attributed to morphology in driving word order change). We hope to be able to show that what at first glance may seem trivial turns out not to be so. We will thus identify some research issues that emerge from the type of approach to word order change here envisaged.
. Clause structure, syntactic movement, and word order change We will start this section with two examples drawn from Portuguese (and borrowed from outside the book). The first example will describe an instance of change in constituent order as the effect of loss of movement, whereas the second will show how the emergence of syntactic movement associated with particular lexical items may give rise to new options for word order. We will then refer to the studies offered in the chapters to in relation to the dichotomy loss/gain of movement (over time) and will try to put forth some descriptive generalizations on the conditions underlying each of the possibilities. These generalizations will in turn lead to some research questions specifically addressing word order change. Contemporary European Portuguese displays VP-scrambling (i.e. short scrambling) as a strategy to establish appropriate information focus configurations by extracting out of the sentence-final position to which focus-stress is assigned the constituents that are not given informational prominence (Costa , ). Object scrambling in contemporary European Portuguese maintains the basic VO order of the language, since the verb regularly moves to T(ense). Old Portuguese, on the other hand, allowed both VP-scrambling and IP-scrambling (i.e. middle scrambling) with parallel information-structural motivation. Old Portuguese middle scrambling derived the type of OV order illustrated in () below. In (a–c) the position of the clitic pronouns (in bold) unambiguously shows that the scrambled objects have not moved to the sentential left periphery; so does the fact that in (d) the object follows the negative word nunca (‘never’), that would trigger proclisis if the sentence contained a clitic pronoun.1 () a. sse pela uẽtujra uos alguẽ a dita vỹa enbargar if by chance you.CL.DAT someone the said vineyard blocks ‘If it happens that someone blocks the said vineyard from you’.2 b. mas nunca achou quem lhe delle novas algũas dissesse but never found who him.CL.DAT of-he news some-PL would-say ‘But he couldn’t find anybody who would tell him any news from him’. c. ainda vos oje muito pesar averrá still you.CL.DAT today much sorrow will-come ‘Today it will still come much sorrow to you’.
1 2
See Martins (, , forthcoming). Legal document, year . Cf. Martins (: ).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso d. E nunca outros panos ouve and never other garments had.SG ‘And he never had any other garments (after then)’.3
Martins (, ) accounts for the loss of middle scrambling in sixteenth-century Portuguese (with the consequence that OV sentences such as (a–d) ceased to be derived) as a consequence of a change in clause structure. Specifically, it is proposed that in Old Portuguese the functional category T would make available multiple specifier positions, which could be targeted by scrambled constituents. When this property of T was lost, middle scrambling was also lost with the result that object scrambling stopped being a possible source of the OV order. Independently of the particular technical implementation of the analysis, what should be retained here is the possible relation between the contraction of clause structure and loss of movement, a type of relation that will reappear in some of the instances of word order change investigated in the book (cf. chapters , , ). The reverse path is a logical possibility and is actually also described in the book (cf. chapter ). Whenever, in the course of time, clause structure expands instead of contracting, new structural positions appear as possible targets for movement operations, which can then emerge anew instead of being lost. If this type of analysis proves to be right, word order change will be giving evidence for the theoretical viewpoint that (certain dimensions of ) clause structure may vary across languages and language varieties, and providing indicators to identify the loci of possible variation. The second case of word order change in the history of Portuguese we will summarily refer to here illustrates the emergence of syntactic movement motivated by the pragmaticization of certain lexical items. Diachronically, it bears an indirect relation to the change relative to object scrambling. In Old Portuguese the deictic locatives lá (‘here’) and cá (‘there’) regularly display locative interpretation and occupy the regular positions of verbal arguments in the clause structure. From the sixteenth century on, they start to reveal other pragmatic/semantic values, each of them associated with a particular higher position in the functional architecture of the clause, thus featuring a case of upward reanalysis through the functional hierarchy in the sense of Roberts and Roussou (). The change is better described as a case of pragmaticization than grammaticalization because the primitive argumental locative value has been preserved while the association with higher positions in the functional system originated the new non-argumental values. The sentences in () below exemplify, from (a) to (d) respectively, the regular locative interpretation of postverbal argumental lá, the emphatic contribution of non-argumental lá (which then surfaces in pre-verbal position in declaratives but in post-verbal position in imperatives because imperative verbs move higher) and the workings of lá as a metalinguistic negation marker. In this last case, illustrated in (d), lá is consistently post-verbal
3 Examples (b–d) belong to the medieval literary text Demanda do Santo Graal, which is part of the thirteenth-century Portuguese translation from French of the Post-Vulgate Arthurian Cycle. Cf. WOChWEL’s POS-tagged and Parsed Old Portuguese texts (A. M. Martins, coord; CLUL/FCT). http://alfclul.clul. ul.pt/wochwel/oldtexts.html.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
(and adjacent to the verb) but at the same time standard tests for verb movement give evidence that the verb has moved into the CP field.4 () a. e viverom la muito and lived there much ‘And they lived there for a long time’.5 b. se não andavam sobre aviso lá ia a cepa e a cepeira if not were.PL on guard lá went the vine and the grapevine ‘If they were not on their guard, they would have lost everything!’ c. conta lá outra história tell.IMPERATIVE.PL lá another story ‘Please, tell us another story!’6 d. Quem foi? Eu sei lá, senhora mãe! who was? I know lá lady mother ‘Who did it? I don’t know, mother. How shall I know?’7 Martins (, ) associates non-argumental emphatic lá with Spec,TP (which is analysed as a dedicated position for deictic items) and metalinguistic negation lá with movement of the deictic locative from Spec,TP to Spec,CP. The relation between the development of the non-argumental values of the deictic locatives and the changes with respect to object scrambling is proposed to be an indirect one. The EPP feature of T(ense) was in Old Portuguese of an unselective nature allowing the projection of multiple specifiers (which could host several scrambled constituents) but evolved into being highly selective, limiting Spec,TP to one single position only accessible to deictic items. Hence, after middle scrambling was lost as a cross-categorial movement operation, deictic locatives could still target Spec,TP, instantiating a residual highly constrained type of middle scrambling (Costa and Martins ). Scrambled pre-verbal deictic locatives in Spec,TP could then be reanalysed as emphatic markers in contexts allowing ambiguity between argumental and non-argumental interpretations of the deictic. Once Spec,TP allowed external merge of non-argumental deictic locatives, it was prone to become the anchor for further movement into higher positions in the functional hierarchy. This case of word order change featuring the development of CP-reaching metalinguistic negation markers from basic deictic locatives indicates that syntactic movement may emerge in relation with the development of polymorphism by particular lexical items. As for changes making a position targeted by syntactic movement become more restrictive (as with middle
4
In Brazilian Portuguese, the corresponding sentence to (d) would display lá in pre-verbal position, which is consistent with the general observation in the literature that Brazilian Portuguese (in contrast with European Portuguese) has lost verb movement into the C domain. Cf. chapter . 5 Thirteenth-century Demanda do Santo Graal. See footnote . 6 Examples (b–c) are taken from the sixteenth-century plays of Gil Vicente. Cf. Gil Vicente, Todas as Obras. CD-ROM. 7 Example taken from a nineteenth-century novel by Júlio Dinis. Cf. Corpus do Português (s–s), Mark Davies and Michael Ferreira (– )—http://www.corpusdoportugues.org.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso
scrambling in Portuguese, that came to be restricted to deictics) or instead less restrictive, the studies presented in this volume show that both directions might indeed be expected (see chapters and ). Most chapters in the book deal with cases of word order change motivated by some change in the properties of a functional category associated with syntactic movement, as will be specified further on. However, none of the chapters handles a case such as the change that the Portuguese deictic locatives underwent over time. This seems to indicate that investigation into the diachronic development of so-called ‘discourse particles’ is the domain more prone to reveal similar cases of emergence of movement to the sentential left periphery (but see Reintges a on Coptic TAM particles). The essays on word order change by Mitrović (chapter ), Cardoso (chapter ), and Bacskai-Atkari (chapter ) all discuss instances of word order change motivated by a specific contraction of clause structure. Mitrović explains the evanescence of the enclitic (Wackernagel’s second-position) type of coordinate construction in IndoEuropean as the effect of the loss of an independent particular kind of coordinate phrase, within an analysis where coordination involves a structure with three coordinate heads. After the change took place, Indo-European languages came to display only initial coordination instead of the earlier two competing structures. Cardoso investigates Portuguese discontinuous noun phrases in the context of restrictive relative clauses. She shows that changes in word order between Old and Modern Portuguese were driven by the loss of a left-peripheral position for contrastive foci within relatives and possibly other types of subordinate clauses. The reduction of the left-peripheral space of relatives significantly constrained the options for linearization of the relevant type of noun phrases. Bacskai-Atkari accounts for changes in the relative order between complementizers in Hungarian subordinate clauses as the consequence of the diachronic reanalysis of the two C-heads originally available for complementizers into one single C-head. As a result of the change, variation in the relative position of complementizers was replaced by a fixed order. Common to all the cases of word order change initiated by contraction of clause structure is the fact that it limits word order flexibility barring certain options for linearization that were available at an earlier period. This is so because clausal contraction goes hand in hand with the loss of syntactic movement. Egedi’s study (chapter ) deals instead with word order change originated by the expansion of the functional structure of the Hungarian DP. Egedi accounts for the change from the Demonstrative-Possessor-Possessum order found in Middle Hungarian into the Possessor-Demonstrative-Possessum order found in Modern Hungarian as the effect of the reanalysis of adjunction positions as new left-peripheral layers of the extended noun phrase. Demonstratives became DP-specifiers instead of adjuncts and a higher Topic phrase emerged at the left periphery of the DP where the Possessor moves. Word order change related to changes in the functional architecture of the clause (or some of its components, like the DP) may bring about new and important evidence to determine what are the possible dimensions of variation in clause structure across languages and within a particular grammar. A better understanding of this problem will in turn help us to correctly diagnose those instances of word order change that truly involve some reconfiguration of clause structure.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
Besides clausal contraction/expansion, a structural position associated with syntactic movement may in the course of time become of a different nature and so trigger movement in a more constrained or less constrained way. Hegedűs (chapter ) and Bernstein (chapter ) deal with this type of change. Hegedűs shows that although Hungarian evolved from an SOV to a SVO language, verbal particles and predicative constituents other than particles still surface in pre-verbal position. The author accounts for this unexpected fact by showing that what was originally a pre-verbal argument position was reanalysed as a pre-verbal predicative position where complex predicates are formed in overt syntax. At a first step of the change, only particles kept moving to the pre-verbal position; then, by Modern Hungarian, the still obligatory movement of particles triggered the generalization of predicate movement, making all predicates pre-verbal in neutral sentences. Hence movement to the pre-verbal position was first restricted, then expanded. Bernstein deals with residual verb second in Appalachian English and African American English. She shows that while in Older Scots verb movement to C was general (like in most current Germanic languages), in Appalachian English (like in African American English) it came to be restricted to modals or finite auxiliaries occurring in transitive expletive constructions and negative auxiliary inversion. Bernstein suggests that over time the verb-second configuration came to be triggered only by features associated with Focus. The study by DiGirolamo (chapter ) also deals with instances of movement to the clausal left periphery, but features diachronic stability instead of change. The author shows that Old Irish sentences displaying non verb-initial orders are the outcome of argument fronting to the sentential left periphery for information-structural purposes just like in Modern Irish (Irish being invariably a VSO language). Thus the chapter handles a case of stability in word order over time, finely uncovering textual evidence in the domain of Information Structure. DiGirolamo’s investigation suggests, against a common claim in the literature, that information structurally marked positions should not be seen as diachronically unstable and thus prone to (neutral) reanalysis over time (cf. the explanations for the hypothetical emergence of V in Late Latin/Proto-Romance proposed by Roberts a; Ledgeway ; Wolfe ). Like Bernstein, Galves and Gibrail (chapter ), Reintges and Cyrino (chapter ), and Postma (chapter ) also discuss cases of word order change involving the loss of verb movement into the CP domain (i.e. ‘verb-second’). Postma attributes it to a change in the properties of TP in Low Saxon, which could no more be crossed by rich verbal forms. Reintges and Cyrino show that TAME particles (encoding Tense, Aspect, Mood, Evidentiality) emerge in Coptic Egyptian as part of the analyticization drift of the language; as these particles are merged in high positions through the functional skeleton of the clause, they block verb movement past the lower T(ense) head.8 An essential question to address in this context is why the loss of verb movement to C should be so common diachronically and the opposite change 8
The chapter by Galves and Gibrail, which combines corpus work with textual and theoretical analysis, does not identify what may have caused the loss of the verb second property attributed to Classical Portuguese.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso
featuring the emergence of verb movement to C apparently so rare across languages (as a cursory review of the ten DiGS volumes published between and would readily illustrate9). If this diachronic contrast is confirmed (which requires broadening the empirical coverage of diachronic generative inquiry beyond the IndoEuropean family), it can be used as an indicator to evaluate diachronically oriented approaches to the verb-second property. Consider, for example, the idea that input generalization (in the process of language acquisition) plays a crucial role in the emergence of V from non V grammars. We would then expect the emergence of V grammars to be more strongly attested because V syntax restricted to particular constructions is quite common across languages, which could thus be the source for analogy-based overgeneralizations.10 Also the idea that information structural positions like topic and focus can be reanalysed as targets of unselective, generalized movement to the clausal left periphery faces a similar problem, as many languages would seemingly have the potential to develop V grammars. Hence, either there is a problem with the proposed models of change to account for the diachronic emergence of V or the problem is our current understanding of V itself. That is to say, perhaps we still do not have a clear understanding of the nature of the features that trigger movement to the left periphery in V grammars and this limits our understanding of diachronic change in this domain. Be it as it may, the fact that changes involving verb-second mostly feature V loss must be telling us something about the verb-second property itself, even if we still cannot put our finger on it.11 Against the apparently more usual diachronic path of loss of verb movement, Garzonio and Poletto (chapter ) demonstrate that the past participle moves higher in Modern Italian than in Old Italian. The authors assume that Old Italian was a verb-second language across the CP, vP and DP phases (a matter of parametric setting). When the verb second property was lost, the focus head within the left periphery of the vP no longer needed to be filled by the past participle, which then starts to move out of the vP and reach the aspectual low IP field. Apart from the dichotomy loss/gain of movement, both Danckaert (chapter ) and Ledgeway (chapter ) invoke different types of movement and constraints on it to deal with changes in Latin word order. But while Danckaert pinpoints a particular (indirect) trigger for syntactic reanalysis (namely the independent evolution of the negator non, from phrasal marker to head to clitic incorporated in T), Ledgeway argues for a fully macroparametric approach. As Latin widely displays surface 9 The acronym DiGS stands for the Diachronic Generative Syntax conference. The published DiGS volumes are: Battye and Roberts (); Van Kemenade and Vincent (); Pintzuk, Tsoulas and Warner (); Lightfoot (); Batllori, Hernanz, Picallo and Roca (); Crisma and Longobardi (); Jonas, Whitman and Garret (); Galves, Cyrino, Lopes, Sandalo and Avelar (); Biberauer and Walkden (); Mathieu and Truswell (), corresponding respectively to DiGS I, III, V, VI, VII, IX, VIII, XI, XII, XV. See http://personalpages.manchester.ac.uk/staff/george.walkden/digs/. 10 Cf. Roberts (), Ledgeway (), Wolfe (), and references therein, on the diachrony of V systems, as well as Kiparsky () on analogical extension in syntactic change, a perspective that diverges from Lightfoot’s () cue-based theory of syntactic change. 11 The ‘merge over move’ economy principle (Chomsky , ) does not seem to be of any help here. See Castillo, Drury, and Grohmann (), and Motut () for a critical appraisal of this economy principle.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
head-final word orders but both Danckaert and Ledgeway adopt Kayne’s () assumption that phrases are universally head-initial (when basic, non-derived, word order is considered), the discussion in both chapters deals with the type of movement operations to which Latin resorts to derive word orders that later will become unavailable. Grammatical constraints on movement (universal or parameterized) are considered to account for the loss of complement-to-specifier movement (Ledgeway) and the VOAux pattern (Danckaert). The success of this type of approach in handling the diachronic drift from head-final to head-initial orders may offer interesting independent support for Kayne’s antisymmetry theory (hence, in a sense, against a true head-initial/head-final parameter). Thus the study of word order change is also a means to test the descriptive adequacy and explanatory potential of competing analyses relative to the dichotomy basic/derived constituent orders, and related ‘movement’ issues.12
. Book structure and chapter outline The chapters in this volume are organized into four parts, motivated by the issues discussed above, namely: I Targets for movement (changes in the functional architecture of the clause); II Triggers for movement (changes in nature or stability); III Verb movement into the left peripheries; IV Types of movement and its constraints (word order change in Latin). An outline of the contents of each part and chapter is provided below. .. Part I—targets for movement The four chapters in the book’s Part I deal with word order change in different types of syntactic structures (coordination, relatives, embedded clauses, determiner phrases) A reviewer finds it surprising that this introductory chapter does not address the subject of ‘competing grammars’ (an E-language issue) and, more broadly, ‘the nature of optionality’ (not clearly an E-language or I-language issue), ‘given its centrality in the domain of word-order change’. There are two reasons for our option of leaving this subject out of the scope of the volume’s introduction. First, as the reviewer acknowledges, the essays in this volume do not discuss or make any significant contribution to further our understanding of ‘competing grammars and the nature of optionality’, a subject that is not specifically tied to the topic of word order. Second, it is our belief that if empirical investigation can lead to advances in the relevant respect, it will come from scrutiny into variation in present-day languages not from the study of syntactic change. What we will be able to learn from current language variation may of course be highly helpful to understand particular instances of diachronic variation. Consider word order. Often, what prima facie appears to be free variation turns out to be the syntactic expression of semantic, pragmatic or discourse distinctions. But the contrasts can be very subtle, so that elucidating them may require access to native speakers’ intuitive knowledge, which is generally not available to diachronic research. As for analyses based on the concept of competing grammars, there is an additional problem. Historical records will usually provide enough evidence for prestige grammars but not for non prestige ones (a classical example is the very limited set of sources for the study of Vulgar Latin, although Latin is extensively documented). A competing grammars hypothesis to deal with syntactic change will therefore be always faced with the problem of extremely lacunar evidence. Last but not least, current syntactic theory is not sufficiently constrained to clearly determine that a certain array of data cannot be derived by one single (internal) grammar. Therefore, we also do not have solid theoretical ground to evaluate between hypotheses based on the concept of competing grammars and other approaches to the nature of (apparent) optionality. 12
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso
and different language families, either focusing in a particular language (Hungarian, European Portuguese) or instead taking a broad cross-linguistic perspective (Early Indo-European languages and their descendants). The common feature to the four chapters is that they all account for changes in word order as the effect of a reconfiguration in clause structure, which either subtracts (chapters –) or adds (chapter ) targets for movement. When considered together, these four studies indicate that this type of syntactic change is not unidirectional, as the functional structure of the clause can both contract or expand over time in certain constructions. Chapter , by Moreno Mitrović, presents a study of word order change in coordinate constructions across a wide range of Indo-European languages. Early IndoEuropean languages had two available patterns of coordination at their disposal: one in which the coordinating particle was placed in first and another in which it was placed in the second position with respect to the second coordinand (so-called ‘Wackernagel effect’). Diachronically, the two competing configurations reduce to the one with the initially placed coordinator that all contemporary Indo-European languages display. This is accounted for as the result of the loss of ‘Wackernagel movement’ (i.e. encliticization to the element realized in first position) and the development of a lexicalized J(unction)-morpheme, which represents the conflation of two previously separate functional heads. The analysis succeeds in explaining the bimorphemicity signature of initial conjunctions by deriving the morpheme count as a fusional exponents of two functional heads. Chapter , by Adriana Cardoso, investigates syntactic change regarding the availability of split noun phrases in relative clauses in the diachrony of Portuguese. In earlier stages of the language, an element that is thematically dependent on the head noun (either as a complement or as a modifier) may not appear adjacent to it but in a relative clause internal position. In Contemporary European Portuguese noun phrase discontinuity also arises in relative clauses, but only with the modifier/complement in the rightmost position. The word order with the modifier/complement at the left periphery of the relative clause is not allowed. The change is explained as being due to the loss of a left-peripheral position for contrastive focus within relative clauses (and possibly other types of subordinate clauses). Hence, the contraction of clause structure and the concomitant loss of movement are taken to constrain the possibilities of phrasal discontinuity found in earlier periods. Chapter , by Julia Backsai-Atkari, examines word order variation and change in the high CP-domain of Hungarian embedded clauses containing the finite subordinating C head hogy ‘that’. It is argued that the complementizer hogy developed from an operator of the same morphophonological form, meaning ‘how’, and that its grammaticalization path develops in two steps. In addition to the change from an operator, located in a specifier, into a C head (specifier-to-head reanalysis), the fully grammaticalized complementizer hogy also changed its relative position on the CP periphery, ultimately occupying the higher of two C head positions (upward reanalysis). Other complementizers that could co-occur with hogy in Old Hungarian eventually underwent similar reanalysis processes. Hence the possibility of accommodating two separate C heads in the left periphery was lost and variation in the relative position of complementizers was replaced by a fixed order.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
Chapter , by Barbara Egedi, studies the determination and the distribution of possessive constructions from Old to Modern Hungarian. The grammaticalization of the definite article in well-defined contexts had structural consequences, the most salient of which is the emergence of a new strategy for demonstrative modification, which is called determiner doubling throughout the chapter. Word order variation arises due to the determiners’ interference with the possessor expressions at the left periphery of the noun phrase. The newly added demonstratives first adjoined to the determiner phrase in a somewhat looser fashion: their combination with the dativemarked possessors resulted in a word order specific only to the Middle Hungarian period (Demonstrative-Possessor). At a later stage, demonstratives got incorporated into the specifier of the DP, with the Possessor undergoing DP-internal topicalization, thus landing in a phrase-initial specifier position. This new structure displays the fixed word order Possessor-Demonstrative, as a result of the left periphery of the DP being expanded to accommodate a Topic position. .. Part II—triggers for movement The issue of what triggers specific instances of movement is an area of generative syntax where little progress has been achieved since the book Triggers was edited by Breitbarth and Van Riemsdijk (). The triggers of observed and well-established displacement operations are often unclear, which turns out to be especially unfortunate for diachronic research. The three chapters in the book’s Part II are relevant to the topic of ‘triggers’, even if this is not the main focus of any of them. These chapters suggest that a better insight into the nature of triggers we help us understand why the nature of the items moving into a certain functional position may change over time (chapter on Hungarian) or why, for example, a hypothetic topic feature shows diachronic instability when it is the trigger behind verb-second (chapter on varieties of English) but not when it is the trigger of movement with (clear) semantic-pragmatic import (chapter on Irish). Chapter , by Veronika Hegedűs, examines the distribution of verbal particles in Old Hungarian, and argues that despite the word order change from SOV to SVO in Hungarian, the particle-verb order did not change because the previous pre-verbal argument position was reanalysed as a pre-verbal predicative position where complex predicates are formed in overt syntax. Predicative constituents other than particles show significant word order variation in Old Hungarian, apparently due to optionality in predicate movement (while variation found with particle-verb orderings can be attributed to independent factors). It is proposed that after the basic word order was reanalysed as VO, internal arguments and secondary predicates could appear post-verbally and it was the still obligatory movement of particles that triggered the generalization of predicate movement, making all predicates pre-verbal in neutral sentences at later stages. This process involves a period of word order variation as predicate movement gradually generalizes to all predicate types. Chapter , by Judy Bernstein, examines verb-second (V) cross-linguistically in closely related varieties of English, namely: Older Scots, displaying general V; present-day Appalachian English and African American English, displaying residual V. Discontinuous subjects (analysed as instances of transitive expletives) and
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso
negative auxiliary inversion are shown to involve verb-movement to Focus in the two present-day varieties of English, unlike the general V found across Germanic, which the author assumes to involve TopicP. The area of overlap among V phenomena in the varieties of English studied is FocusP, which encodes the V associated with wh-elements in all three varieties (Older Scots distinguishes between Topic, for regular V and transitive expletives, and Focus, for wh-elements). It is suggested that the loss of general V might be understood as a change in triggers: ‘In some varieties of English, such as Appalachian English and African American English, Topic triggers may have given way to Focus triggers’. Chapter , by Cara DiGirolamo, deals with the interface between Syntax and Pragmatics by examining argument fronting in Old Irish non-poetic Glosses. The author argues, against previous literature, that Old Irish (a VSO language like Modern Irish) had a set of productive argument fronting positions with distinct information structural properties, which can be analysed in terms of a Rizzian articulated left periphery. These fronting positions are shown to be the direct ancestors of fronting positions in Modern Irish. Relying on lexical and contextual indicators of discourse function, three Information Structure patterns are identified: aboutness topic, contrastive topic and focus. Aboutness and contrastive topic appear to be both left dislocation structures. Focus is most commonly expressed through clefts, although clefts in Old Irish can be morphologically opaque. Modern Irish has all these structures besides a non-clefted focus structure (i.e. focus-fronting), which possibly originated in the Old Irish morphologically opaque clefts. Hence the author concludes that Irish maintained the pragmatics and syntax of its topic and focus structures, although simple focus fronting has become a more productive structure over time. .. Verb movement into the left peripheries The four chapters in the book’s Part III are quite diverse in their goals but all handle instances of verb-second at some stage of the languages investigated (Portuguese, Egyptian, Dutch, and Italian). None of the authors is concerned with the issue of triggers for movement and all assume invariable clause structure. Loss of verb movement into the left peripheries is understood as a side effect of other changes (chapters , ) or is just not the issue (chapters , ). Interestingly, chapter proposes that the loss of verb-second in Italian frees the past participle to move higher than the low (vP) left periphery, where in Old Italian it satisfied the V property of the language at the v-phase level. Therefore, the effect on word order of the loss of verb-second can, in certain cases, be the reverse of what might be expected. Indirectly, the issue of triggers comes to mind again. If the past participle must move to the aspectual low IP field in Modern Italian, as the authors assume following Cinque (), was the current trigger for movement inexistent in Old Italian or could it be satisfied in some other way? Chapter , by Charlotte Galves and Alba Gibrail, focuses on Classical Portuguese and its change to Modern European Portuguese, bringing to the debate new data concerning transitive sentences, with the aim to prove that Classical Portuguese was a verb-second language. The data are drawn from the Tycho Brahe Parsed Corpus of
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
Historical Portuguese (texts written by Portuguese authors born –). It is argued that both constituent order syntax and the information structure functions of word order in transitive sentences (SVO, VSO, VOS) support the characterization of Classical Portuguese as a verb-second language. Specifically, the authors argue that it is the fact that the verb occupies a high left-peripheral position in clause structure that makes a high position for post-verbal subjects available as well. This can explain why post-verbal subjects in Classical Portuguese are not obligatorily associated with an information focus interpretation, but may receive a familiar topic interpretation. The empirical evidence discussed in this chapter is intended to support the claim that there was a syntactic change from Classical to Modern European Portuguese, rather than a discursive reinterpretation of the same syntax. Chapter , by Chris Reintges and Sonia Cyrino, discusses the syntactic consequences of analyticization from a macroparametric perspective, focusing on the restructuring of the verbal tense system in two genetically unrelated languages: Coptic Egyptian and Brazilian Portuguese. The global effects of synthetic-to-analytic drift are thus examined in two different diachronic scenarios: one in which the process has almost, though not entirely been completed (Coptic Egyptian), and another one in which the process is still under way (Brazilian Portuguese). Coptic has gone very far in abandoning its former synthetic features and thus exhibits a high degree of analyticity. In Brazilian Portuguese, the analyticization process is in an advanced state, with synthetically inflected tenses exhibiting a decreasing productivity and gradually being replaced by the corresponding auxiliary verb constructions in the spoken language. The restrictions on verb movement observed diachronically in both languages are a side effect of ongoing analyticization, as the presence of auxiliaries or TAME particles in the IP space bars verb movement to these or higher positions. The shrinking of the available movement space leads to the exclusion of word orders previously derived by verb movement and may also diminish positional options of adverbial modifiers. Chapter , by Gertjan Postma, is a theory-informed quantitative corpus study of infinitival fronting in a (peculiar) type of Infinitival V construction found in OldFrisian and Middle-Dutch (and preserved in some coastal Dutch dialects until the nineteenth century). Based on the idea that infinitival fronting is the non-finite counterpart of the embedded subjunctive constructions (which the quantitative investigation confirms), the author proposes that both the loss of subjunctives and the rise of laten-support (the parallel of English do-support) are tied to changes in CP/TP (viz. ‘Rich verbal forms {may not/may} cross TP’). Concretely, before the fifteenth century, in Dutch, subjunctives and infinitives found in the relevant constructions move out of TP reaching C or Mod. In the second half of the fifteenth century, infinitives are being reanalysed as sitting in T. Hence, in infinitival fronting constructions, a separate verbal auxiliary form (laten) is created as a spellout of C. Although Laten-support is a transient phenomenon (‘failed change’), it has been the trigger of the reanalysis of auxiliaries as ordinary verbs in Dutch. Chapter , by Jacopo Garzonio and Cecilia Poletto, considers the distribution of VO and OV orders in Old Italian when the object is represented by a quantified constituent. The investigation takes into consideration cases of VO/OV variation
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Ana Maria Martins and Adriana Cardoso
with complex analytic verb forms where V is the past participle and O contains a universal or a negative quantifier. It is shown that while OV with non-quantified DPs and complex QPs is optional, universal bare quantifiers always precede the past participle. The authors propose that bare quantifiers undergo obligatory movement to a dedicated aspectual projection in the IP space, which is a function of their internal structure. Moreover, the authors argue that the modern stage of the language has preserved the movement of the quantifier, but this is not always visible because of a change in the movement properties of the past participle. In generalized verbsecond Old Italian, the past participle remains trapped inside the vP left periphery, while in Modern Italian it moves outside the VP raising to a position inside the aspectual layer of projections, which crucially is higher than the one targeted by bare quantifiers. .. Types of movement and constraints on it The two chapters in the book’s Part IV bring us back to the issue of directionality in syntactic change. Both chapters deal with Latin and use roll-up movement (i.e. movement from complement to specifier) to derive certain orders that characterize Latin syntax. But whereas chapter deals with the loss of roll-up movement (comparing Latin with Romance DPs), chapter deals with the emergence of roll-up movement (comparing VAux structures in Classical and Late Latin). So, again (as with the chapters in Part I), syntactic change with effects on word order is not unidirectional. Even a typologically less common type of movement, like complement to specifier, can either be lost or emerge over time. Chapter , by Lieven Danckaert, investigates the loss of the word order pattern ‘VOAux’ in Latin. This order was fully productive in Classical Latin, but in the Late Latin period (from until AD) the relevant pattern is only rarely attested, despite both of its ‘ingredients’, i.e. a head-initial VP (VO) and a head-final TP (VPAux) still being productive. The author proposes that in Classical Latin such clauses are derived by EPP-driven A-movement of the extended VP, targeting a position FP in the middle field, which is higher than TP and NegP. The decline of VOAux is then explained by postulating that VPAux-orders in Late Latin are reanalysed as being derived through roll-up movement, which is the result of an independent change in the grammar, namely the incorporation of the pre-verbal negator non ‘not’ into the T-head, through a process of procliticization. That is to say: in Classical Latin, negated clauses with an auxiliary and a non-finite lexical verb, which typically feature the order ‘VP-Neg-Aux’, provided the language learner with evidence that VPAux-clauses did not involve roll-up movement. Once the negator non incorporates into the verbal head in T, there is no longer any strong indications that a VPAux-clause does not involve roll-up movement. This opens the way for reanalysis. Exclusion of VOAux then follows from the Final-Over-Final Constraint (Biberauer, Holmberg, and Roberts’s ), which applies to roll-up movement and predicts that head-final projections dominate only other head-final projections. Chapter , by Adam Ledgeway, discusses under a macroparametric perspective discontinuous syntactic constituents produced by edge-fronting, which represent one distinctive feature of Latin regarding Romance. The author proposes that the head
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word order change from a generative syntax perspective
parameter can account for the observed contrast: whereas Romance is consistently head-initial, Latin fluctuates between different settings because it occupies an intermediate position in the gradual shift from head-finality to head-initiality. This difference in the head parameter is responsible for the observed variation in edgefronting, since its setting determines the application of antilocality in constraining movement. Concretely, if head-finality is the output of a roll-up operation raising the complement to the specifier to the left of its head, this indicates that suspension of antilocality must be allowed in head-final languages like Latin. In head-initial Romance languages, by contrast, antilocal movement never arises. This investigation thus derives from the different settings of the head parameter a concomitant parametrization in the role of antilocality.
. Closing up This collective volume provides through in-depth language-internal or comparative studies new perspectives on the relation between word order change and syntactic movement (the theoretical concept that ties the book’s different parts together). Each chapter in the book raises a number of stirring questions and issues that go well beyond those chosen as illustration in this introduction and should feed further research in the fascinating topic of word order variation and change. We hope this book will demonstrate that the generative theory of grammar is particularly well equipped to deal with the specific questions that a diachronic inquiry in this domain needs to address. The matters of ‘directionality of change’ and ‘triggers for movement’ present themselves as salient topics for further investigation.
Acknowledgements This volume contains a selection of papers presented at the fourteenth edition of the Diachronic Generative Syntax (DiGS) conference held in Lisbon (– July ). DiGS was hosted by the Center of Linguistics of the University of Lisbon (CLUL) and supported by funds provided by Fundação para a Ciência e a Tecnologia (FCT). We would like to thank these institutions as well as the organizers, the invited speakers, the participants and the scientific committee for making the conference a great experience. We are also grateful to the three anonymous reviewers whose comments and suggestions helped improve and clarify earlier drafts of this manuscript. The present volume was prepared within the research project WOChWEL— Word Order and Word Order Change in Western European Languages (FCT PTDC/CLE-LIN//).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Part I Targets for movement Changes in the functional architecture of the clause
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
2 Configurational change in Indo-European coordinate constructions M O R E N O M I T R O V IĆ
. Introduction This chapter investigates the word order change in Indo-European (IE) coordinate construction. Across the entire IE family, two morphosyntactic patterns of coordination are found as Agbayani and Golston () have investigated most recently. In one type of coordinate construction, the coordinator occupies the peninitial—that is, enclitic in second—position with respect to the internal (second) coordinand, while in another type, the coordinator is initially placed between any two, or more, coordinands, as the the minimal representative pair from Homeric Greek shows in (). Diachronically, the change from the two competing structures with peninitial and initial positions to the initial type is uniform across IE. () a. aspidas eukuklous laisēia te pteroenta shields round pelt and feathered ‘the round shields and fluttering targets’ b. keīs’ eīmi kaì antiō polemoio there go and meet battle ‘go thither, and confront the war’
(Il., book M: . )
(Il., book M: .)
The proposed synchronic analysis of the two coordinate structures, represented in (a) and (b), identifies two coordinate positions: I will show that enclitic (peninitial) coordinators occupy one of those positions, while the orthotone (initial) coordinators occupy both coordinator positions. By looking into the fine-grained structure of coordination synchronically in IE languages, a diachronic account resting on the feature-checking mechanism will suggest itself straightforwardly. The morphosyntactic
Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Moreno Mitrović . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Moreno Mitrović
change in word order patterns in coordination will be shown to not only have ramifications in terms of linearisation (change from peninitial to initial position), but to be tightly related to the semantics underlying the two positions. I show that the alternation between the two (a) and (b) constructions is not free and random but rather that it obeys the phasal ‘logicality’ of derivation. In the following section, I outline the theoretical foundations regarding the syntax of coordination that my analysis rests on. Section . then proceeds to present the Indo-European data, for which an analysis is proposed in Section .. Section . provides some conclusions and outlooks.
. From coordination to junction Once the binary, phrase-structure compliant, idea is laid out in §.., I theoretically and empirically motivate an enrichment of the this structure in §.. by appealing to a more fine-grained model of coordinate construction. .. Background assumption The syntactic structure for coordination is taken to be binary as most notably argued for by Kayne () and Zhang (). Earliest arguments for a binary-branching model of coordinate syntax go back to, at least, Blümel () followed by subsequent substantiations (see Munn for an exhaustive list of references). Following Kayne (), we will assume that coordinators are heads, merging an internal argument (coordinand) as its complement, and adjoining an external argument (coordinand) in its specifier, as per (). ()
&P XP
coordinand1
&0
YP
coordinator
coordinand2
In what follows, I motivate a revision of (): instead of one coordinator position, two are additionally proposed to accommodate some theoretical and empirical facts. .. An enriched structure ... Den Dikken’s J(unction) Assuming a binary branching structure for coordination (), Den Dikken (b) argues that exponents such as and and or do not in fact occupy the coordinator-head position as indicated in () but are rather phrasal subsets of the coordinator projection, with their origin in the internal coordinand. The actual coordinator head, independent of conjunction and/or disjunction, which originate within the internal coordinand, is a junction head, J0.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Configurational change in IE coordinate constructions ()
JP XP J0
coordinand1
andP
coordinator
and0
YP
(silent)
and
coordinand2
The core motivation for Den Dikken’s postulation of the silent presence of J0 is to capture the distribution of the floating either in English. As Myler () succinctly summarizes: () Den Dikken’s either is a phrasal category and can be adjoined to any XP as long as: a. XP is on the projection line of the element focused in the first disjunct; and b. XP is not of C category; and c. no CP node intervenes between either and the focused element in the first disjunct; and d. either surfaces to the left of the aforementioned focused element at PF. This characterisation of either predicts its floatation (optional height of adjunction in a more standard coordination analysis), which is, in Den Dikken’s (b: , exx. ,) words, either too high () or too low (). ()
a. John ate either rice or beans. b. John either ate rice or beans. c. Either John ate rice or beans.
() a. Either John ate rice or he ate beans. b. John either ate rice or he ate beans. ()
JP 〈either〉 …
XP (…) 〈either〉 …
J0
YP or …
0
Employing (in his words, the abstract head) J , Den Dikken’s account explains not only the either . . . or coordinate constructions but also the whether . . . or and both . . . and, which are unified under the structural umbrella of JP structure. Den Dikken (a: ) takes the head introducing the internal (second) coordinand not
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Moreno Mitrović
as the lexicalisation of J0 but as a phrasal category establishing a feature-checking relationship with abstract J0 instead. For an implementation of Den Dikken’s (a) J-system of coordination, see also Slade (), Mitrović (), and Szabolcsi (). An updated and enriched structure of () that I propose is given in (): ()
JP µP µ0
coordinand1
J0
µP µ0
coordinand2
There is no principled reason in his account according to which J0 would resist or be banned from lexicalisation. For Den Dikken, J0 is an abstract ‘junction’ category inherently neutral between conjunction and disjunction for which no overt evidence is provided since his account rests on J0 not being lexicalized. I take it as a reasonable hypothesis that there may be languages, which overtly realize this junctional component of coordination. In §..., empirical justification for () is provided. Section . will show that IE syntax of coordination was of the same type. ... Lexicalised J: Avar There are empirical arguments substantiating the finegrained (double-headed) structure for coordination (). The structure I propose for coordination supposes there are three heads involved (a J and two μ heads). Mutatis mutandis, the theory predicts that there may be languages that realise all three (J+ μ/μ) heads simultaneously. Avar, a northeast Caucasian language of Daghestan, provides such evidence.1 Avar allows coordinate constructions of the polysyndetic (Latin que/que) type (), which, according to the system proposed here, involves two overt μ heads and a silent J. () keto gi hve gi cat μ (J) dog μ ‘cat and dog’ Aside from the polysyndetic type, Avar also allows an English-like construction with an initially placed coordinator, which we take to be a phonological instantiation of J: () keto va have cat (μ) J dog (μ) ‘cat and dog’ The third and last type of construction allowed in Avar is most relevant for our purposes since it shows a union of phonological realisations in () and (). In this construction type, both μ heads as well as J are realised simultaneously.2 1
This novel data was provided by Kamil Sharapudinovich Ramazanov (p.c.). Other languages which allow ‘triadic exponence’ of two-argument conjunction include Hungarian, Bulgarian, Macedonian, and Ser-Bo-Croatian. For a detailed empirical overview, see Mitrović () and references therein. 2
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Configurational change in IE coordinate constructions
() keto gi va hve gi cat μ J dog μ ‘cat and dog’ There is currently no alternative syntactic model of coordination, which could explain the third () option without further stipulations. The fine-grained system in (), however, can not only handle () without any problem, it even predicts its existence.
. Indo-European Having motivated a fine-grained J-μ complex for coordinate construction, both theoretically and empirically, I now address the central concern of this chapter, the IE coordinate construction. The existence of two types of construction with respect to the pen/initial positioning of the coordinator does not only correlate with (i) the alternation in linear placement of coordinator but also (ii) the very morphological structure of the the two types of coordinators heading pen/initial constructions. In sections ..–.., we take each of the two (i, ii) properties in turn. .. Alternation in linear placement I start the discussion with a diachronic perspective on IE syntax of coordination, which shows linear alternation in coordinator placement. The earliest IE languages show that there existed two syntactic types of coordinate structures. One in which the coordinator occupies the initial, and another in which the coordinator occupies the peninitial position with respect to the internal coordinand. Klein (a, b) has show for Ṛgvedic that the alternation between initial and peninitial placements of the coordinator patterns with the category of the coordinands, whereby the peninitial (enclitic) coordinators generally cannot coordinate clauses which the initial coordinators can. Given in Table . is a clausal/subclausal conjunct distribution with respect to occurences with initial/peninitial coordinators in Ṛgvedic (see section . for an explanandum). T . Categorial distribution of sub/clausal conjuncts for Ṛgvedic pen/initial coordinators (N = , ; numbers from Klein a,b) Coordinator
Distribution
[+CP] conjuncts
[CP] conjuncts
utá (initial)
.%
.%
.%
ca (peninitial)
.%
.%
.%
For a semantic explanandum, see Mitrović and Sauerland (, ) and Mitrović (). I now turn to sketching the synchronic morphosyntactic status and diachronic decline of the double system of coordination in the three representative classical IE languages (Sanskrit, Latin, Greek). Ṛgvedic, the oldest variety of Indic, shows a
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Moreno Mitrović
double system of coordinate expression: there existed an initial (a) and a peninitial (b) conjunction marker: ()
Vedic Sanskrit: a. mā´ no mahā´ntm utá mā´ no arbhakám not us great and not us small ‘[Harm] not either the great or the small of us.’ (ṚV, ..ab) b. vā´yav īndraś ca cetathaḥ sutā´nāṃ vājinīvasū vayu Indra and rush..DL rich srength.bestowing ‘Vayu and Indra, rich in spoil, rush (hither).’ (ṚV, ..a)
While both initial and peninitial markers were part of the Ṛgvedic grammar of conjunction, the initial conjunction declines in the post-Vedic period, as the statistical data given in Table . and plotted in Fig. . show.
T . Development and loss of the double system of coordination in Indic PERIOD
uta
ca
archaic
.%
.%
early
.%
.%
epic
.%
.%
classical
.%
.%
medieval
.%
.%
late
.%
.%
Relative occurence (%)
100 80 60 40 20
uta
te La
al iev
al
ed
sic M
c
Cl as
Ep i
ly Ea r
Ar
ch
aic
0
ca
F . The loss of the double system of coordination in Indic
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Configurational change in IE coordinate constructions
The synchronically dual status of conjunction is also observable in Classical Latin, as the pair of examples in () show. ()
Classical Latin: a. ad summam rem pūblicam atque to utmost weal common and ‘to highest welfare and all our [lives]’
ad to
omnium all
b. vīam samūtem que life safety and ‘the life and safety’
nostrum [ . . . ] of us (Or., .VI.–)
(Or., .VI.–)
Diachronically, the peninitial marker is lost with the initial et becoming the predominantly single device for conjunction, as shown in Table . and Fig. .. The role of the bimorphemic and que-containing coordinator atque will become relevant in §.. and in §. where an analysis is provided. T . Grammatical change from first century BC to fourth century CE in the conjunction system of Latin: et, que, and atque et
que
atque
st c. BC
.%
.%
.%
th c. AD
.%
.%
.%
Relative occurence (%)
100 80 60 40 20 0 –1
0 et
1 2 Century
3
que
atque
4
F . Grammar of conjunction in Latin: et, que, and atque from first century fourth century CE
BC
to
Identical synchronic and diachronic pattern is found in Ancient Greek where the double system of coordination in Homeric () declines in the post-Homeric period, as shown in Table. . and Fig. .. See Goldstein (), or indeed Mitrović (), for details.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
()
Moreno Mitrović Homeric Greek: a. keīs’ eīmi kaì antiō polemoio there go and meet battle ‘Go thither, and confront the war.’ b. Aspidas eukuklous laisēia te pteroenta Shields round pelt and feathered ‘The round shields and fluttering targets.’
(Il., book M: )
(Il., book M: )
T . Grammatical change in the Greek conjunction system of Greek from eighth century BC to fifteenth century AD καὶ (N )
τε (N )
Σ(καὶ, τε)
καὶ (%)
τε (%)
th c. BC
,
,
,
.%
.%
th c. BC
,
,
,
.%
.%
nd c. AD
,
,
.%
.%
th c. AD
,
,
.%
.%
PERIOD
100
Relative occurence (%)
80
60
40
20
ϰαὶ
15th ad
2nd ce
5th bc
8th bc
0
τε
F . Grammar of conjunction in Greek: καὶ and τε from eighth century BC to fifteenth century AD
The syntactic duality of the double placement of the coordinator extends beyond the three classically representative IE languages. It is clear from examples in ()–()
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Configurational change in IE coordinate constructions
that IE had prepositive (a) and a postpositive (b) series of coordinators. The two types of configurations could be distinguished by positing that the peninitially placed (enclitic) coordinator induces some form of movement, either syntactically or postsyntactically, but that the difference lies only in the linearisation of the surface placement of the coordinator. The data in ()–() expands the empirical coverage beyond the three classical IE languages. In section .., I proceed to motivating a difference between (a) and (b) which extends beyond linear placement of the coordinator. ()
Avestan: a. uta mazdå huruϑma haoma and wisdom.M.SG.GEN increase.M.SG.NOM haoma.M.SG.VOC raose gara paiti grow..SUBJ.MID mountain.SG.M.LOC toward ‘And [thus] may you grow upon that mountain, O Haoma, [bringing] the increase of wisdom, [ . . . ].’
(YH, .)
b. yūžəəm aēibiiō ahurā aogō you..SG.NOM them.PL.DAT lord.M.SG.VOC strength.N.SG.ACC dātā aṣ ̌ā xṣ ̌aϑrəm cā give..PL.AOR.IMP truth.N.SG.INST power.N.SG.ACC and ‘o Lord, may you give strength to them through truth and that power [ . . . ]’ (YH, .) ()
Hittite: a. nu kán Mursilin kuennir nu eshar ieir nu and PRT Mursilis.ACC they.killed and blood shed..PL and Hantilis nahsariyatati Hantilis feared..SG.M ‘And they killed Mursilis and they shed blood and Hantilis was afraid.’ (BoTU, ..–) b. anšu.kur.ra.meš lú.meš.is.guškin ya humandan charioteers golden.grooms and all ‘Charioteers and all the golden grooms.’ (StBoT, .ii.–)
()
Old Church Slavonic: a. sŭvĕtŭ sŭtvori sŭ knęzi counsel.M.SG.ACC took..SG.AOR with princes.PL.INST svoimi i s Moravlęny POSS.REFL.INST and with the Moravians.PL.INST posla kŭ carju Mixailu sent..SG.AOR to tsar.SG.DAT Mixailu.SG.DAT ‘the Prince of Moravia took counsel with his Moravian princes and appealed to Emperor Michael, [ . . . ]’ (VC, 2) b. Azŭ že gljo vamŭ . . . I but tell..SG.PRES you.DAT ‘But I tell you . . . ’
(CM, Mt. :)
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Moreno Mitrović
()
Gothic: a. ak ana lukarnastaþn jah liuteiþ allaim neither on candle.DAT.SG and light.IND..SG all.DAT.PL þaim in þamma garda. it.DAT.PL in that.M.DAT.SG house.M.DAT.SG ‘Neither do men light a candle, and put it under a bushel’. (CA, Mt. :) b. (galaiþ in praitauria aftra Peilatus jah) came.PRET..SG in judgment hall.ACC.SG again P.NOM and wopida Iesu qaþ uh imma. called.PRET..SG Jesus.ACC said.PRET..SG and him.M.DAT.SG ‘(Then) Pilate entered into the judgment hall again, and called Jesus, and said unto him.’ (CA, Jn. :)
()
Old Irish: a. boí Conchuḃur ocus maithi was..SG.AOR Conchobar.M.NOM.SG and the nobles.PL.NOM UlaḋN iN nEmuin Ulstermen.M.PL.GEN in Emain Macha ‘Conchobar and the nobles of the Ulstermen were in Emain Macha.’ (CCC, .) b. ba ch ri Temrach COP and king Tara.GEN ‘And he was king of Tara.’ (Laws, .)
()
Tocharian: a. ṣerśkana ñi aiścer ce sisters.F.PL.VOC me give.Q.PRES.PL.IX these.M.OBL.SG pintwāt epe se ññissa śpālmeṃ alms.M.OBL.SG or who.M.SG.NOM than me better tākaṃ cwi aiścer be..PL.SUBJ him.M.SG.GEN give.PL.PR.IX ‘Sisters, will you give me these alms or will you give (them) to him who would be better than me?’ (TB; THT, .) b. mā empeles omskeṃ sac mā pe not terrible.M.PL.OBL evil.M.PL.ALL not and tampewātsesac powerful.M.PL.ALL ‘Not for the terrible, the evil, and not for the powerful’ (TA; PJ, b)
The enclitic series is generally and freely prone to reduplication. As Gonda () and Dunkel () note, a peninitial connective like *kwe is traditionally reconstructed with a twofold syntax: both monosyndetic (X Y *kwe) and bisyndetic, or indeed polysyndetic, (X *kwe Y *kwe) constructions are freely available in early IE languages, as the pairs in ()–() representatively show.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Configurational change in IE coordinate constructions
()
Vedic and Classical Sanskrit: a. dharma ca arthe ca kāme ca dharma/law.LOC and commerce.LOC and pleasure.LOC and mokṣe ca bharata ṛṣabha yad iha asti tad liberation.LOC & Bharata giant which here is..SG that anyatra yad na iha asti na tat kvacit elsewhere which not here is..SG not that anywhere ‘Giant among Bharatas whatever is here on Law, and on commerce, and on pleasure, and on liberation is found elsewhere, but what is not here is nowhere else.’ (Mbh, ..) ́ āṃ vājinīvasū b. Vā ý av īndraś ca cetathaḥ sutā n Vayu Indra and rush..DL rich strength.bestowing ‘Vayu and Indra, rich in spoil, rush (hither).’ (ṚV, ..a)
()
Homeric Greek: a. os ede tá te eonta tá te which were (=know.PLUP) the and exist.PART the and essomena pró te eonta exist.FUT before and exist.PART ‘That were, and that were to be, and that had been before.’ (Il., book A: ) b. aspidas eukuklous laisēia te pteroenta shields round pelt and feathered ‘The round shields and fluttering targets.’
()
(Il., book M: )
Classical Latin: a. iam tum tendit que fovet que already then pursue and favour and ‘Already then, she both pursued it and (also) favoured it.’ b. vīam samūtem que life safety and ‘the life and safety’
(Aen., .)
(Or., .VI.–)
The polysyndetic pattern of enclitic coordinators in (a), (a) and (a) seems to have carried an emphatic component, akin to the modern English emphatic conjunction with both . . . and. The same reduplicative pattern with emphatic/focal semantics is found in Old Church Slavonic (OCS), which survives in contemporary Ser-BoCroatian, among other modern Slavonic languages. It is OCS, and its diachronic descendants, that shows the independence of linear placement and semantic force behind the coordinator. Proto-Slavonic has independently syncretized the prepositive (initial atque-type) and postpositive (peninitial/enclitic que-type) coordinators but only lexically. As the OCS example in () shows, conjunctor i has both the conjunctive semantics of the initial atque-type coordinators in IE as well as the emphatic/focal semantics of the enclitic que-type coordinators. While the dual semantics (Mitrović, ) is retained in Slavonic, the morpholexical difference between the two classes of coordinators has been collapsed. We will return to the syntax of this collapse in §..
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Moreno Mitrović
In ()–(), (a) shows (reduplicative) polysyndetic coordination with emphatic/focal meaning, while (b) illustrates a monosyndetic construction ()
Old Church Slavonic: a. boite že sę pace mogǫštaago i dšǫ i tělo pogubiti fear but REFL rather which may and soul and body destroy ‘But rather fear that which is able to destroy both soul and body.’ (CM, Mt. :) b. bǫděte že mǫdri ěko zmjię i cěli ěko golǫbĭe be but wise as serpents and harmless as doves ‘Rather be wise as serpents, and harmless as doves.’ (CM, Mt. :)
()
Contemporary Ser-Bo-Croatian: a. Bojte se više onoga koji može i dušu i tijelo pogubiti fear REFL more that which may and soul and body destroy ‘Butratherfear thatwhich isabletodestroybothsoulandbody.’ (CM,Mt.:) b. budite dakle mudri kao zmije i bezazleni kao golubovi be therefore wise as serpents and harmless as doves ‘Rather be wise as serpents, and harmless as doves.’ (CM, Mt. :)
.. Morphemicity There is one additional, and for the present purposes crucial, fact distinguishing the initial and the peninitial types of coordinators. The difference also lies in the morphological structure of the two series. The generalisation I motivate in this subsection is that (i) peninitial coordinators found across the earliest IE languages are monomorphemic, and the (ii) the initial coordinators are bimorphemic, at least in their original form. Initially placed coordinators are bimorphemic and as such are decomposable synchronically or diachronically into two coordinators, each underlying a morpheme. Greek kai, for instance, derives from *kati, itself being a concatenation of *kwe + *te (Beekes : , Boisacq : ). Conversely, Indo-Iranian (Ilr.) uta comprises coordinator u + ta ( [CP [NP [N0 ]i ] [C0 . . . b. Relexicalization: [CP [NP D [N0 ]] [C0 . . .
Split noun phrases have also been accounted in terms of remnant movement. The normal instantiation of this configuration starts with the movement of an element X out of a constituent. Then, the whole constituent, which contains the trace of X, rises to its designated position, as is depicted in (). An analysis along these lines is suggested by Androutsopoulou () for Modern Greek, Sekerina () for Russian, Franks and Progovac () and Bašić () for Serbo-Croatian, among others. ()
[YP . . . Y . . . ti]j . . . Xi tj
Finally, based on the copy theory of movement (Chomsky ), according to which moved elements leave copies behind that are subsequently deleted, some authors claim that split noun phrases can be derived from the mechanism of distributed
7 In this context, imperfect splits correspond to noun phrases that have no well-formed source in a simple movement account.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Adriana Cardoso
deletion. Under this approach, it is assumed that a complete noun phrase is copied to the left and that the splitting results from the fact that the deletion operation may partially affect both the upstairs and the downstairs copies of the moved constituent. According to Fanselow and Ćavar (), it is the pragmatic structure that determines the occurrence of both continuous and split noun phrases. Turning now to the base-generation analyses, the main idea that underlines these approaches is that the parts of a discontinuous noun phrase are merged independently of each other in different slots of the sentence (Hale ; Jelinek ; Fanselow ). This hypothesis has been proposed for adjectives by Baker () and for PPs by Koster (), among others. Baker () claims that adjectives can be generated in any syntactic position (such as VP) as long as that position permits free adjunction. Koster (: ) argues that PPs introduced by of do not necessarily have their source inside a noun phrase. According to him, if this were always the case, (a) would have an ungrammatical source, as shown in (b). ()
a. Of the students in the classj I like [Mary tj] better than anyone else b. *I like [Mary of the students in the class] better than anyone else (Koster : )
Some objections that can be raised against the base-generation analyses are discussed in section ... .. Towards an analysis of remnant-internal relativization In this study, I assume that the adjunction analysis of RRCs cannot account for the properties of remnant-internal relativization. Apart from several problems that this derivation raises, the strongest objection is that remnant-internal relativization with the modifier/complement following the relativizer would involve lowering to a nonc-commanding position (Fiengo ) and hence should be rejected in view of such requirements as the Proper Binding Condition or the Empty Category Principle. Concomitantly, I assume that the base-generation analyses of discontinuous noun phrases cannot account for the properties of remnant-internal relativization. There are at least three objections that can be raised against these analyses. First, in contrast to Koster’s example in (), when remnant-internal relativization is involved, the head and the PP could have a grammatical source, as shown in (b) (repeated from (b)). () a. os livros que eu compus da philosaphia the books that I wrote of.the philosophy b. os livros da philosaphia que eu compus the books of.the philosophy that I wrote Second, the idea that the second split-parts function as event modifiers (being adjoined to the VP) rather than noun modifiers is not plausible. Although some modifiers can be related to entities and events, not all have this ability. For instance, in (), if the book of philosophy was written by me, it is the book that is of philosophy and not the event of writing. Thus, the modifier is clearly interpreted in relation to the noun (and not in relation to the event).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Discontinuous noun phrases in the diachrony of Portuguese
Finally, as already mentioned, it seems reasonable to assume that the semantic dependency between the head and its modifier/complement require these elements to be in a structural relation at some point at the derivation.
. The proposal In this section, I propose an analysis of remnant-internal relativization that combines the raising analysis of RRCs with the movement analysis of discontinuous noun phrases. However, a simple movement approach cannot account for some of the core properties of remnant-internal relativization described in section ... In particular, it cannot derive the cases where the first split-part contains the head associated with other elements, such as an adjectival modifier, as in () (repeated from ()). () da qual cousa ellas dizem que som hisentas e que nuc̃ a from.the which thing they say that are free and that never a pagarõ per príuelegio antigo que tẽem do papa it.CL paid.PL by privilege old that have.PL from.the pope ‘as for it, they say that they never paid it because they have an old privilege from the pope’ (th c., Martins : ) These complex first split-parts are problematic for simple movement approaches because they seem to involve non-constituent movement. Take, for instance, sentence (); if we assume that the PP originates within the DP in a structure like [DP D [NP N PP]] and that the adjectival modifier also originates within the DP as a specifier of a functional projection (Cinque , among others), then the problem is that there is no constituent that includes the head and the adjective but excludes the PP and that can undergo leftward movement, displaying remnant-internal relativization. The proposal that will be put forward in this section circumvents the abovementioned problem by assuming that the (apparent) movement of non-constituents can be taken as an effect of phonological deletion. This hypothesis gains strength when considered in the light of the insights of Bošković and Nunes () on the copy theory of movement. However, before going on to the proper analysis of remnant-internal relativization, it is useful to make a short excursus on (i) the relationship between focus and prosody and on (ii) the basics of the copy theory of movement. .. Focus and prosody Several authors (Costa , ; Frota , , among others) have proposed that prosody plays an important role in the identification of focus in CEP. This chapter adopts Zubizarreta’s (, ) view of the relationship between prosodic prominence and focus. Starting with a slightly impressionistic generalization, the crucial factor seems to be that a focused constituent always carries the nuclear stress. However, there is no one-to-one correlation between the subtypes of focus and the subtypes of stress: information focus is identified by neutral stress, whereas
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Adriana Cardoso
emphatic/contrastive focus may be identified by either emphatic or neutral stress (Zubizarreta : , ).8 Neutral and emphatic stresses are, in turn, assigned by different rules: the Nuclear Stress Rule (NSR) (see ()) and the Emphatic/Contrastive Stress Rule (E/CSR) (see ()). The NSR assigns prominence to the rightmost/lowest sentential constituent, whereas the E/CSR assigns prominence to any position. () Nuclear Stress Rule (NSR)9 Given two sisters Ci and Cj, the one lower in the asymmetric c-command ordering is more prominent. (Zubizarreta : ) () Emphatic/Contrastive Stress Rule (E/CSR) A word with contrastive stress must be dominated by every F[ocus]-marked constituent in the phrase. (Zubizarreta : ) In addition to these rules, Zubizarreta () proposes the Focus Prominence Rule (FPR) (see ()), which aims to capture the relationship between the focus-structure of a sentence and its prosody. The idea is that the focused elements in a sentence are marked with a feature [+F], whereas the presupposed ones are marked with a feature [–F]. The FPR dictates that the nuclear stress must target an [+F]-element. () Focus Prominence Rule (FPR) Given two sister nodes Ci (marked [+F]) and Cj (marked [–F]), Ci is more prominent than Cj. (Zubizarreta : ) The coexistence of the FPR and E/CSR in the grammar does produce any conflict: because the E/CSR assigns emphatic/contrastive stress to any element, no principle prevents an [+F]-element from receiving it. Some conflicts may arise, however, between the FPR and the NSR. The FPR may force the stress on a non-final position, whereas the NSR requires the stress to fall on the rightmost clausal position. According to Zubizarreta (), languages seem to vary in the way they solve this conflict. In languages such as English and French, defocalized [–F] elements are treated as extrametrical in the sense that they are skipped by the NSR. Other languages (such as Spanish and Italian) employ prosodically motivated movement (p-movement), which moves the post-focal [–F] elements out of the rightmost clausal position. Then, the NSR applies and puts the stress on the sentence-final position.
8
In CEP the ambiguity between information focus and emphatic/contrastive focus can only arise if the focused constituent is rightmost. Following Frota (, and other related work), I propose that the aforementioned ambiguity is eliminated by differences in peak alignment (or choice of pitch accent). See Cardoso () for more details. 9 Actually, this rule consists of two parts: one sensitive to asymmetric c-command relations (C-NSR) and the other sensitive to selectional ordering (S-NSR). Languages differ in the way in which the NSR applies; both C-NSR and S-NSR are active in German and English, but only C-NSR is active in languages such as Spanish and Italian (and Portuguese). Here, I consider only the part of the rule that is relevant to CEP.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Discontinuous noun phrases in the diachrony of Portuguese
.. The copy theory of movement This section briefly presents two approaches that are crucial for the analysis of remnant-internal relativization presented in this chapter: the version of the copy theory of movement proposed by Bošković and Nunes () and the analysis of post-verbal subjects in Serbo-Croatian put forth by Stjepanović (). ... Bošković and Nunes () Following Chomsky’s () approach to movement, where the raising of elements leaves copies behind that are subsequently deleted, Bošković and Nunes () (based on previous work by Bošković , , a,b; Nunes , ) propose a revised version on the copy-theory of movement, according to which traces (i.e. copies that are structurally lower in the syntactic representation) may be phonetically realized as the result of a copy deletion mechanism that works at Phonological Form (PF). The basic idea is that PF has a preference for pronouncing the highest copy of a chain, but a lower copy may be pronounced to avoid a PF violation. Technically, this statement requires further clarification. The first point in need of clarification concerns the preference for deleting lower copies. The explanation Nunes (, ) provides for this fact is based on an economy principle that prefers fewer applications of deletion in later computations of the phonological component. As J. Nunes (: ) puts it: Exploring the null hypothesis regarding the copy theory of movement, the above proposal thus takes the position that both heads of chains and traces should in principle be subject to phonetic realization. According to the logic of the proposal, there is nothing intrinsic to lower copies that prevents them from being pronounced. If Chain Reduction proceeds in such a way that only a trace survives, the derivation may eventually converge at PF. The fact that in most cases such a derivation yields unacceptable sentences is taken to follow from economy considerations, rather than convergence at PF. Since the highest chain link is engaged in more checking relations, it will require fewer applications of F[ormal]F[eature]–Elimination than lower chain links, thereby being the optimal candidate to survive Chain Reduction and be phonetically realized, all things being equal.
This brings us to the second point needing clarification: why the deletion of a lower copy in PF is just a preference and not the only option. The reason is clear: a lower copy can be produced to avoid a PF violation. Some of the PF factors that have been considered to induce the pronunciation of lower copies are stress assignment processes, intonational requirements, and morphological restrictions on identical elements. However, PF constraints may also block full copy deletion. In this case, a lastresort mechanism is admitted on the PF side: the so-called scattered deletion. It consists of the deletion of different pieces of different copies, as represented in (). ()
[X Y]i . . . [X Y]i
Scattered deletion has been successfully applied to a range of languages and phenomena, namely to cliticization in Bulgarian and Macedonian (Bošković ), participle-auxiliary order in Bulgarian (Lambova ), and split phrases (Fanselow and Ćavar ).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Adriana Cardoso
... Stjepanović () A concrete example of the implementation of Bošković and Nunes’ () version of the copy theory of movement is provided by Stjepanović () to explain different patterns of word order in Serbo-Croatian. Serbo-Croatian is characterized by great freedom of word order. Depending on the information structure involved, sentences containing a new information focus can exhibit the following word orders: SVO, SOV, OSV, and OVS. For the sake of illustration, consider the OVS word order in (b), which can be produced as an answer to the question in (a). () a. [Who is catching the mouse?] b. Mis̆a hvata mac̆ka mouse catches cat ‘A cat is catching a mouse.’ (Stjepanović : ) Stjepanović () shows that in sentences with the neutral intonation pattern the constituent bearing new information focus (such as mack̆ a ‘cat’ in (b)) follows elements that represent old information. In order to explain the final sentential position of focalized elements, the author proposes that (i) the subject moves in overt syntax even in the cases where it surfaces post-verbally; (ii) the subject surfaces post-verbally because a lower copy of the chain is pronounced in PF. The factor that Stjepanović takes to induce the pronunciation of the lower copy is sentential stress assignment. This implies that in the output of syntax sentences have a focus structure whereby each element is associated with an [F]-feature. For example, (b) would have the syntactic output in () (taken from Stjepanović : ). () [AgrSP mis̆a [AgrSP mac̆ka [TP mac̆ka hvata[AgrOP mis̆a hvata [V mac̆ka –F +F +F –F –F –F +F hvata [V hvata mis̆a]]]]]] –F –F –F Stjepanović additionally assumes that in Serbo-Croatian the NSR applies just after Spell Out, assigning prominence to the rightmost/lowest sentential constituent (Zubizarreta , ). If the element that receives the nuclear stress has an [+F]-feature, no problem arises. On the contrary, if the rightmost element is [–F], a conflict situation emerges between the NSR and the FPR. To resolve this conflict, Serbo-Croatian renders defocalized [–F] elements extrametrical for the application of the NSR. Under this approach, post-verbal subjects in Serbo-Croatian (as in (b)) are a result of lower copy pronunciation. This is due to the requirements on sentential stress assignment, which force the copy associated with the nuclear stress to be pronounced. If this PF requirement is not satisfied, the derivation does not converge. .. The derivation of remnant-internal relativization As already noted, to account for the remnant-internal relativization found in ESP, it is necessary to derive three distributional patterns (see ()–()), which will be discussed in sections ..., ..., and ..., respectively. Pattern I. Regular relativization The head and its modifier/complement are adjacent.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Discontinuous noun phrases in the diachrony of Portuguese
() os livros da philosaphia que eu compus the books of.the philosophy that I wrote Pattern II. Remnant-internal relativization (with a final modifier/complement) The head and its modifier/complement are not adjacent; the modifier/complement surfaces in the rightmost sentential position. () os livros que eu compus da philosaphia the books that I wrote of.the philosophy Pattern III. Remnant-internal relativization (with a non-final modifier/complement) The head and its modifier/complement are not adjacent; the modifier/complement surfaces in a non-final position, following the relativizer. () os livros que da philosaphia eu compus the books that of.the philosophy I wrote ... Pattern I Along the lines of the copy theory of movement, I assume that the head and its modifier/complement are merged together in the relativization site. Then, in order to satisfy the wh-feature on C, the entire noun phrase is copied and merged in [Spec, CP], leaving a copy behind. philosaphia]i]] () [os [livros da philosaphia]i [que eu compus [livros da the books of.the philosophy that I wrote books of.the philosophy Next, I consider, in line with Nunes ( and subsequent work), that the syntactic object in () cannot be linearized because it is not in accordance with Kayne’s () Linear Correspondence Axiom (LCA). The LCA dictates that at PF, an element cannot asymmetrically c-command and be asymmetrically c-commanded by the same element in a structure. Because the two instances of os livros da philosaphia ‘the books of philosophy’ are non-distinct, the verb compus ‘wrote’ is required to precede and be preceded by the same element. This induces a violation of asymmetry, cancelling the derivation. The deletion of copies may then yield the outputs in (a, b), with a single application of deletion, and the ones in (c–f) and (g–h) with two and three applications of deletion, respectively. ()
a. b. c. d. e. f. g. h.
[os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]]
If there were no convergence problems resulting from these reductions, the derivations would be eligible for economy comparison, and the derivations from (c) to
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Adriana Cardoso
(h) would be excluded for employing more operations of deletion than necessary. That is, economy principles would ensure that deletion applies as few times as possible. An economy-based explanation is also available to cancel the derivation yielding (b). Because the formal features associated with the higher copy have already been rendered invisible for PF upon checking, the deletion of the lower copy will employ fewer applications of deletion in later computations of the phonological component. Therefore, all else being equal, (a) is the most economical way of deriving (). Under these conditions alone, remnant-internal relativization would never be derived. The most economical derivation in (a) would always be preferred and, as a result, Pattern I would be the only option. ... Pattern II However, as in any economy approach, if the most economical option does not lead to convergence, a less economical option may be chosen, as in Pattern II. According to the mainstream version of the copy theory of movement adopted here, lower copies can be pronounced if there is an independent wellformedness PF requirement that precludes the pronunciation of a higher copy. The same line of reasoning applies to scattered deletion: if full deletion does not satisfy PF requirements, deletion may apply within different chain links. In the light of these assumptions, the derivation yielding remnant-internal relativization (with the modifier/complement in a final position) can also be a legitimate outcome. In this case, the phonological system resorts to scattered deletion, with part of the noun phrase pronounced in the higher copy and part in the lower one, as shown in (). () [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] the books of.the philosophy that I wrote books of.the philosophy I submit that this is the method to which the computational system resorts in order to accommodate the PF requirements in () and () (to be detailed in turn). () PF requirement I In non-corrective contexts,10 emphatic stress must be rightmost. () PF requirement II The EPP feature on C dictates that the relative head must be pronounced in the higher copy. A. PF requirement I Let me start by showing how Bošković and Nunes’ () version of the copy theory of movement can handle PF requirement I. Sticking to the principle that synchrony can inform historical approaches to languages, I presume that ESP are like CEP in that (narrow) information focus 10
Corrective contexts aim to set right a poorly transmitted or wrongly received part of a message. In this case, the emphatic stress can fall on any item: a phrase, a word or even an individual syllable of a word (Enkvist : ).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Discontinuous noun phrases in the diachrony of Portuguese
always appears in the rightmost position of the sentence11 (see Costa for CEP, and Martins for ESP). Moreover, I assume that ESP pattern with CEP in that an emphatic/contrastive focus in non-corrective contexts must either move to a dedicated left peripheral focus-position or be clause-final (just like narrow information focus); its location in a non-final position leads to ungrammaticality (see (b)–(b)).12 () a. Comi prego no prato raspado. ate.SG steak in.the dish scraped ‘I ate scraped steak in the dish.’ b. *Comi prego RASPADO no prato. ate-SG steak scraped in.the dish () a. Paguei de multa cem euros. paid of fine one.hundred euros ‘I paid a fine of one hundred euros.’ b. *Paguei CEM EUROS de multa.13 paid one.hundred euros of fine The contrast between the examples a and b is not surprising under the system developed thus far. As already mentioned, if emphatic/contrastive focus is assigned by prosodic prominence alone, it is freely assigned in corrective contexts but not in non-corrective ones. In the latter case, which corresponds to examples ()–(), the nuclear stress always targets the rightmost constituent (being assigned by the NSR). The PF facts just discussed nicely illuminate our general understanding of remnant-internal relativization, providing us with the tools to handle Pattern II (repeated here as ()). () [os [livros da philosaphia]i [que eu compus [livros da philosaphia]i]] the books of.the philosophy that I wrote books of.the philosophy First, recall that remnant-internal relativization emerges in non-corrective contexts where the second split-part is assigned emphatic/contrastive focus. In this case, emphatic/contrastive focus is marked prosodically via the NSR. Adopting Stjepanović’s () view of the relation between PF deletion and stress assignment, the deletion of the higher copy of da philosaphia in () can be explained if we assume that in the output of syntax () has the focus structure in ().
11 This is due to the fact that the (narrow) information-focused constituent bears a nuclear neutral stress assigned via the NSR (Zubizarreta , ). 12 Examples () and () are adapted from Martins (). Making use of the notational conventions in Zubizarreta (), the emphatic/contrastive focus identified by the E/CSR rule is indicated by small caps, whereas the (rightmost) emphatic/contrastive focus identified by the NSR is underlined (see Zubizarreta : , among others). 13 Example (b) would be possible in CEP if de multa ‘of fine’ were understood as a kind of afterthought, involving a prosodic contour with a pause between cem euros ‘one hundred euros’ and de multa ‘of fine’.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Adriana Cardoso
() [os [livros da philosaphia]i [que eu compus –F –F –F –F –F –F –F [livros da philosaphia]i]] –F +F +F In ESP, just like in CEP, the NSR applies just after Spell Out, assigning prominence to the rightmost/lowest sentential constituent (i.e. to philosaphia in ()). Since the element that receives the nuclear stress has an [+F]-feature, the FPR does not conflict with the NSR. As a result, the higher copy of da philosaphia is deleted, and the lower one is pronounced. Recall that a lower copy must be pronounced if it carries the nuclear stress. If this PF requirement is not satisfied, the derivation does not converge. B. PF requirement II Consider now PF requirement II, according to which ‘The EPP feature on C dictates that the relative head must be pronounced in the higher copy’. The question that has to be answered is why the head noun is not pronounced in the lower copy as well; in other words, why the more economical option of full copy deletion is not allowed and the last-resort mechanism of scattered deletion is used instead. Adopting as my point of departure Chomsky’s () view on EPP, I assume that core functional categories, such as v, T, and C, can have an EPP feature, which requires that their Specifier position be filled. This is the case of the relative C in CEP and in ESP. In this language, there are no head-internal relative clauses, a fact that clearly indicates that the relative C must have its Specifier position filled. Following Landau () (and much related work), I take the EPP to be a PF constraint that requires the Specifier of a functional category to be filled with an overt element at PF. This idea has been implemented in different ways in the literature. In line with Landau (), I will assume that (i) the EPP is a selectional feature governing PF configurations; (ii) it is locally satisfied by some element with phonological content; and (iii) the satisfaction of the EPP selectional requirement is not evaluated until PF.14 The approach to EPP just outlined constitutes a good basis for explaining why the head is pronounced in the higher copy in Pattern II: because the relative C has an EPP feature, failure to pronounce the head in [Spec, CP] would represent a PF violation. ... Pattern III To account for remnant-internal relativization constructions with the second split-part in a non-final position (as in (), repeated from above), I propose a remnant movement approach interpreted in terms of the copy theory of movement (Bošković and Nunes : ). () os livros que da philosaphia eu compus the books that of.the philosophy I wrote
14 It is worth clarifying that, under this view, the EPP does not trigger movement on its own. It is always parasitic on an anchoring feature (in this case the wh-feature), which triggers movement.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Discontinuous noun phrases in the diachrony of Portuguese
In order to derive this distributional pattern, I assume that the modifier/complement is interpreted as emphatic/contrastive focus, which is marked syntactically. Furthermore, I claim that in ESP there is a designated Focus position (see Rizzi and much related cartographic work) in the left periphery of dependent and nondependent clauses.15 Under this view, the (emphatic/contrastive) focused constituent (da philosaphia ‘of philosophy’ in ()) undergoes leftward movement to check a Focus feature in FocP.16 Then the noun phrase (containing a copy of the moved modifier/complement) moves to [Spec, CP] (or [Spec, ForceP] in the new terms) for wh-checking. Under these assumptions, at the output of syntax the structure of () would be as in ().17 ()
[os [livros da philosaphia]i [que [da philosaphia]j eu compus [livros [da philosaphia]j]i]]
The particular configuration dictated by remnant movement requires PF deletion to apply to two different chains: the PP chain formed by the movement of da philosaphia ‘of philosophy’ and the noun phrase chain formed by the movement of livros da philosaphia ‘books of philosophy’. Note that if the lower copy of these two chains were deleted (see ()), the resulting structure would not be linearized due to the presence of more than one copy of da philosaphia ‘of philosophy’. ()
[os [livros da philosaphia]i [que [da philosaphia]j eu compus [livros [da philosaphia]j]i]]
Adopting the representational hypothesis proposed in Nunes (), I assume that spell out sends the whole structure in () to the phonological component. Let me furthermore suppose that the functional category Foc has an EPP feature that requires its Spec to be filled with phonological material (and that failure to pronounce the material in Spec would cause a PF violation). Given these requirements, Chain Reduction inspects the PP chain and instructs the phonological component to delete the occurrence of da philosaphia ‘of philosophy’ that is a sister of livro ‘book’. As there are two elements that satisfy this description, Chain Reduction ends up deleting the two copies that satisfy this instruction, as represented in () (see Bošković and Nunes for more technical details). ()
[os [livros da philosaphia]i [que [da philosaphia]j eu compus [livros [da philosaphia]j]i]]
As for the noun phrase chain formed by the movement of livros da philosaphia ‘books of philosophy’, I adopt a solution similar to that proposed for Pattern II.
15 To account for the presence of a Focus projection in relative clauses, I assume, in line with Bianchi (), that the complementizer que ‘that’ is spelled out in Force and that there is a Focus projection below Force that hosts interrogative and focalized phrases in its Spec. 16 From a typological point of view, the internal movement of head-related elements is not surprising. Basilico () shows that in some languages internally-headed relative clauses display movement of the head to a sentence-internal position. 17 For simplicity, I am abstracting away other syntactic movements involved in this sentence.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Adriana Cardoso
According to this proposal, the EPP is a selectional feature governing PF configurations, which is locally satisfied by some element with phonological content. Because the relative C has an EPP feature, failure to pronounce phonological material in [Spec, CP] represents a PF violation. The higher pronunciation of the head noun is thus necessary to avoid a PF violation, as shown in (). ()
[os [livros da philosaphia]i [que [da philosaphia]j eu compus [livros [da philosaphia]j]i]]
. ESP and CEP in contrast In CEP, remnant-internal relativization is possible with the modifier/complement in the rightmost position, as in (a)–(a). The occurrence of the modifier/complement in the left periphery of the relative clause leads to ungrammaticality, as illustrated in (b)–(b). () a. uma filha que eu tenho pequena a daughter that I have young.F.SG ‘a young daughter that I have’ (Martins : ) b. *uma filha que pequena eu tenho a daughter that young.F.SG I have () a. uns touros que aqui tinha agrestes some bulls that here had.SG wild.PL ‘some wild bulls that I had here’ (TV-show Liga dos Últimos) b. *uns touros que agrestes aqui tinha some bulls that wild.PL here had.SG () a. uma casa que eu comprei de cinco assoalhadas a house that I bought of five rooms ‘a house with five rooms that I bought’ b. *uma casa que de cinco assoalhadas eu comprei a house that of five rooms I bought () a. um rapaz que eu conheço a boy that I know ‘a boy from Leiria that I know’ b. *um rapaz que de Leiria a boy that of Leiria
de Leiria of Leiria eu conheço I know
Somewhat tentatively, I would like to suggest that the contrast between CEP and ESP can be reduced to an independent change that took place in the history of Portuguese, namely the loss of a focus-dedicated position in the left periphery of embedded clauses. In ESP there is a focus position available in the left periphery of dependent and non-dependent clauses, whereas in present-day Portuguese this position might no longer be available in dependent clauses or, at least, in some types of dependent clauses. This hypothesis has to be worked out in future research.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Discontinuous noun phrases in the diachrony of Portuguese
Nevertheless, it is worth noting that it is consistent with the quantitative results of Martins (, ), which indicate that the frequency of preposed contrastive foci structures decreased significantly after the seventeenth century.
. Conclusion This chapter provides an account of remnant-internal relativization, showing how it fits in with the more general phenomenon of discontinuous noun phrases attested in ESP. From a theoretical point of view, the analysis developed throughout this chapter is that remnant-internal relativization provides new evidence for the raising analysis of RRCs. Diachronically, it is hypothesized that the contrast between ESP and CEP (i.e. the loss of remnant-internal relativization with a modifier/complement after the relativizer) can be explained by the loss of a focus dedicated position in the left periphery of embedded clauses.
Primary Sources Corpora P.S.: CLUL (ed.) (). P.S. Post Scriptum. Arquivo Digital de Escrita Quotidiana em Portugal e Espanha na Época Moderna. TYC: Galves, Charlotte, and Pablo Faria (). Tycho Brahe Parsed Corpus of Historical Portuguese.
Editions Coelho, António Borges (). Inquisição de Évora: Dos primórdios a . Vol. . Lisboa: Caminho. Martins, Ana Maria (). Documentos Portugueses do Noroeste e da Região de Lisboa: Da produção primitiva ao séc. XVI. Lisboa: Imprensa Nacional—Casa da Moeda. Piel, Joseph (). Livro dos oficios de Marco Tullio Ciceram, o qual tornou em linguagem o Ifante D. Pedro, duque de Coimbra. Edição crítica, segundo o MS. de Madrid, prefaciada, anotada e acompanhada de glossário. Coimbra: Universidade de Coimbra. Pimpão, Álvaro Júlio da Costa (4 []). Os Lusíadas, de Luís de Camões. Lisbon: MNE, Instituto Camões.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
4 The relative cycle in Hungarian declaratives J U L I A BA C S K AI - AT K A RI
. Introduction The aim of this chapter is to investigate the diachronic development of the Hungarian complementizer hogy ‘that’ and to account for its contribution to word order changes in the left periphery of various subordinate clauses. In Modern Hungarian (from the end of the eighteenth century), hogy typically introduces finite complement clauses and is located in the higher C head position (cf. Rizzi ). Furthermore, it also appears as part of certain complex complementizers, such as merthogy ‘because that’. By contrast, in Old Hungarian (ninth century– early sixteenth century) hogy had various other functions as well, e.g. hogy was the original comparative complementizer. The reverse order of present-day complex complementizers also existed, e.g. hogymert ‘that because’; also, hogy could potentially appear in relative clauses, together with relative operators (e.g. hogy ki ‘that who’). As far as the meaning of these combinations is concerned, it must be noted that it was transparent in most cases: hogy did not contribute to the meaning of the ultimate combination, i.e. for any complementizer combination hogy+X or X+hogy, the meaning was—originally—invariably that of ‘X’. The word order variation with respect to complementizers is summarized in Table .. As can be seen, hogy could be combined with other complementizers, namely ha ‘if ’, mert ‘because’ and mint ‘as/than’ and if a given combination existed in the XY order, then it seems to have invariably developed the mirroring YX order as well. In addition, the original meaning of such combination pairs was exactly the same. One may therefore be tempted to say that there was free word order variation in Hungarian complementizers in that Old and Middle Hungarian not only allowed the co-presence of two C heads but the order of these heads was not fixed. By contrast, in Modern Hungarian there seems to be no such variation allowed: out of the combinations given in Table ., only the ones marked in grey survive (i.e. hogyha ‘that if ’, merthogy ‘because that’, mintha ‘as if ’ and minthogy ‘than that’). Essentially, such an Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Julia Bacskai-Atkari . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
T . Complementizer word orders ha ‘if’
hogy ‘that’
mert ‘because’
mint ‘as/than’
–
hahogy
–
hamint
hogyha
–
hogy mert
hogymint
mert ‘because’
–
merthogy
mint ‘as/than’
mintha
mintogy
ha ‘if ’ hogy ‘that’
– –
– –
analysis would allow the co-presence of two C heads in one left periphery both in Old/Middle and in Modern Hungarian and the only difference between the two stages would be that while the former arrangement involved interchangeable C heads, in the latter configuration the relative positions of the individual complementizers are fixed. However, such a proposal would fail to account for the observation that though hogy is the last element in most combinations, it appears as the first one in hogyha, hence one would treat hogy as a higher C head in most cases but a lower one in this exceptional case. Furthermore, it should also be explained what constraint rules out free variation in Modern Hungarian but not in the previous periods. To avoid having to encounter these theoretical problems, I will show that the chief difference between Old/Middle and Modern Hungarian lies in the fact that while the former allowed the co-presence of two C heads, the latter does not. In turn, I will show that this restriction derives from the grammaticalization of all complementizers into higher C heads, hence these complementizers are no longer base-generated in the lower C position. In other words, Old/Middle Hungarian truly allowed two C heads in a single clause but the word order in these combinations was fixed—and in such a way that the relative position of the individual complementizers is predictable from the different timing of their grammaticalization into proper C heads. However, with the upward movement of lower C heads to the higher C position the reverse surface order of the combinations also came into being, resulting in the apparent word order variation between C heads. These latter configurations ultimately grammaticalized into single (but morphologically complex) C heads and are still preserved in Modern Hungarian. Hence the apparent free word order variation is actually a result of fully predictable grammatical processes. My proposal relies on two basic assumptions: on the one hand, I will show that hogy developed from an operator into a complementizer, which is in keeping with the general mechanism of the relative cycle (see, for instance, Van Gelderen ). On the other hand, I will argue that hogy became a general marker of subordination
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
in Old and Middle Hungarian. Both of these are of crucial importance in understanding word-order variation and change in the left periphery: depending on what stage of the relative cycle a given element—a complementizer or an operator—was in, it could occupy different positions in the left periphery. Hence both the absolute syntactic position of hogy and its relative position to other elements was subject to change. In addition, the fact that hogy appeared in a wide range of structures implies that it was likely to combine with several other complementizers too. In the following, I will show that the word order variation (hogy+C vs C+hogy) in Old and Middle Hungarian was mainly a result of the upward movement of the C heads other than hogy, and the fact that Modern Hungarian exhibits mostly C+hogy orders is a result of these combinations being fully grammaticalized. In other words, although the various orders at first sight may seem to be a result of free word order variation, it can be shown that both variation and change are fully predictable in terms of general economy principles. Though the present investigation will chiefly concentrate on Hungarian data, it has to be stressed that the observations are relevant also for a more general understanding of complementizer ordering and combinations.
. The data First of all, let us consider the basic data in terms of functional and structural differences between Modern Hungarian and earlier periods. Since hogy ‘that’ was by far the most significant complementizer taking part in combinations, the issue of why it was easily combined with other elements must be addressed and the answer lies fundamentally in the functional flexibility of hogy. Most functions of the complementizer hogy can be observed in Old and Modern Hungarian alike. First, the most basic function is that hogy introduces simple embedded declarative clauses, in examples like (): () a. Láttam, (hogy) esik az eső. saw.SG that rains the rain ‘I saw it was raining.’ (Modern Hungarian) b. & felkèlè hog o̗ hazaiaba mēnè o̗ kèt and up.rose.SG that she.homeland.POSS.SG.ILLATIVE go.COND.SG she.two menėvèl Moabitidiſnc̣ videkebo̗ l/ daughter-in-law.POSS.SG.COM Moab.DAT country.POSS.SG.ELATIVE Mert hallotta vala hog vr tèkėntėttė volna for heard.SG was.SG that Lord looked.SG be.COND.SG o̗ nėpėt & adot volna o̗ nèkic ėtkèkèt he.people.ACC and gave.SG be.COND.SG they.DAT dishes.ACC ‘Then she arose with her daughters in law, that she might return from the country of Moab: for she had heard in the country of Moab how that the Lord had visited his people in giving them bread.’ (Vienna Codex , th century)
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
As indicated, in this function the overt presence of hogy is optional: it alternates freely with a zero complementizer. Second, hogy may also introduce embedded imperatives, as in (): () a. Azt mondták, (hogy) menjek Portóba. that.ACC said.PL that go.SUBJ.SG Porto.ILLATIVE ‘They told me to go to Porto.’ (Modern Hungarian) b. & kèʒdec kėrni hog èltauoʒnec o̗ and began.PL ask.INF that off.leave.COND.SG he videkecbo̗ l country.POSS.PL.ELATIVE ‘And they began to pray him to depart out of their coasts.’ (Munich Codex ra, ) Again, hogy may freely alternate with a zero in this function. Third, hogy is also the complementizer (optionally) introducing embedded wh-interrogatives, resulting in sequences of hogy + an interrogative pronoun: () a. Azt kérdeztem, (hogy) mikor indulsz. that.ACC asked.SG that when leave.SG ‘I asked when you were to leave.’ (Modern Hungarian) b. valobiʒō o̗ n ko̗ ʒo̗ tto̗ c aʒ vtban veto̗ ko̗ dtec vala indeed self among.POSS.PL the way.INESSIVE contested.PL was.SG hog ki o̗ ko̗ ʒo̗ tto̗ c nagob volna that who he among.POSS.PL greater be.COND.SG ‘for by the way they had disputed among themselves, who should be the greatest’ (Munich Codex rb, ) Fourth, hogy introduces clauses of purpose: () a. Elmentem, hogy vegyek kenyeret. off.went.SG that buy.SBJV.SG bread.ACC ‘I went to buy some bread.’ (Modern Hungarian) b. & monda aʒocnac Meńńètec a rokon falucba / & and said.SG those.DAT go.SBJV.PL the cognate villages.ILLATIVE and varoſocba hog ot eſ p̄dical’l’ac mert arra towns.ILLATIVE that there also preach.SBJV.SG because that.SUBLATIVE io̗ ttèm came.SG ‘And he said unto them, Let us go into the next towns, that I may preach there also: for therefore came I forth.’ (Munich Codex ra, ) Note that in this use, hogy cannot be replaced by a zero; the same is true for its fifth function, which is that of introducing resultatives: in these structures, the matrix
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
clause contains a degree element (úgy ‘so’ or olyan ‘so’), which selects for a subclause headed by hogy. This is illustrated in (): () a. Mari úgy elesett, hogy két hétig kórházban Mary so off.fell.SG that two week.TERMINATIVE hospital.INESSIVE volt. was.SG ‘Mary fell so badly that she spent two weeks in hospital.’ (Modern Hungarian) b. & ſokan go̗ lekeʒenᶜ egbè / ug hog ſem a haʒba and many gathered.PL together so that neither the house.ILLATIVE ſem aʒ aitohoz nē fėrnėnᶜ neither the door.ALL not got.PL ‘And the multitude cometh together again, so that they could not so much as eat bread.’ (Munich Codex rb, ) In addition to these functions, which are shared between Old/Middle and Modern Hungarian, in Modern Hungarian there is a new environment, which is unattested in the previous periods: the introduction of embedded yes/no questions. In embedded yes/no questions, as illustrated in () below, the interrogative subclause invariably contains the question marker -e (usually attached to the verb), which is responsible for the marking of [+wh]; however, the overt complementizer hogy may also appear overtly, which shows that hogy is underspecified for the feature [wh], see É. Kiss (: , ). Consider: () Azt kérdeztem, (hogy) éhes vagy-e. that.ACC asked.SG that hungry be.SG-Q ‘I asked whether you were hungry.’ (Modern Hungarian) On the other hand, there is an obsolete function of hogy, which appeared in earlier periods but not in Modern Hungarian: hogy was the complementizer introducing comparative subclauses, in subclauses expressing inequality followed by the element nem ‘not’, giving the sequence hogy nem ‘that not’: () Mert iob hog megfog’doſuā algukmėg’ vrat because better that PARTICLE.catch.PTCP bless.SUBJ.PL.PARTICLE Lord.ACC hog nė mėg-hal’l’ōc that not PARTICLE-die.SUBJ.PL ‘because it is better that we should bless the Lord when we are caught than to die’ (Vienna Codex , th century) This function was present both in Old and Middle Hungarian. Apart from functional similarities and differences, the issue of combinations with other complementizers must also be addressed since this is the domain where complementizer word order variation may be observed. In Modern Hungarian, the following combinations are available: minthogy ‘than that’, merthogy ‘because that’ and hogyha ‘that if ’ (note that here I intend to give meanings that reflect the
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
T . Some changes affecting hogy Munich Codex ()
Káldi translation ()
Neovulgata ()
hogynem ‘that not’
–
–
hogyha ‘that if ’
hogy ‘that’ Ø alternates
morphological setup of the given combinations since this will be crucially important for the present investigation). Historically, however, the reverse order of all of these combinations existed, alongside the ones mentioned above, hence: hogymint ‘that than’, hogymert ‘that because’ and hahogy ‘if that’. Before turning to the detailed analysis of how hogy took part in word order variations in the left periphery, let us see some preliminary data that show basic facts about the history of hogy. The research I carried out was based on a small corpus analysis whereby I examined the Gospel of Mark in three translations: the Munich Codex () from the Old Hungarian period, György Káldi’s translation () from Middle Hungarian and the so-called Káldi-Neovulgata (), which is from Modern Hungarian. I examined altogether loci: in all of these cases at least one of the translations contained the complementizer hogy or a combination thereof.1 The most important results are summarized in Table .. As can be seen, the number of hogy substantially increased over the periods in question; moreover, the number of zero alternates in these loci also increased. Both follow from the fact that finite subordinate clauses became more frequent as opposed to non-finite clauses (cf. Haader ). Second, the combination hogynem ‘that not’ is present only in the Old Hungarian text but not in Middle or Modern Hungarian: this is in line with the fact that hogy was used as a comparative complementizer fundamentally in the Old Hungarian period and this function gradually came to be lost in later periods. Third, as can be seen, the combination hogyha ‘that if ’ was already present in Old Hungarian and survives until the Modern Hungarian period; hence at least some combinations of hogy originating in Old Hungarian are preserved in the language.
. Grammaticalization and the relative cycle In order to provide an adequate syntactic analysis for the changes affecting hogy ‘that’, let us first discuss the notion of the relative cycle, as introduced by Van Gelderen (), following the ideas of Hopper and Traugott () or Heine and Kuteva (), among others. The relative cycle is a grammaticalization process Note that I found instances only of the combinations hogynem ‘that not’ and hogyha ‘that if ’, hence there are no data for other combinations that were or are possible otherwise in a given period). I will return to the changes attested for complementizer combinations in section .. 1
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
whereby a pronoun first becomes an operator moving to [Spec,CP] and subsequently this operator is reanalysed as the head of that CP. In addition, there is a further possibility in terms of the grammaticalization of C heads, which is the reanalysis from lower C to higher C. Both this process and the relative cycle are attested for English that, as discussed by Van Gelderen (). The same is true for Hungarian hogy; the processes are summarized in () below: ()
CP C’ CP
C hogy
hogy
C’ C
hogy As can be seen, hogy was originally an operator moving to the lower [Spec,CP] position—at this stage, the meaning of hogy was ‘how’ (cf. Juhász : –, : , –, ; Haader : –, : –).2 In line with the general mechanism of the relative cycle, this operator was reanalysed as a complementizer, i.e. instead of an element moving to the C-domain it was reinterpreted as an element base-generated there, hence as a (lower) C head.3 Second, from a lower C head it was reinterpreted as a higher one, which is responsible for marking the Force of the clause (cf. Rizzi ).4 In order for operators to develop into complementizers they have to have features that are compatible with C heads; in Hungarian, for instance, C heads are not allowed to have person and number features and hence ordinary relative operators (e.g. aki ‘who’) cannot be reanalysed as C heads, while operators such as hogy could since they had essentially the same features as complementizers. Note that both steps of reanalysis are motivated by economy, which can be summarized in the form of two principles, as described by Van Gelderen (): the Head Preference Principle (HPP) and the Late Merge Principle (LMP), both originally going back to the idea that Merge is preferred over Move (cf. Chomsky
Evidence for this comes from early Old Hungarian data, see section .. It is worth mentioning that reanalysis does not necessarily mean the complete loss of a previous function: it is possible to have a split between the original function and the newer one. This is also the case for hogy, which is preserved as an interrogative operator ‘how’ even in Modern Hungarian (with the possible alternate hogyan); the relative operator, however, developed a distinct form with a- during Late Old Hungarian and Early Middle Hungarian, hence Modern Hungarian has the forms ahogy and ahogyan. 4 Note that while the analysis of Rizzi () suggests that there is a clear-cut difference between the higher and the lower C (Force and Fin, respectively), it is evident from Rizzi’s work that this is not entirely the case: in fact, all Force heads identified by Rizzi () are unambiguously associated with [+finiteness]. Furthermore, it is also possible to mark clause type lower than the highest CP node, as described by Rizzi () for interrogatives. Given this, I will refer to the two CPs as higher and lower CP rather than Force and Fin. 2 3
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
). The HPP states that being a head is preferable to being a phrase—hence the reanalysis from operator to complementizer. The LMP states that it is more economical to be base-generated in a higher position than to be moved to that position— hence the reinterpretation of the original lower C as a higher one. The reason behind this latter step is simply that it is the higher C head that is responsible for defining the Force of the clause and the fact that certain overt lower C heads become associated with carrying Force implies that these elements also start moving up to the higher C head. This again leads to a choice between movement and base-generation at a higher point in the structure—and just as in the case of the HPP, the latter configuration is preferred. Note that the preference of Merge over Move follows from the general principles of language acquisition, as described by Roberts and Roussou (: –): the structure involving Merge is simpler both in derivational and in representational terms, since the element in question does not demonstrate feature syncretism. Therefore, when an element X can potentially be analysed as merged at a given point, or as merged at a lower level and moved subsequently higher up, the former option will be preferred by the language learner, as it is more transparent. In this sense, the HPP and the LMP are merely epiphenomenal results of the actual driving forces underlying grammaticalization processes; still, since there are some typical paths that grammaticalization processes follow (as described by Roberts and Roussou in detail), the HPP and the LMP are useful notions for the description of the various changes that can be observed in the history of a language.
. Simplex complementizers Before turning to the actual complementizer combinations and word order variations in the left periphery, let us first briefly discuss the changes affecting other complementizers, which is necessary to understand how these could be combined with hogy ‘that’. In addition to hogy, Hungarian has three other complementizers that also developed by way of the relative cycle; these are ha ‘if ’, mint ‘as/than’ and mert ‘because’. All of them were originally operators (cf. Juhász : –, : , –, ; Haader : –, : –). The operator function of hogy from Early Old Hungarian is illustrated in (): () furiſcte muſia!|| etetý ýmletí. ug hug ana ſciluttet. bathes washes feeds breastfeeds so how mother child.POSS.ACC ‘she bathes, washes, feeds and breastfeeds him as a mother does her child’ (Königsberg Fragment, th century) However, it is important to stress that there are crucial chronological differences between the individual operators. The functional split between operator and complementizer for hogy and ha took place mostly in Proto-Hungarian: there are very few instances of ha used as an operator (and this function gradually disappeared altogether from the language), and while the operator hogy is attested in Old Hungarian too (and even later), the complementizer use can be detected already in
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
the early documents. Furthermore, there is also reason to believe that they were already higher C heads as complementizers (though hogy could rarely be a lower one as well): this is shown by their relative positions in complementizer combinations (they come first in base-generated orders), as will be shown in the next section, and they also precede relative operators (see section .); in addition, hogy also preceded the negative polarity head (located in between the two C heads) in comparative subclauses (see Bacskai-Atkari : –). By contrast, for mint and mert the split took place only during the Old and Middle Hungarian periods and hence these were either operators in the lower [Spec,CP] or were already base-generated in the lower C head as complementizers. The possible Old Hungarian positions for present-day complementizers are summarized in () below (note that ha and hogy had already split from the original operator uses by early Old Hungarian, hence these operator uses are not included here): ()
CP C’ C
CP
mint ha hogy mert
C’ C
…
mint mert (hogy) As can be seen, the various present-day complementizers could take various positions historically; ultimately all of them came to be base-generated in the higher C head position.
. Multiple complementizers Due to the difference in the positions of complementizers mentioned above, the prediction is that it should be possible to have two overt complementizers—in predictable order—within one left periphery as long as the positional differences still held, i.e. in Old and Middle Hungarian. This prediction is borne out: it was possible to have a higher C head combined with a lower C head, ultimately deriving from original combinations of a higher C head and an operator located in the lower [Spec,CP] position. The example in () shows the combination of hogy ‘that’ with mint ‘than’: () edesseget erze nagÿoban hogÿmint annak elo̗ tte sweetness.ACC felt.SG greater that.than that.DAT before.POSS..SG ‘(s)he felt sweetness even more than before’ (Lázár Codex , after )
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
The possible structures for such combinations are shown in (): ()
CP
CP
C’
C’
C hogy
CP mint
C hogy
C’ C Ø
CP
…
C’ C
…
mint
The left-hand diagram shows the configuration where the higher C head is headed by an overt complementizer (here: hogy) and the specifier of the lower CP is filled by an operator (here: mint), while the head of the lower CP is not filled by overt material. The operator was later reanalysed as the head of the lower CP, as shown in the right-hand diagram. Hence examples such as () have either of these structures underlyingly. While it is often difficult to decide whether a particular example shows an operator or a complementizer use of mint or mert, there are clear indicators that the operator use was lost towards the end of the Old Hungarian period, as mint and mert started to behave differently from ordinary relative operators (such as ki ‘who’). On the one hand, complementizers took part in head movement (see section .), while relative operators did not. On the other hand, ordinary relative operators developed distinct forms from their interrogative operator counterparts, all relative operators starting with a- (hence aki ‘who-Rel.’ versus ki ‘who-Int.’); however, this did not affect mint and mert, which indicates they were categorically distinct from ordinary relative operators by the end of the Old Hungarian period. Given that all complementizers had their respective typical positions in the left periphery, the expectation is that this should have a bearing on the possible word orders. This is indeed so: the word orders of C+C combinations—as well as combinations of a higher C head with an operator—were fixed. Since hogy was typically in the higher C head, it normally appeared as the first element in C+C sequences, hence the combinations hogymint ‘that than’ and hogymert ‘that because’. On the other hand, ha ‘if ’ was invariably a higher C head; consequently, in combinations with hogy, ha was the first element, hence hahogy ‘if that’. I will return to this issue in section .; for the time being, suffice it to say that the word order of complementizers in combinations is fully predictable on the basis of their typical positions in the left periphery.
. Complex complementizers The question arises how the reverse word orders can be accounted for, given that the complementizer word order in C+C combination was fixed. As has been said, lower
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
C heads were ultimately all reanalysed as higher C heads: this naturally involved the upward movement of these original lower C heads. Movement was motivated by the preference for marking the Force of the clause on the higher C head: however, complementizers first had to be reanalysed from operators, which are (and were) always located in the lower [Spec,CP] in Hungarian, hence complementizers had to appear in the lower C head first. While operators may indeed be the overt markers of clause type, they are not responsible for defining the Force of the clause, and hence are not necessarily associated with the higher CP cross-linguistically: as far as Hungarian is concerned, they are in the lower CP in all periods, and they do not take part in upward movement, unlike lower C heads, which were preferably interpreted as not only overt markers of clause type but also as Force heads in language acquisition. The fixed order of original C+C combinations (as described in section .) hence stems chiefly from the fact that certain complementizers grammaticalized later than others, which is attested independently from combinations, too. The upward movement of lower C heads was possible also when the higher C head was filled by another overt complementizer. If so, the lower head joined the higher one via adjunction: however, adjunction took place so that it produced the reverse linear order of the two heads, due to the Linear Correspondence Axiom (Kayne ; but cf. also the Mirror Principle of Baker , ). Hence in the case of an original C+C combination such as hogymint ‘than that’, the complex C head is of the form minthogy ‘than that’: () semi nagob nem mondathatik: mint hogh lego̗ n istenek nothing greater not say.PASS.COND.SG than that be.SBJV.SG God.DAT ania mother.POSS ‘nothing can be said to be greater than that she be the mother of God’ (Tihanyi Codex , ) The relevant structures of complex complementizers are shown in (): ()
CP
CP
C’
C’
C
CP
C
CP
minti hogy
C’
minthogy
C’
C ti
…
C
…
Ø
The left-hand diagram shows the actual derivation of complex complementizers from multiple complementizers: the lower C head moves to the higher one (just as
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
when it is a single overt C head) and adjoins to it from the left, resulting in the reverse of the original linear word order. The right-hand diagram shows the result of grammaticalization: complex complementizers are base-generated as a single— though morphologically complex—unit, which produces a derivationally simpler structure than the one involving movement. Since, as has been argued for, the word order of C+C combinations was fixed, it is predictable that the order of C heads in a complex C is fixed as well; hence there are combinations such as minthogy ‘than that’ and merthogy ‘because that’, as well as hogyha ‘that if ’ (note that I will return to the particular issue of hogyha in the next section). The variation in the linear word order between C+C combinations and complex C heads, however, is no longer attested in the language. The reason behind this is that Modern Hungarian no longer has the possibility of accommodating two separate C heads in one left periphery: this is because all complementizers have been reanalysed as higher C heads and consequently there is no overt complementizer to be base-generated in the lower C position. Naturally, this should rule out the original C+C combinations, which is indeed the case: these are all extinct. On the other hand, the reverse order combinations are preserved: these are morphologically complex units base-generated as single heads.
. The position of hogy Let us now further investigate the role of hogy ‘that’ in word order variations in the left periphery. As should be obvious from the discussion so far, the underlying order was typically of the form hogy+X (X referring to a complementizer other than hogy); this was so because hogy was typically in the higher C head. Consequently, as all complex C heads represent the reverse order, generally it is the combinations of the form X+hogy that remained in the language up to Modern Hungarian. The only exception to these generalizations is the case of hogy and ha ‘if ’. Here the underlying order was ha+hogy, as demonstrated by the fact that intervening elements could also potentially appear between the two heads: () Ha késen hogy el nyugot az nap, hamar eso̗ t váry if late that off set.PST.SG the sun soon rain.ACC expect.SBJV.SG ‘if the sun has set late, expect rain soon’ (Cisio of Cluj-Napoca. G, ) As can be seen, there is an adverb (késen ‘late’) appearing between the complementizers ha and hogy, which are located in one and the same left periphery: it is still just one conditional clause that could not be potentially analysed as two separate clauses. It is true that Hungarian does not generally prefer elements (i.e. topics, foci) to appear as high as between the two C positions, and this seems to be valid for earlier periods as well (see É. Kiss a on the evolution of the clausal left periphery). Still, the fact that it is possible to have an element there clearly indicates that ha and hogy must be in two distinct positions. Hence in combinations with ha, hogy was originally base-generated as a lower C head, contrary to other combinations. The explanation behind this is simply that
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
T . The distribution of multiple complementizers Source
hogymint ‘that than’ hogymert ‘that because’ hahogy ‘if that’
Jókai Codex around tokens:
–
Bod Codex first half of the th century tokens:
–
–
Székelyudvarhely Codex (first quarter) – tokens:
–
–
Kazinczy Codex – tokens:
–
–
TOTAL
ha was always in the higher C head and hence hogy could only be base-generated in the lower C head, which was still an available option for hogy (see section .). However, since typically hogy was also a higher C head already, movement generally took place, and thus the complex C head showing the reverse order (hogyha ‘that if ’) was more frequent even in Old and Middle Hungarian than the base-generated order (hahogy ‘if that’). Regarding the frequencies of the individual combinations discussed here, I conducted a corpus search based on the normalized part of the Old Hungarian Concordance corpus.5 The results are summarized in Tables . and . below.6 Table . shows the distribution of multiple complementizers involving hogy, see section .. Table . shows the distribution of complex complementizers involving hogy, see section .. The number of occurrences is higher for multiple complementizers in the case of hogy and mint only; even for hogy and mert, the complex complementizer structure is more frequent. This is in line with the hypothesis that once an original operator is reanalysed as a complementizer, it preferably moves up to a position that is associated with Force. The difference is most striking in the case of hogy and ha: there 5 The corpus is available from http://omagyarkorpusz.nytud.hu/en-intro.html, where the metadata (date, token numbers) for the individual texts are also available. As of December , the normalized part of the corpus includes the texts given in Tables . and ., and some shorter texts that contain no instances of complementizer combinations (many of these shorter texts do not contain complex sentences at all). Note also that the corpus is restricted to Old Hungarian only, and currently there is no searchable Middle Hungarian corpus at all. 6 The token numbers always refer to the normalized version without punctuation marks.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
T . The distribution of complex complementizers minthogy ‘than that’
merthogy ‘because that’
hogyha ‘that if ’
Jókai Codex around tokens:
–
Vienna Codex middle of the th century tokens:
–
–
Birk Codex tokens:
–
–
Festetics Codex – tokens:
–
–
Guary Codex before tokens:
–
Czech Codex tokens:
–
–
Booklet tokens:
–
Bod Codex first half of the th century tokens:
–
–
Miskolc Fragment tokens:
–
–
Song on the Conquest of Pannonia around tokens:
–
–
Praise of Saint Nicholas tokens:
–
–
Kazinczy Codex – tokens:
–
TOTAL
Source
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
are no instances of hahogy in the normalized corpus, while hogyha is far more frequent than any other complementizer combination dealt with here, indicating that the frequency of the combination of ha and hogy in itself cannot be a reason for the lack of hahogy. However, if one takes into account that hogy was preferably a higher C head already (unlike mint and mert), it follows that the preference for the complex complementizer order was more significant than in the case of the other two combinations.
. Relative clauses One of the most compelling questions in terms of the word order variations observed in connection with hogy ‘that’ is whether the analysis presented so far can be extended to other combinations involving hogy in the left periphery. A possible extension line is that of relative clauses, which could contain the sequence of hogy + a relative operator in Old and especially in Middle Hungarian (cf. Galambos ); however, the reverse word order (i.e. relative operator + hogy) is not attested. Hence the analysis should be able to account for the lack of surface word order variation here. Consider the example for the combination hogy + operator in () below: () olÿaat tezo̗ k raÿtad hog kÿto̗ l felz such.ACC do.SG you.SUPERESSIVE that who.ABL fear.SG ‘I will do such a thing on you that you are afraid of ’ (Sándor Codex , th century) As can be seen, the structure contains hogy besides the relevant form of the relative pronoun ki ‘who’; the representation is given in (): ()
CP C’ C hogy
CP kÿto ˏl
C’ C
…
∅ The higher C head is filled by hogy and the specifier of the lower CP hosts the relative operator. Note that this configuration is identical to the left-hand diagram in (), that is, the configuration when a higher C head co-occurs with an operator that will later be reanalysed as a C head. However, in the case of relative operators such as ki, there is no such reanalysis: these operators never developed into complementizers since they never lost, for example, their person and number features that are incompatible with complementizers in Hungarian—on the other hand, the option was available for those operators that had essentially the same features as complementizers.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
A further consequence of the lack of reanalysis in this case is of course that while sequences such as hogy ki ‘that who’ did occur, there are no instances of the reverse order, i.e. *ki hogy ‘who that’: since these operators were not reanalysed as C heads, they did not (and could not) take part in head movement from lower C to higher C either.7
. Further combinations So far I have mainly been concerned with combinations involving two CP projections. Another point where the analysis can be extended is combinations that include negative-like Pol heads, which appear between the two CPs. The prediction is that if these combinations show surface word order variations (and changes), then these should be in line with the general mechanisms attested in original C+C combinations. The base-generated C+Pol+C combinations were hogynemmint ‘that not than’ and hogysemmint ‘that neither than’. Consider the following example containing hogynemmint: () az mentól alsobÿkban is tob angÿal uagon honnem mÿnth az the more down.INESSIVE also more angel is that.not than the napnak feneben sun.DAT light.POSS.INESSIVE ‘there are more angels in the basest one of them than in the sun’s light’ (SándK. v) Assuming that the negative-like element nem is a Pol head that marks the negative polarity of the comparative subclause (cf. Bacskai-Atkari ; on the negative polarity of comparative subclauses, cf. Seuren ; Gergel ; Matushansky ), the structure of the left periphery containing hogynemmint or hogysemmint should be the one given in () below:8 7 Note that the motivation behind the upward movement of C heads is the marking of Force in the highest CP; since operators do not define Force, there is no motivation for an operator located in the lower [Spec,CP] to move further to the higher [Spec,CP] either. 8 It has to be stressed that the negative-like element in comparatives like () is merely a polarity marker, and does not imply the presence of clausal (predicate) negation. Hence such combinations are different from Modern Hungarian nehogy ‘lest’, which contains a negative element (ne) and hogy: such purpose clauses are indeed instances of clausal negation. Note also that nehogy is not attested in Old Hungarian (see Haader : ; accordingly, a corpus search on the normalized part of the Old Hungarian Concordance corpus, as of June , did not yield any instances of nehogy either). It is highly unlikely that the combination involves the present-day declarative complementizer hogy, since nehogy in negative purpose clauses can be preceded by hogy (the presence of which is optional, just like in positive purpose clauses, see () in section .):
(i) Elküldtem neked a kiállítás meghívóját, (hogy) nehogy lest off.sent-SG you-DAT the exhibition invitation-POSS.ACC that lemaradj róla. it-DELATIVE miss-SBJV.SG ‘I sent you an invitation to the exhibition, lest you should miss it.’
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
()
CP C’ C
PolP
hogy
Pol’ Pol
CP
nem/sem
C’ C
…
mint Hence there are two CPs and in between the two there is a PolP headed by the negative-like element nem ‘not’ or sem ‘neither’; while the higher C head is responsible for defining the Force of the clause, the overt, unambiguous marking of negative polarity and [+compr] is associated with lower functional heads. This configuration is no longer present in Modern Hungarian, which is predictable considering the assumption that complementizers are no longer base-generated in the lower C head. However, the reverse order combination survives in the form of mintsemhogy ‘than neither that’, which can be derived from the underlying order in the following way: ()
CP C’ C
PolP
minti semj hogy Pol’ Pol
CP
ti,j
C’ C
…
ti
Given these differences, I will not venture to provide an analysis for nehogy here, but since nehogy seems to be a single syntactic element below the C head expressing negation, it is highly unlikely that nehogy features a Pol+C combination.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
The mechanism represented in () is in line with the proposal argued for in the present chapter. That is, the lower C head mint ‘than’ first moves up to the intervening Pol head, filled by sem and adjoins to it from the left, resulting in the combination mintsem ‘than neither’. As a second step, the complex mintsem moves up to the higher C head since mint ultimately targets this position as the final landing site; the complex again adjoins to hogy ‘that’ from the left, giving the final combination mintsemhogy, which is thus the reverse order of the underlying configuration. Again, this complex head grammaticalized as such and hence is preserved even in Modern Hungarian.9
. Functions of hogy Finally, let us briefly examine the changes in the functions of hogy ‘that’, as reflected by its role in complementizer combinations. It should be obvious that hogy was readily combined with several other elements (either C heads or operators) and thus played a crucial role in word order variations and changes in the left periphery. The question is what allowed for the high potential of hogy to appear in several types of subclauses and combinations. Recall first the data from Table ., which data show that the frequency of hogy gradually increased over the periods in question, and while certain combinations (e.g. hogynem ‘that not’) disappeared from the language, others (e.g. hogyha ‘that if ’) survive until Modern Hungarian. The major changes in the functions of hogy can be summarized as follows. Originally, in addition to introducing various embedded clauses that survive into Modern Hungarian, hogy had the specific function of introducing comparative subclauses, which was lost alongside the appearance of mint ‘than/as’. On the other hand, hogy became the marker of subordination, which is also indicated by the fact that in Modern Hungarian its appearance is extended to embedded yes/no questions.10 The significance of marking subordination was in line with a general increase in the number of finite clauses. Table . showed that the number of clauses containing hogy—and its zero alternant—increased; I examined all the loci in question in all the three texts and looked at the various possible constructions that could stand instead of hogy-clauses in the Munich Codex and in Káldi’s translation. The results are summarized in Table .. 9 Note that the same mechanism is not attested in the case of nem: this rather cliticized onto the preceding element hogy, and the resulting combination (hogynem, which was after phonological assimilation honnem) was probably interpreted as a complex head itself (for further details, cf. Bacskai-Atkari , : –). While hogy was not an unambiguous overt marker for comparative Force, hogynem was, and it seems likely that there was a trigger for the upward movement of mint in the former case but not in the latter. Since the further discussion of this issue would require an investigation of comparatives, I will not venture to examine the question here any further. 10 Just as in the case of clause typing and finiteness, the marking of subordination also has a preferred relative position in the left periphery, which is the highest functional projection. The order of the three is identified as Sub > Force > Fin by Haegeman (). Since I do not wish to adopt a strict cartographic approach, I do not label CPs as any of these more specific categories, but the analysis presented here is compatible with the general theory.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Julia Bacskai-Atkari
T . Structures standing for finite subclauses Munich Codex ()
Káldi translation ()
coordination (és ‘and’)
participial clause
infinitival clause
mert ‘because’
–
nominal expression
As can be seen, in Old Hungarian the significance of non-finite structures, as well as coordination and phrasal—i.e. mostly nominal—equivalents of finite subclauses was high and it decreased in the Middle Hungarian period (cf. Haader ). The possibility of having mert ‘because’ as a simple declarative subordinator was absent already from the Middle Hungarian period (cf. Haader : ). The hypothesis that hogy became a general marker of subordination is supported by two phenomena. First, it appeared in a wide range of clauses, i.e. conditionals, clauses of reason, relative clauses, hence it does not seem to have been restricted to specific functions. Second, the meaning of a combination hogy+X or X+hogy did not initially differ from the meaning of X on its own (X standing for any complementizer different from hogy and also for ordinary relative operators), which indicates that hogy did not induce any semantic change by its appearance and thus its function was merely that of marking subordination. The fact that hogy was fundamentally a subordination marker also explains why it was able to appear in a wide range of structures and, consequently, why it frequently took part in surface word order variations in the left periphery.
. Conclusion The aim of this chapter was to provide an analysis for the diachronic development of the Hungarian complementizer hogy ‘that’ and to show that the word order variations and changes attested in connection with it can be explained by considering the general grammaticalization processes responsible for the changes affecting all complementizers in Old and Middle Hungarian. As was shown, hogy was reanalysed from an operator into a C head, the process of which can be traced in the case of other Hungarian C heads too. Regarding the combinations of various complementizers, it was seen that the order in base-generated combinations follows from the differences in when each C head grammaticalized from an operator. In turn, the reverse order combinations can all be derived from the base-generated ones via head adjunction, which predicts that the linear order will be reversed. These complex complementizers fully grammaticalized and they remain in the language even in Modern Hungarian, while C+C combinations disappeared because complementizers are no longer base-generated in the lower C head.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The relative cycle in Hungarian declaratives
Though all of these changes affect the entire complementizer system as such, the importance of hogy in this respect must be stressed since most combinations involved this particular element: this, as was seen, is because hogy was a general subordinator head that thus had the possibility to appear in a wide range of structures and to combine with other elements.
Acknowledgements This research was funded by the German Research Fund (DFG), as part of the research project The syntax of functional left peripheries and its relation to information structure (BA /–), and formerly also the Collaborative Research Centre Information structure: The linguistic means for structuring utterances, sentences and texts (SFB-).
Textual sources Old Hungarian Concordance corpus. URL: http://omagyarkorpusz.nytud.hu/en-intro.html: Birk Codex. . Bod Codex. First half of the sixteenth century. Booklet on the dignity of the apostles. . Cisio. Cluj-Napoca . Czech Codex. . Festetics Codex. –. Guary Codex. Before . Jókai Codex. Around . Kazinczy Codex. –. Königsberg Fragment and its Ribbons. Middle or second half of the fourteenth century. Lázár Codex. After . Miskolc Fragment. . Munich Codex. . Praise of Saint Nicholas. Written by Ferenc Péchy. . Sándor Codex. The first quarter of the sixteenth century. Song on the Conquest of Pannonia. Written by Demeter Csáti. Around . Székelyudvarhely Codex. –. Tihanyi Codex. . Vienna Codex. Middle of the fifteenth century.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
5 Word order change at the left periphery of the Hungarian noun phrase BARBARA EGEDI
. Introduction This chapter aims to give an account of the change that can be observed in the structure of the Hungarian noun phrase from the first written sources to the present day, with a special focus on determination and possessive constructions, i.e. the nominal left periphery. Analysing the historical data, a possible scenario will be offered as to how the functional extension of the noun phrase develops in a language, apparently inevitably, after the overt marking of definiteness grammaticalized and its use expanded. Three steps (or states) of the so-called definiteness cycle can clearly be identified from Old Hungarian to modern times: the appearance of the article in welldefined contexts, the emergence of a new type of demonstrative modification, and the reintegration of the adjoined demonstrative pronouns into the DP domain. Word order variation arises due to the determiners’ interference with the possessor expressions at the left periphery of the noun phrase. Following this general introduction, the essential characteristics of the noun phrase in present-day Hungarian will be discussed in section ., relying on current generative analyses. In the subsequent section, the noun phrase in Old Hungarian will be described and its structure reconstructed. This section also reviews the historical language stages of Hungarian to provide the fundamental background information about the sources on which this study has been built. Finally, through the presentation of some peculiar Middle Hungarian word order phenomena, an overall reconstruction of the structural change at the nominal left periphery will be proposed and, accordingly, the word order change from Demonstrative-PossessorPossessum into Possessor-Demonstrative-Possessum will be accounted for. The study concludes by summarizing the word order variation and change thus far discussed in a schema of the various patterns. Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Barbara Egedi . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
. The noun phrase in Modern Hungarian The literature on the Hungarian noun phrase is rather rich: a detailed analysis in a generative framework can be found in Szabolcsi (), which has been modified to some extent by É. Kiss () and Bartos (, ). Of course, not all aspects of the noun phrase structure will be discussed here; several features, such as the inner agreement properties, or the derivation of attributes and quantifiers, will be ignored. The focus of this study primarily falls on possessive constructions and demonstratives, and their interaction at the left periphery of nominal constructions. .. Demonstratives Demonstrative modifiers behave rather interestingly in Modern Hungarian, that is to say, they are not prototypical deictic determiners. Demonstratives co-occur with the definite article, and unlike other Hungarian modifiers, they agree in number and case with the head noun. ()
a. ez-ek a könyv-ek this-PL the book-PL ‘these books’ b. az-t a könyv-et that-ACC the book-ACC ‘that book’ c. az-ok-at a könyv-ek-et that-PL-ACC the book-PL-ACC ‘those books’
According to the above-mentioned generative analyses there are good reasons to assume that the demonstratives (ez/az ‘this/that’) are located in the specifier position of the DP projection (), rather than being adjoined constituents, even though the definite article is also spelled out in the head of the same phrase. This goes against the well-known economy principle established for functional projections, and still there are good arguments for a doubly filled DP analysis: nothing can intervene between the demonstrative and the article, the construction is not grammatical without spelling-out the article, and the two determiners also form a prosodic unit (cf. Bartos , ; as far as I know this placement of the demonstrative modifier was first pointed out by Kenesei : ): ()
a. [dp ez/az [d’ a [NumP három [np könyv]]]] b.
DP Spec
D’ D
NumP Spec
ez/az a három this/that the three ‘these/those three books’
NP könyv book
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
As will be shown at a later point in the discussion, the conditions, however, were slightly different when this pattern arose in the Middle Hungarian period. .. Possessives In Hungarian, there are two types of possessor expressions: a dative-marked (a) and a nominative/unmarked (b)—with no actual difference in meaning. (The gloss ‘POSS’ in the examples stands for the suffix on the head noun that encodes its being possessed.) ()
a. a szerzetes-nek a könyv-e the monk-DAT the book-POSS b. a szerzetes könyv-e the monk book-POSS
The syntactic positions of the possessor expressions are claimed to be different. This is suggested by two main features (as has been pointed out by Anna Szabolcsi : –): (i) by the presence/absence of the article between the two members of the constructions; and (ii) by the fact that dative-marked possessors can be extracted from the noun phrase. This latter means that the dative-marked possessor can appear outside the noun phrase, disconnected from the possessed noun by other constituents of the sentence, as a kind of external possessor. Unmarked possessor expressions cannot be extracted in this way. Following Szabolcsi’s analysis, É. Kiss (: ) claims that the possession marker on the noun licenses the possessor as an argument. This possessor-argument is realized either as a dative case marked complement, or as a pre-nominal specifier bearing no case at all. The caseless possessor can only survive if it has a [+determiner] feature and moves to the DP domain to have it checked. The presence/absence of the article follows from the difference between the two constructions. The caseless possessor in (b) moves directly into Spec,DP and, consequently, the head of the DP cannot be filled by the article. Actually, we never find a definite article in the constructions of type (b).1 On the other hand, the possessor in (a) has to leave its post-nominal complement position for the so-called Case Constraint2 and must undergo noun phrase internal topicalization, landing in a phrase-initial specifier position (Spec,TopP). This derivation can be observed in (). Note, that the notion of a DP-internal Topic phrase was introduced by Katalin É. Kiss in her paper only, while in her earlier works this movement operation was considered to be an adjunction to DP. 1 The asymmetry is obvious: in the case of lexical possessors the DP cannot be doubly filled, while in the case of demonstratives the simultaneous filling of the head and the specifier even seems to be required, as has been presented above. This phenomenon remained an unsolved contradiction in the generative literature on present-day Hungarian. Even though it shows to be a challenge in the synchronic description, however, viewing the problem from a diachronic perspective may shed some light on the question as to how this asymmetry emerged (see section .) 2 The Case Constraint is formulated as follows: (i) the case suffix must cliticize to the right edge of the noun phrase; (ii) the case suffix cannot cliticize to a case marked stem. These constraints can only be satisfied if a post-nominal complement is extraposed, or if it is realized in the form of a modifier (É. Kiss : ). Not only are possessive constructions subject to this rule, but any nominal expressions that can take a argument will meet the case.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase ()
TopP Spec
(After É. Kiss 2000:139) DP
D
PossP NP
Poss N Jánosnaki a -e könyv John-dat the -poss book
KP ti
In this case, the definite article freely co-occurs with the possessor, and so does the demonstrative, since the dative-marked possessor is located in Spec,TopP, rather than in Spec,DP. It is important to note, and it also follows from the derivation in (), that if both a demonstrative and a possessor are present in the construction (), the possessor must always precede the demonstrative which is located in Spec,DP. (b) is therefore ungrammatical. So is (c) with its unmarked/caseless possessor expression, since this latter competes with the demonstrative for the same structural position (SpecDP). The structure in () corresponds to the grammatical word order exemplified in (a). ()
a. a szerzetes-nek ez a könyv-e the monk-DAT this the book-POSS ‘this book of the monk’ b. *ez a szerzetes-nek a könyv-e this the monk-DAT the book-POSS Intended meaning: ‘this book of the monk’ c. *a szerzetes ez a könyv-e the monk this the book-POSS Intended meaning: ‘this book of the monk’
()
TopP Spec
DP Spec
a szerzetes-neki ez the monk-dat this
D’ D
PossP
a the
könyv-e ti book-poss
. The noun phrase in Old Hungarian Before turning to the historical data, a few basic facts are to be mentioned with respect to the nature and the date of the sources which are used throughout this
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
study. The following table summarizes the most important information related to the historical linguistic stages of Hungarian: the time intervals and the types of sources we have from the various periods. T . Historical language stages of Hungarian and types of sources Proto-Hungarian
Old Hungarian
BC– AD no written documents
Early Old – Hungarian
sporadic records, glosses, a few short texts, e. g. Funeral Sermon and Prayer (c.)
Late Old – Hungarian
mainly codices; translations of Latin religious literature + original Hungarian compositions
Middle Hungarian
–
Book printing New secular genres Significant increase in quantity
Distinguishing between the Early and the Late Old Hungarian period is relevant because of the nature of the sources that come from these two periods. The first written sources can be dated to the Early Old Hungarian period, but these are sporadic records, mostly names of people and places and other glosses embedded in Greek or Latin documents and charters. The first continuous texts from the same period (dating from between the end of the twelfth century to the end of the thirteenth century) are short and do not form a uniform corpus either in time or in space. One of the four records is fragmentary and two of them are not even narrative. Accordingly, the first texts of considerable length and suitable for a syntactic investigation come from the first half of the Late Old Hungarian period, in the form of the first Hungarian codices.3 These, for the most part, contain translations of Latin religious literature,4 but they are long enough, each forming a closed, uniform corpus of its own. .. Demonstratives and articles The grammatical encoding of referential identification in the Old Hungarian period appears to comprise a different system, as compared to the one used at the present time (for a detailed study on this topic, see Egedi ). Modern Hungarian makes 3 For more details about the first continuous texts from the Early Old Hungarian period (beginning with the Funeral Sermon and Prayer) and about the earliest codices, see Egedi (a: –). 4 The possible influence of the source language on Hungarian word order phenomena has been kept in mind throughout this research. Moreover, it has been concluded for the language of the Old Hungarian codices in general, that the influence of Latin can only be detected in terms of quantity rather than in terms of quality. This means that if Hungarian grammar allowed the use of alternative structures to express a certain content, the early translators were inclined to choose constructions that appeared to be closer to the original text. No doubt, some of the passages read quite unnaturally in a stylistic sense, but the grammar itself seems not to have been affected by the Latin syntactic structures. (The issue has been summarized in a paper written in Hungarian: Egedi b.)
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
extensive use of the definite article. As can be observed in examples () and (), the article obligatorily occurs together with demonstratives and with a dative-marked possessor. The corresponding Old Hungarian examples, however, look rather different since the definite article is absent in the same contexts. Consider () and () to see the contrast. Modern Hungarian ()
ez/az a könyv this/that the book ‘this/that book’
()
a szerzetes-nek a könyv-e the monk-DAT the book-POSS ‘the book of the monk’
Old Hungarian ()
e(z)/a(z) ø könyv this/that book ‘this/that book’
() a(z) szerzetes-nek ø könyv-e the monk-DAT book-POSS ‘the book of the monk’ As a result of my earlier investigations, it may be concluded that the marking of pragmatic definiteness on a syntactic level is already obligatory at the time of the first codices, but the definite article only appears in the cases where the referential identification is not encoded otherwise. Therefore, among other contexts, it is absent with nouns modified by a demonstrative, which encodes directly accessible reference, and in possessive constructions, in which the referent of the possessed noun is existentially presupposed and is identified through its relation to the referent of the possessor.5 To put it differently, the presence of a demonstrative () or a possessor expression () can imply the definiteness of the noun phrase as a whole, and the use of the article is not yet required. The situation was changing quite rapidly, although not in an equal manner in all the possible contexts, as recent research has demonstrated (Egedi and Simon ).6 An automatic query carried out in five normalized codices from the Old Hungarian Corpus showed that the use of the definite article proportionally increased already within the Old Hungarian period (cf. Table .), but the spreading can only be detected in certain contexts (such as generics or before pronominal possessors),
5 There are detailed arguments for the Old Hungarian definite determiner functioning as a true article and being a fully grammaticalized category encoding pragmatic definiteness in the already cited paper (Egedi ). The identification of definite contexts in which the article is still absent is also discussed there. 6 The results of the query presented in have been updated several times, which slightly modified the numerical data. Table . has been synchronized with the table that can be found in Egedi (a: ).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
T . The proportion of definite determiners (a/az) in five Old Hungarian codices Codex
Date
Tokens
a/az
Jókai Codex
after /c.
,
Vienna Codex
after /c.
,
,
.
Guary Codex
before
,
,
.
Booklet
,
.
Kazinczy Codex
–
,
,
.
% .
while the contexts relevant here (nouns with demonstratives and lexical possessors) resisted the determiner all along the period. Constructions such as those illustrated in () and () only appear during the Middle Hungarian period and therefore will be discussed in the next section. It is to be noted that the Hungarian definite article developed from a demonstrative modifier, as also happened in many other unrelated languages, but in this early period these two elements, the distal demonstrative and the definite article cannot be distinguished merely on formal grounds. The modern reader sometimes remains uncertain how to interpret an Old Hungarian phrase such as (a–b) because the article and the distal demonstrative modifier look exactly the same, have the same position and also show some functional overlap (e.g. anaphoric use). ()
a. az the
kapu gate
b. az kapu that gate Depending on the context, the morpheme ‘az’ sometimes seems to behave as a demonstrative, sometimes as a definite article, and often it is impossible to decide between the two options.7 The only exceptions to this ambiguity are the clear semantic or pragmatic contexts in which only an article can appear, such as the larger situational use, or the associative-anaphoric use, in terms of Himmelmann (). Despite this homophony, the article and the distal demonstrative can be assumed to occupy distinct structural positions, D and Spec,DP, respectively. The emergence of the definite article may be reconstructed similarly to what Giuliana Giusti (: ) proposes for the reanalysis of the Latin determiner ille in the Romance
7 Hence it is not completely unexpected that there is a tendency in the descriptive tradition of Old Hungarian to consider these determiners as ‘pre-articles’, or ‘pronoun-articles’, representing a transitional word-class with dual nature (Bakró-Nagy : ; I. Gallasy : –). For the same problem in Old English, see, among others, Sommerer () with further references.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
languages. As is illustrated in (), ille, originally located in the specifier of the Determiner Phrase as a demonstrative, is reinterpreted as an element in D. ()
a.
DP DemP (il)le
b. ⇒
D’ D
DP Spec
…
D’ D
(il)le
…
This kind of reanalysis perfectly fits with one of the economy principles formulated by Van Gelderen (, b), the so-called ‘Head Preference Principle’, according to which speakers prefer to build structures where an element is merged directly into the head position instead of moving it to the specifier from below. The two constructions in () may also have coexisted in the language for several generations. Nevertheless, to avoid ambiguity, new lexical items developed to fulfil the demonstrative function in Romance: an adverbial reinforcer has been added to the phonologically weakened demonstrative head resulting in a new series of deictic elements. Interestingly enough, in Hungarian, the homophonous structures split in a different way, resulting in a somewhat surprising word order variation on the nominal left periphery. This problem will be addressed in section . on Middle Hungarian. .. Possessives As was presented in section .., Hungarian has two kinds of possessor expressions, dative-marked and unmarked. Although the same holds for Old Hungarian, it is worth examining whether the characteristics of these possessors correspond to those established for their Modern Hungarian descendants. As far as possessor extraction is concerned, the existence of extracted constituents could be tested and justified by an automatic query in the digitized Old Hungarian Corpus. More precisely, the search has been completed for one of the early codices, namely in Jókai Codex, which has also been morphosyntactically annotated, thus it can be more reliable to probe. After having also examined manually all the examples in the selected manuscript, out of possessive constructions in which the possessor is dative-marked, external possessors were found.
T . Positions of dative-marked possessors in Jókai Codex Constructions with dative-marked possessors in the text
Possessor and possessed noun are adjacent
Possessor and possessed noun are separated by attributes, or quantifiers
Possessor extracted out from the NP
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
The presence/absence of an article between the two members of the possessive construction is not a distinctive feature in Old Hungarian. As was stated in the previous section, the use of the definite article is more restricted in this period, being redundant in the contexts where reference is identified by another entity. Thus, as expected, the definite article never appears before the possessed noun in either of the constructions. This phenomenon has also been tested in the five normalized codices listed in Table .. I attested zero occurrence of the article between a caseless possessor and its possessum in constructions such as (b), and fewer than ten occurrences between a dative-marked possessor and its possessum in constructions such as (a). ()
a. a szerzetes-nek ∅ könyv-e the monk-DAT book-POSS ‘the book of the monk’ b. a szerzetes ∅ könyv-e the monk book-POSS ‘the book of the monk’
So the opposition shown in () for Modern Hungarian does not hold in Old Hungarian. Undoubtedly, the morphologically marked possessor must have emerged in ProtoHungarian (or rather reanalysed from an ‘affected’ constituent, cf. É. Kiss for modern Hungarian) to encode extracted or external possessors. The external position of the dative-marked possessor can still be observed in sporadic examples, where an overt and coreferent NP/DP-internal possessive pronoun co-occurs with the lexical possessor. By the documented era, however, the variation presented in () is widespread, with no interpretive difference. Accordingly, there is no reason to assume, on a synchronic level, that the Old Hungarian dative-marked possessor could not occupy Spec,DP in these cases. (Of course, when the dative-marked possessor is separated from the possessum, there are still two possibilities: it has been extracted out of the NP/DP and is co-indexed with a pro, or else, it is really external, which means that it has been generated outside the DP.) What the data suggest at this point of the analysis is that while unmarked possessors always occupy the specifier of the DP, dativemarked possessors can freely appear both in an external and in an internal position and, in the latter case, they are located in Spec,DP as well. The assumption that both types of Old Hungarian possessor expressions prefer the Spec,DP position is supported by a remarkable observation: possessives and demonstratives seem not to be attested together in this period, suggesting that they mutually exclude each other. Does it mean that these expressions occupy the same structural position? It is really interesting to observe how the Old Hungarian codices and an early printed book from the beginning of the Middle Hungarian period translated the Latin phrase of Matthew :, which contains the rare combination of a possessor and a demonstrative modifier. They all apply various avoidance strategies by ignoring the deictic element, or by transforming the possessive relation into an attributive construction or a relative clause:
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
Gospel of Matthew ::8 ()
a. Latin Et praedicabitur hoc evangelium regni in universo orbe and will.be.preached this gospel kingdom.GEN in whole.ABL world.ABL b. King James Version of the Bible And this gospel of the kingdom shall be preached in all the world c. Munich Codex [] ! ignores the demonstrative És prédikáltatik az ország-nak evangéliom-a mend ez világ-ban and be.preached the kingdom-DAT gospel-POSS all this world-INE d. Jordánszky Codex [–] ! attribute instead of a possessor És prédikáltatik ez mennyei szent evangéliom mind ez teljes and be.preached this heavenly holy gospel all this whole világ-on world-SUP e. New Testament of Pesti Gábor [] ! relative clause És ez evangéliom, ki mennyből vagyon, hirdettetik and this gospel which heaven.ABL is be.preached
. Middle Hungarian phenomena .. Demonstratives By the expansion of the use of the article in semantically definite contexts, such as generic readings and nouns with an inherently unique referent, the increased frequency of the spelling out of the D head led to successive reanalyses at the left periphery of the noun phrase. During the sixteenth century, in the Middle Hungarian period, a new pattern arises for the noun phrases containing a demonstrative. In this structure a sort of determiner doubling can be observed, which seems to be identical—at least at first sight—to the construction we have in Modern Hungarian, where the functional projection for definiteness is doubly filled. This means that the demonstrative is spelled out in the specifier position, while the determiner functioning as an article is in the head of the DP (cf. examples in () and the tree in ()). Observing the historical data, however, a distinct synchronic state can be identified in which word order patterns at the nominal left periphery differ from those established in section .. In the sixteenth and seventeenth centuries, the newly added demonstrative can be separated from the rest of the construction by various elements, e.g. particles, conjunctions, sometimes by the verb itself, which means that it links more loosely with the noun phrase in Middle Hungarian than it does today. (a) and (a) are examples from the first half of the seventeenth century (taken from the
8 The Latin source is included in the text edition of the Old Hungarian translations. The King James Version of the Bible is available at .
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
descriptive grammar of Simonyi : –), while in (b) and (b) alternative word orders are presented in order to show how these phrases would look in Modern Hungarian. In the Middle Hungarian data, the sequence of the demonstrative and the definite article is broken: in (a) a scalar particle intervenes, while in (a) an interrogative particle comes in between. (Pázmány Five letters )
()
a. azok is az esküvések those also the swearings ‘also those swearings’ b. azok az esküvések is those the swearings also ‘also those swearings’
()
a. Mit választasz inkább? ezt-e a pénzt? what choose.you more? this.ACC-INT.PRT the money.ACC ‘What do you prefer to choose? This money?’ (Káldi Preachings ) b. ezt a pénzt-e this.ACC the money.ACC-INT.PRT
To account for the Middle Hungarian data and for their different appearance with respect to their modern counterparts, it seems plausible to assume that the determiner doubling construction is the outcome of an adjunction operation, where the demonstrative pronoun as a whole adjoins to the DP headed by an article: ()
[DP [DP ez-t] [DP a pénz-t]] this-ACC the money-ACC
It is to be noted that the spread of the new pattern in the Middle Hungarian period does not show the same degree with the proximal and the distal demonstrative modifier. The proximal demonstrative ez differs from the article in its form, causing no ambiguity of the kind illustrated in (). As the figures in Table . indicate, the doubling construction with ez is much less frequently attested in the entire period and its spreading is definitely slower than that with the distal demonstrative.
T . The spread of the constructions ‘ez/az az N’ th c.
th c.
th c.
For the entire period
‘az az N’
.%
.%
.%
.%
‘az N’
.%
.%
.%
.%
th c.
th c.
th c.
For the entire period
.%
.%
.%
.%
.%
.%
.%
.%
‘ez az N’ ‘ez N’
(Figures after Dömötör : –)
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
.. Possessives In line with the tendency of the more and more frequent spelling out of the D head by means of an article, the definite article started to precede a possessed noun with a dative-marked possessor as well. This means that patterns exemplified in (a) and () are getting to be attested from the Middle Hungarian period. As for the structure behind this emerging pattern, dative-marked possessors may be assumed to adjoin to the DP as a whole (just the same way as demonstratives do, cf. ()), if one accepts the claim that the simultaneous filling of the D head and its specifier is not preferred. As demonstrated in section ., dative-marked possessors can regularly appear detached from the noun phrase as early as in the first continuous manuscripts, thus an extracted or external possessor can easily be reanalysed as an adjoined constituent in Middle Hungarian (a). Nevertheless, the Old Hungarian possessive construction with no article before the possessed noun (b) remains available and survives besides the new pattern. ()
a. [DP [DP a szerzetes-nek] [DP a könyv-e]] the monk-DAT the book-POSS b. [DP
a szerzetes-nek the monk-DAT
[D’ ∅ [könyv-e]] book-POSS
Remarkably, unlike in Old Hungarian, the combination of demonstratives and possessives becomes well attested in this period, but at the same time shows a peculiar distribution in which the demonstrative precedes the possessor expression, as exemplified in () to ()—the data in () and () are taken again from Simonyi (): ()
azok a szerzetes-nek könyv-e-i that.PL the monk-DAT book-POSS-PL ‘those books of the monk’
()
az-ok az Angliá-nak nemes-e-i that-PL the England-DAT nobleman-POSS-PL ‘those noblemen of England’
(Faludi )
()
ar-rul is az bibliá-nak rész-é-rűl that-DEL also the Bible-DAT part-POSS-DEL ‘also about that part of the Bible’
(Pázmány Kal. )
These word orders, with the demonstrative on the left and the possessor inside, would be completely ill-formed in Modern Hungarian (cf. examples in ()). Because of the fact that, at this point of the research, Middle Hungarian texts have not been digitized and data could only be collected manually, it is not clear when the Modern Hungarian word order (Poss Dem) appeared. Neither can we provide any numerical data about the proportion of the peculiar Dem Poss word order. (The combination is extremely rare in every period for obvious semantic reasons.) The question to be raised in the present analysis is how this word order variation can be accounted for. The possibility of a more extended noun phrase evidently emerged in this period, but while the D head is more and more systematically occupied by the definite article,
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
the position of the dative-marked possessor and the demonstrative, two expressions of quite different nature, have not become fixed yet. The situation may be reconstructed in the following way. When the determiner doubling construction emerged in Middle Hungarian, the demonstrative (showing agreement in case and number with the noun) adjoined either to the noun phrase headed by the definite article (as in ()), or to the old, articleless and a more condensed possessive structure (b), but it did not adjoin to the possessive structure in which the possessor expression was also ‘external’ (a) —at least, the author does not know of such occurrences. The combination of the new loosely adjoined demonstrative with (b) is illustrated below (an accusative case ending has been added to make the agreement noticeable): ()
a. [dp [dp az-t] [dp a szerzetes-nek [d’ ∅ [ könyv-é-t]]] that-acc the monk-dat book-poss-acc ‘That book of the monk’ b.
DP DP
DP Spec
D’ D
az-t a szerzetes-nek that-acc the monk-dat
PossP könyv-é-t book-poss-acc
In section .., it could be observed how Old Hungarian sources had struggled with translating a phrase from the gospel of Matthew (Matthew :), in which a demonstrative modifier should have been combined with a possessor expression. In the so-called Bible of Vizsoly, compiled in , the problem is, however, practically solved by the application of this new combination: Matthew : in the Bible of Vizsoly (): ()
Es predikáltatic ez Isten ország-á-nak Evangeliom-a mind az and be.preached this God kingdom-POSS-DAT gospel-POSS all the egész föld-ön whole land-SUP
The question remains how we get to the present-day distribution where, according to the strict word order rule, dative-marked possessors always have to precede demonstrative modifiers. The simultaneous appearance of a demonstrative and a possessor must always have remained infrequent, but simple demonstrative modification is definitely and universally a common construction. At a certain point in the history of the Hungarian language, which cannot be defined in lack of empirical data but must be placed between Middle Hungarian and present-day Hungarian, the position of demonstratives was established in the specifier of the DP. When looking for an
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
explanation, one might rely on Van Gelderen’s (b: ) third universal economy principle, the so-called ‘Specifier Incorporation’ principle, that claims for elements coming from outside to tend to be a specifier rather than an adjunct.9 In accordance with this principle, the originally adjoined demonstrative phrase must have been integrated into the DP domain, as its behaviour in Modern Hungarian suggests (see section ..). In this way, the demonstratives’ rather atypical nature in the presentday system becomes more reasonable: they appear in a doubly filled DP projection, and their morphology is also somewhat exceptional, since they are marked for case and plurality unlike other modifiers within the Hungarian noun phrase. All these can be due to their appositive origin and to the assumption that, diachronically speaking, they arrived from outside the DP. However, a final question may be raised: why is it the demonstrative that has become fixed in Spec,DP rather than the possessor expression, as the word orders presented in ()–() would suggest? The answer might be found by considering the different nature of demonstratives and possessors. Whereas demonstratives were (and are always) closely associated with the noun they modify, the option for a dative-marked constituent to be interpreted as an external constituent (with an ‘affected’ theta-role, cf. É. Kiss ), or to be extracted or even completely omitted always remained available. At the same time, of course, in unmarked cases, the possessor appeared adjacent, syntactically adjoined, to the determined noun phrase. After the demonstrative had been integrated into Spec,DP, the target position of the adjoined dative-marked possessors may have been reanalysed as a new layer of the extended noun phrase, namely a Topic projection—in line with what É. Kiss () proposes.
. Summary In this chapter, by studying the determination and the distribution of possessive constructions from Old Hungarian to the present day, a peculiar word order change phenomenon has been accounted for. The grammaticalization of the definite article in well-defined contexts had structural consequences at the nominal left periphery, the most salient of which is the emergence of a new strategy for demonstrative modification. This strategy is called determiner doubling throughout the chapter. It has also been observed that the newly added demonstratives first adjoined to the noun phrase in a somewhat looser fashion: their combination with the dative-marked possessors resulted in a word order specific only to the Middle Hungarian period (Dem Poss). Demonstratives, however, at a later stage, got incorporated into the specifier of the DP as the Modern Hungarian fixed word order (Poss Dem) clearly testifies.
9 Van Gelderen, providing a general overview of the DP cycle in several languages (, and Chapter in ), points out that after the reanalysis of the demonstrative as an article, the renewal of demonstratives may happen in one of the following ways: either from above, by the incorporation of an appositive demonstrative pronoun, or from below, through the addition of a locative adverb. It is to be noted, that reinforced forms of demonstratives (emez, amaz, ezen, azon) also emerged in the history of Hungarian, but today their use is highly limited and stylistically marked. For further discussion, see Egedi (a).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Barbara Egedi
The word order variation and the change discussed are summarized in the following schema of the various patterns and their combinations: ()
Schema of variation and change
[dp az [ ∅ könyv]]
NEW!
[dp a szerzetes-nek [ ∅ könyv-e]]
Old Hungarian
NEW!
Middle Hungarian
[dp [dp az] [dp a könyv]] [dp a szerzetes-nek [ ∅ könyv-e]] [dp [dp a szerzetes-nek] [dp a könyv-e]]
Combination: [dp [dp az] [dp a szerzetes-nek [∅ könyv-e]]]
NEW!
[dp az a [könyv]]
Present Day Hungarian [dp [dp a szerzetes-nek] [dp a könyv-e]]
Combination: [dp [dp a szerzetes-nek] [dp az a könyv-e]]
[TopP a szerzetes-nek [dp az a könyv-e]] [TopP a szerzetes-nek [dp a könyv-e]]
The complex figure in () presents how the Old Hungarian patterns without definite article were replaced by new constructions involving adjunction sites for demonstratives and dative-marked possessors. However, as the old possessive construction had still been preserved, its combination with the new determiner doubling strategy produced a word order that did not survive the Middle Hungarian period (see the combination in the centre of the figure). The lower part of the schema shows that the incorporated demonstrative cannot precede the possessive expression any more, and from this point on, dative-marked possessors occupy the outermost position of the extended noun phrase.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Change at the left periphery of the Hungarian noun phrase
Acknowledgements The chapter forms part of a complex research project devoted to the study of the diachronic syntax of Hungarian as well as to the building of an online searchable historical language corpus of Old Hungarian. The project has been supported by the Hungarian Scientific Research Fund (OTKA No. and No. ), and the first theoretical results have been summarized in an edited volume, including the contribution of the present author (cf. Egedi a). The focus of the latter book chapter and the present chapter considerably differ, but as their topics are closely related, some of the arguments and the data will necessarily overlap.
Sources Bible of Vizsoly: Szent Biblia (=The Holy Bible). Translated by Gáspár Károli. Vizsoly, . (Facsimile published by Helikon, Budapest, .) Faludi: Toldy, Ferenc (ed.): Faludi Ferencz minden munkái (=Collected Works of Ferenc Faludi). Nemzeti Könyvtár, . Jordánszky Codex: Toldy, Ferenc and György Volf (eds): A Jordánszky-kódex bibliafordítása. (Critical edition of the original Old Hungarian translation from –) Buda, . Káldi Preaching: Káldi György: Az innepkre-valo predikatzioknak első része. (=Preaching for feast days, Part One) Pozsony, . Munich Codex: Nyíri, Antal (ed.), Müncheni kódex. (Critical edition of the original codex form , with the corresponding Latin text) Budapest, Akadémiai Kiadó, . New Testament of Pesti Gábor: Pesti Mizsér Gábor: Új Testamentum magyar nyelven (=The New Testament in Hungarian). Facsimile edition of the original from . Budapest, MTA []. Pázmány Kal.: Pázmány Péter: Kalauz. (=A Guide) Pozsony, .
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Part II Triggers for movement Changes in nature or stability
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
6 Particle-verb order in Old Hungarian and complex predicates V E RO N I K A H E G E D Ű S
. Introduction The aim of this chapter is to examine where we find verbal particles (also called preverbs or prefixes) in the Old Hungarian clause, and to give a syntactic analysis that accounts for their properties. An analysis of particles should also be extendable to the distribution of the other elements that appear in the same surface position as particles in Modern Hungarian. The syntactic properties of particles in Old Hungarian are very similar to how they behave now. The slight variation they show in their orderings with respect to the verb can be attributed to independent syntactic factors, and this variation is marginally still present today. Although the basic word order of Hungarian had changed from SOV to SVO by the early written period of Hungarian, this change did not affect the particle (PRT)-Verb ordering. I will argue that this is because the original pre-verbal argument position of particles was reanalysed as a predicative position where particles obligatorily move in overt syntax in order to form a complex predicate with the verb. This (predicate) movement has been extended to all other predicative elements in the clause as evidenced by Modern Hungarian word order. However, predicative constituents other than particles show significant word order variation in Old Hungarian.
. On Modern Hungarian word order The word order of Modern Hungarian has been claimed to be mostly determined by information structural considerations, which often change the underlying constituent order. The constituent order has been argued to be SVO (Horvath ; Surányi ; É. Kiss ). The skeletal structure of the Hungarian sentence is sketched in (), following É. Kiss (, ). The left periphery of the clause includes positions for topics, distributive quantifiers and structural focus, as well as a position where predicative complements surface to the left of the verb. Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Veronika Hegedűs . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
()
Veronika Hegedűs [CP [TopP [DistP [FocP [PredP [VP . . . ]]]]]]
The predicative elements which appear left-adjacent to the verb in neutral sentences have been collectively called Verbal Modifiers (VMs; see Komlósy ; É. Kiss ). A sentence is neutral if there is no focus, wh-item or negation in it (see Kálmán a,b and subsequent literature on word order); and in non-neutral sentences, the verb precedes the VMs. The name ‘Verbal Modifier’ is indicative of the fact that these are not regular arguments but have a different semantic relation to the verbal predicate of the clause. The elements that we find here include non-verbal predicates of copular clauses, secondary predicates, verbal particles, goal arguments of motion verbs, and also bare nominal internal arguments, which have been argued to be predicative as well (Komlósy ; Farkas and de Swart ). Some examples are given in (). () a. János {okos / tanár} volt. John smart / teacher was ‘John was {clever / a teacher}.’ b. A kovács laposra kalapálta a vasat. the smith flat.SUB hammered the iron.ACC ‘The smith hammered the iron flat.’ c. A labda {be- /a kapu-ba} gurult. the ball into / the goal-ILL rolled ‘The ball rolled {in / into} the goal.’ d. Mari könyvet olvas. Mary book.ACC read ‘Mary is reading a book.’ It has been proposed in various places that the verb and the pre-verbal element form a complex predicate in Modern Hungarian. There have been different analyses with respect to the derivation of complex predicates. On the one hand, it has been proposed that the complex predicate is a lexical unit inserted into syntax as one (Ackerman and Webelhuth ) or that they are base-generated complex elements (Horvath ). Other analyses, however, assume movement of the predicative constituent into its pre-verbal position. The most recent analyses take this surface position to be a predicative position and the movement of the VM to be predicate movement (into PredP by É. Kiss , Surányi ; or into VP by Broekhuis and Hegedűs ). Old Hungarian exhibits most of the word order properties mentioned here, however, the pre-verbal VM position does not seem to be fully developed yet: particles are pre-verbal in most of the cases (with some minor variation), but the other VMs show much more variation in their surface order. This word order variation is interesting in light of the fact that Hungarian has been argued to have undergone a word order change from SOV to SVO before the written period, with some remnant OV properties still present in Old Hungarian.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates
. Word order change and consequences .. SOV vs SVO in Old Hungarian É. Kiss (, ) argues that Hungarian word order changed from (S)OV to (S)VO by the Old Hungarian period. Old Hungarian, however, still preserved some fossils of the OV order. One such property is that we can find the OV order with a caseless object in participial clauses (e.g. ()), although objects are normally marked for accusative in Old Hungarian and in later stages as well.1 () mel myse haluan vymadgozzonk harmad ydeyglen which sermon hear.PART pray.SUBJ.PL third time.until ‘Upon hearing which sermon, let us pray three times’ (Jókai Codex , É. Kiss (a)) Another word order found in Old Hungarian that É. Kiss argues to be a remnant of the old OV order is related to negation: in negative sentences we frequently find the order PRT-NEG-V, as opposed to the now predominant NEG-V-PRT. The former order is still marginally present today but it has a marked meaning which it did not seem to have in the old texts. () Es az Ver touaba ky nem futott and the blood moreover out not ran ‘and the blood didn’t run out anymore’ (Jókai Codex , É. Kiss ()) One more property that can be analysed as a remnant of the old OV order is the presence of sentence-final question particles. The Q-particle -e is at the end of the clause in examples such as (). () Nēde tu̇ incab nagobac vattok azocnal ė? not.but you rather bigger.PL are those.ADE Q ‘Are ye not much better than they?’ (Munich Codex vb, É. Kiss (b)) The position of the Q particle was in transition from being clause-final to appearing on the verb or the focus of the clause in Old Hungarian. The process was in progress from the beginning of the written period, indicated by the word order variation in the texts. These changes indicate the final steps of a word order shift whereby the headfinal structures had changed into head-initial ones. At the same time the (discourse) functional, information structural left periphery of the clause was mostly developed by this period: we find topic, focus and quantifier phrases preceding the verb, while regular arguments can easily be post-verbal (when they do not have a discourse function and are not operators). This raises the issue of change in the distribution of elements such as particles, which are not regular referential arguments themselves.
1
Most of the other Uralic languages still have the same OV order with unmarked objects.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
If arguments remain in the post-verbal field in the newly VO Hungarian, the different distribution of particles has to be accounted for. From a crosslinguistic point of view, it is also interesting to see what happened in other languages that also changed from SOV to SVO and have particles. .. The Germanic parallels Diachronically, English provides an interesting parallel to the Hungarian word order changes. The change from SOV to SVO in the history of English is well documented and has been shown to involve a change from PRT-V order to V-PRT order (cf. Van Kemenade and Los ; Elenbaas ). As far as the position is concerned, Kroch and Taylor () use the surface position of the particle as a diagnostic to determine whether the language is already a VO language or not at a given point. Since particles are light elements, they would not be rightward-moved (would not undergo heavy NP-shift or rightward topicalization), so if the particle appears to the right of the verb we can assume that it is base-generated there. Thus, at the point when particles appear to the right of the verb, we can analyse the language as having VO order. On the other hand, when we look at synchronic data, Germanic languages with SOV word order usually have the particle pre-verbally. However, this does not necessarily mean that it is base-generated there. Under the assumption that all languages start out with an SVO base order (cf. Kayne ), the surface position of the particle is derived from a post-verbal base-position. To account for the properties of Dutch secondary predicates (including particles), Zwart () and Koster () claim that secondary predicates move into a functional projection on top of the VP during the derivation of the clause. They surface to the left of the verb, in the specifier of a Predicative Phrase (PredP), while the verb is in Pred0, this is what accounts for the fact that particles and other secondary predicates are pre-verbal.
. Particle-verb constructions in Old Hungarian .. The class of particles The class of verbal particles was already quite large in Old Hungarian and it was gaining new members then as well (D. Mátai , , , etc.). The oldest verbal particles are meg ‘orig. back’, el ‘away’, ki ‘outDIR’, fel ‘upDIR’, le ‘downDIR’, bel/be ‘into’; these all have uses as particles from the early sources. We can also find newly grammaticalizing elements, such as egybe ‘together’ (lit. ‘one.into)’, által ‘over, through’, alá ‘underDIR’, elé ‘beforeDIR’, össze ‘together’, vissza ‘back’, elő ‘lit. to the front’. Out of the oldest ones, meg is sometimes only a functional element and has no real, directional (i.e. lexical) meaning, as illustrated in (), but sometimes it is still used with the original directional meaning ‘back’, like in () with the verb ‘go’. () zent ferenc mene zent Jacabot meg-latny saint Francis went saint Jacob.ACC PRT-see.INF ‘Saint Francis went to see Saint Jacob’
(Jókai Codex )
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates
() Es fèlèlet vėuėn almocban hog ne mennenèc meg herodèshėz and reply taken dream.POSS.PL.INE that not go.COND.PL back Herod.ALL ‘And being warned [of God] in a dream that they should not return to Herod’ (Munich Codex ra) The original, directional use of meg slowly disappears, and other particles take over its function in this meaning. This particle has lost its spatial meaning and is a functional element in Modern Hungarian, a telicity marker. The other particles still have directional meanings in at least some of their uses. The spread of particles is apparent in this period; they appear in more and more contexts. On the one hand, the old ones are used more extensively. For example, meg ‘back’ or el ‘away’ newly appear in a lot of contexts. Consider the examples from parallel texts which were translated less than a hundred years apart (Munich Codex from / vs Jordánszky Codex from –). The newer texts in () and () have particles where the older ones in () and () lack them, that is, particles seem to be spreading into new contexts. ()
es mutata nèki mend èz vilagnac orzagit and showed to.SG all this world.DAT country.POSS.PL.ACC ‘and showeth him all the kingdoms of the world, and the glory of them’ (Munich Codex ra)
()
es meg mwtata hewneky ez vylaghnak mynden orzaghyt and PRT showed he.to.SG this world.DAT all country.POSS.PL.ACC ‘and showeth him all the kingdoms of the world . . . ’ (Jordánszky Codex )
() es èrèztuē oket betlechembe and send.PART them Betlehem.ILL ‘And having sent them to Bethlehem’ () es el-ereztwen hwket betlehembe and away-send.PART them Betlehem.ILL ‘And having sent them to Bethlehem’
(Munich Codex vb)
(Jordánszky Codex )
On the other hand, particles in general became more frequent as a consequence of the grammaticalization of new particles as well. For example, által ‘through, over’ is considered to be new in Old Hungarian, although it already appears—with a spatial meaning—in the oldest translated codex from the beginning of Late Old Hungarian. () Myre kesel altal-mened az vyzen what.SUB delay.SG over-go.INF.SG the water.SUP ‘Why are you delaying going across the water?’
(Jókai Codex )
The most frequent meg is sometimes substituted by other particles with more obvious spatial meanings. Parallel texts from two translations sometimes use different particles, and in several cases meg is replaced by el ‘away’ when referring to leaving a place.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
() Tahat meg-haga ọtèt az ọrdọg then PRT-left him the devil ‘Then the devil leaveth him’ () Ottan el-hagӳa hewtet az erdeӳg there away-left him the devil ‘Then the devil leaveth him’ () es nazaret nèuọ varos m̅g-haguan and Nazareth named city PRT-left ‘And leaving Nazareth’ () es el hagywan nazarethnek varosat and away left Nazareth.DAT city.POSS.ACC ‘And leaving Nazareth’
(Munich Codex ra)
(Jordánszky Codex )
(Munich Codex rb)
(Jordánszky Codex )
This suggests that meg was losing its spatial meaning already at this stage, and the more obviously spatial elements replaced it whenever the spatial meaning was relevant and the particle was not simply indicating the telic nature of the event. .. The distribution of particles The word order properties of particles in present-day Hungarian give us a clue that it is worth taking a look at their distribution in neutral and non-neutral sentences separately. What we find in the Old Hungarian corpus is that particles display a distribution that is very similar to their present one: they generally appear right in front of the verb in neutral sentences and post-verbally in non-neutral sentences. I will show examples of the two sentence types in this section including not only the general pattern but the typical exceptions as well. ... Neutral sentences
In neutral sentences we find particles pre-verbally.
() De frater ylyes meg haragwan: sebesseguel bel-teue aytayat but brother Elias PRT angry.got speed.INSTR into-put door.POSS.SG.ACC S el-tére and away-went.SG ‘But having got angry, Brother Elias quickly closed the door and left’ (Jókai Codex ) () Es ọ a haiocskaba fėl-mēuen and he the ship.DIM.ILL up-go.PART ‘And when he was entered into a ship’
(Munich Codex va)
However, sometimes we find the particle post-verbally when the sentence seems neutral otherwise. This order is unexpected based on our present-day intuitions. () De menenek el pyspekseg here. but went.PL away bishopric place.SUB ‘but they went to the bishopric’
(Jókai Codex )
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates () Tehat zent ferenc mene el az erdewbe then saint Francis went away the forest.ILL ‘Then Saint Francis went to the forest’
(Jókai Codex )
It is not always easy to decide if a sentence is indeed neutral since we do not have information about intonational patterns, so we can solely rely on linguistic context and interpretation. I will briefly return to these examples in section .. ... Non-neutral sentences Particles are usually post-verbal in non-neutral sentences. As was mentioned above, non-neutral sentences are of several types, so let us take a look at the different types one by one. A sentence is non-neutral when it includes structural focus, that is when the identificational focus of the clause is in the pre-verbal position. The verb has to immediately follow the focused constituent in such sentences, therefore the particle surfaces to the right of the verb (and other VMs do as well), in the postverbal field.2 In (), it is the manner adverb that is in focus and the particle is post-verbal. () ky ezkeppen fordolot vala meg who this.way turned PAST PRT ‘who turned around this way’
(Jókai Codex )
Negation is also pre-verbal in Hungarian, thus in negative sentences, the particle is expected to be post-verbal, and we find this word order in Old Hungarian in many cases; for example, in (). () nem fyzettel telyesseguel meg not paid.SG completely back ‘you did not pay completely back’
(Jókai Codex )
It is with negation where we find most of the exceptions to the generalization that particles are post-verbal in non-neutral sentences. The particle quite often appears before the negative particle (nem/ne) in front of the verb. () De aldott cristus sonha hyw embernek zyuett but blessed Christ never faithful man.DAT heart.POSS.ACC meg nem kemeneyty PRT not harden.SG ‘But blessed Christ never hardens the heart of a faithful man’ (Jókai Codex ) () es sem urroc ki nē assac, sē èl-nē uroziac and neither thief.PL out not dig.PL neither away-not steal.PL ‘and where thieves do not break through nor steal’ (Munich Codex va)
2 The word order of the post-verbal field seems largely to be free. For analyses of this free word order, see Surányi (); É. Kiss ().
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
While the standard analysis of negative sentences (in Modern Hungarian) is that the verb moves to the Neg head and that is why the particle is post-verbal (É. Kiss ), these sentences cannot be derived that way. É. Kiss (, ) assumes that this order is actually a remnant of the old OV word order and the derivation of negative sentences in Old Hungarian should proceed differently from their present derivation. Her proposal is that Neg is adjoined to the V head in these sentences; therefore the particle will surface to the left of both of them. Another instance of non-neutral sentences are imperative sentences. They typically have the order V-PRT in modern Hungarian and the same appears to be the case in Old Hungarian. () Ha akarz tekelletes leny men-el if want.SG perfect be.INF go.SUBJ.SG-away s-arwld-el mendenedett and-sell-SUBJ.SG-away all-POSS.SG.ACC ‘If you want to be perfect, go away and sell all your possessions’ (Jókai Codex ) However, similarly to the sentences that involve negation, the particle sometimes stays before the verb in imperative sentences. () bè-men˙ te hailakodba into-go.SUBJ you house.POSS.SG.ILL ‘enter into thy closet’
(Munich Codex rb)
If we analyse imperative sentences as involving V-movement to a higher functional projection in the CP-domain—as has been proposed for Modern Hungarian (see Varga )—we can derive the post-verbal position of the particle. We must assume, however, that in these cases with PRT-V order, the verb stays low. Indeed this has been suggested for similar Modern Hungarian examples by Dömötör and Varga (), who argue on the basis of semantic and pragmatic evidence. I will return to this word order briefly in section ...
. The syntactic properties of Old Hungarian PRT-V constructions .. The syntax of particles Particles are functional adpositions; they originate within the extended Adpositional Phrase (PP). I call their projection pP, following Van Riemsdijk (), who first analysed particles as being base-generated in a functional layer on top of the PP. Much research has been done since then on the fine-grained structural analysis of PPs, as a result of which the structure expanded into various projections hosting place and path denoting elements and modifiers within the PP. Here I am concerned with the syntactic properties of particles, and I assume that when they appear with an additional PP in the clause, they form a constituent in base structure, as in (), the structure of which is depicted in ().
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates
() el az erdő-be away the forest-ILL ‘away into the forest’ ()
pP p’
Spec P
PP
el ‘away’
P
NP
-be ill
az erdő the forest
The functional projection pP is essentially analogous to Ramchand and Svenonius’ () PrtP, an adpositional Relator Phrase (Den Dikken a) or a Small Clause (Den Dikken ). The subject of the SC is introduced in Spec,pP. ()
pP
Spec zent Ferenc ‘Saint Francis’
p’ p el ‘away’
PP P -be ‘ill’
NP az erdő ‘the forest’
From a diachronic point of view, the grammaticalization process of adpositional elements involves the so-called preposition cycle (see Waters for English), and particles (i.e. functional adpositions) develop as part of this cycle (cf. Hegedűs ). Particles grammaticalize from (semi-)lexical adpositional elements and become functional heads. In Hungarian, these functional elements form a complex predicate with the verb in overt syntax. .. Complex predicates In SOV languages, the VP is verb-final, that is, both referential and non-referential arguments appear to the left of the verb. If Proto-Hungarian was indeed an SOV language, we can assume that SC complements of the verb were pre-verbal as well. It was not always the case, however, that the arguments were pre-verbal. É. Kiss (, ) has shown that referential XPs often appeared post-verbally due to rightward
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
topicalization or heavy-NP shift; there are a large number of verb initial clauses at the beginning of the written period. É. Kiss argues that the old base order got reanalysed as SVO on the basis of this cue. There are, however, sentences in Old Hungarian where we find both a caseless object and a particle in front of the verb. This only occurs in participial clauses and not in finite clauses. Examples are given in () and (). () es feye le hayttuan tystesen fogada zent and head.POSS.SG down turn.PART respectfully vowed.SG saint attyanac engedelmesseget father.POSS.SG.DAT obedience.ACC ‘And having bowed his head, he respectfully pledged obedience to his Saint Father’ (Jókai Codex ) () es azoc legottan haloioc meg-haguā and those immediately net.POSS.PL PRT-leave.PART ‘And they straight away left their nets’ (Munich Codex rb) I take these examples to show that the whole complement, which is a Small Clause, is in front of the verb here. This may be a further remnant of the old OV order similar to the ones in section .., with not only the internal argument but also the secondary predicate appearing pre-verbally. () [VP [SC INT.ARG PREDICATE ] V ] Particles are different from regular arguments since they are not referential and are very light, so they are less likely to undergo rightward movement or shift (cf. Kroch and Taylor ). What we find in the majority of cases in Old Hungarian is that particles appear to the left of the verb but regular (referential) arguments do not. I assume that this order could have been the cue to the language learner that the position in front of the verb is not an argument position, that movement into this position is not obligatory in order to check φ-features locally. The position is not one of the information structure related functional projections at the left periphery either, although they had developed by this time (É. Kiss , ). The fact that this position was the host of particles and other non-referential elements at this stage could have been a cue that this is a predicative position. There are various possibilities for the analysis of the surface position of the particle (and other predicates), and it is a central topic of works on the syntax of Modern Hungarian. Surányi () has shown that a two-step analysis may be necessary to derive the surface positions, but for our purposes, the first step, which derives the linear order and adjacency of the particle and the verb, is the crucial one.3 I assume that the original preverbal position of the internal argument was reanalysed as a predicative position.
3
Surányi () argues that in the second step of the derivation the verb moves into T and the already pre-verbal VM moves into Spec,TP. This step does not change the linear order of the verb and the VM in neutral sentences.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates
Referential arguments do not have to be moved, since there is no need to check φ-features locally, but predicates have to raise in order to form a complex predicate. That is to say, an A-movement driven by the need to check φ-features locally was replaced by a movement that is motivated by the need to form complex predicates in overt syntax, which seems to be fully generalized in neutral sentences in Modern Hungarian. We can account for the word order and the necessary adjacency of the particle (and other predicates) and the verb if we adopt the analysis which assumes a universally available functional projection, PredP (as proposed by Zwart and Koster for Dutch), the specifier of which is the position where predicates move, while the verb raises into its head; see the tree structure in (). ()
PredP Pred’
Spec meg ‘prt’
Pred haragwan ‘angry.got’
VP V
SC
haragwan
…meg…
The papers in É. Kiss (), as well as Surányi (), adopt this proposal for Modern Hungarian, and I propose that the projection and movement into its specifier is present in Old Hungarian as well. This PredP may be thought of as the locus of complex predicate formation and the movement into its specifier as predicate movement.4 Old Hungarian particles seem to have undergone predicate movement almost obligatorily in neutral sentences but this movement was not fully generalized to all predicative elements. As will be shown in section ., other VMs show much more word order variation in the Old Hungarian data. Of course, if we assume that particle movement was obligatory in Old Hungarian, we have to account for the cases where it did not seem to have taken place. The exceptional cases that we saw in () and () may fall under one of two scenarios. One possibility is that in these sentences, the post-verbal element is not a particle yet but a simple directional adposition. Structurally, the two are different, directional adpositions occupy the Path head, while particles are higher functional heads (p0) in the extended PP. As we will see in the next section, directional complement PPs did not always end up pre-verbally in Old Hungarian, although they are mostly preverbal today.
4 Broekhuis and Hegedűs () also take this movement to be predicate movement, but they propose that the landing site is Spec,VP. Under the PredP account, however, we expect the lack of movement in the case of referential internal arguments (e.g. definite objects), that is why I adopt this analysis for the Old Hungarian data as well.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
The other possibility is that these sentences have different properties from regular neutral sentences, for example, they are aspectually different. It may be that the event is still atelic and moving the particle would telicize the complex predicate, therefore there is no particle movement.5 At the same time, we also find examples where the particle is pre-verbal, but we are dealing with a non-neutral sentence (where we would expect the verb to precede the particle), as we have seen in section .... In negative sentences the particle is not adjacent to the verb, since negation intervenes. This construction is still possible today (see ()). É. Kiss (, ) claims that this is an archaic construction that is marginally present in Old Hungarian (see () and ()–()), and it seems to live on up to now as a marked semantic/ pragmatic choice. () El nem árulom senkinek a titkot! away not sell.SG nobody.DAT the secret.ACC ‘I will not tell the secret to anyone (I promise).’ It is also still possible to use imperatives with non-inverted word order, similarly to the examples we find in Old Hungarian, as was illustrated in (). () Aztán meg-edd a spenótot! then PRT-eat.SUBJ.SG the spinach.ACC ‘Do eat the spinach.’ In an analysis of imperative sentences in Modern Hungarian, Varga () assumes that the verb moves to a higher functional projection, namely into FinP, thereby deriving the regular V - PRT word order. In sentences like (), the verb must stay lower. Since these sentences have a special meaning, Dömötör and Varga () claim that they are not real imperatives and analyse them as matrix subjunctive clauses, where the verb does not move into the C domain but remains lower in its regular surface position.6 They make this claim based on observations concerning the restricted use of this construction (in present-day Hungarian) both syntactically and pragmatically. This accounts for the distribution of particles both in neutral sentences and in non-neutral ones. Particle movement is assumed to be an instance of predicate movement as a result of a reanalysis of the old pre-verbal argument position into a pre-verbal predicate position. Predicate movement takes place in order to create complex predicates with the verb in overt syntax; however, it was not obligatory for all predicative elements in Old Hungarian.
5 It could be argued that particle movement is movement into an aspectual position (e.g. AspP) and it would account for the movement of most particles. However, not all VMs have an obvious relation to aspect, therefore, their being predicates seems to me to be a more decisive factor. It is for this reason that I have not opted for an analysis where aspect drives the movement of VMs (see also É. Kiss ). 6 Subjunctive involves V - PRT word order in embedded clauses as well.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates
. Extending predicate movement: other predicative elements Particles, bare noun arguments (which are always internal arguments, see É. Kiss ), predicative PPs, predicative nominals and adjectives, and secondary predicates behave alike in Modern Hungarian in that they are pre-verbal in neutral sentences. This is the group of elements and the position that is often referred to as VMs. We examined the behaviour of particles in Old Hungarian and found that they are pre-verbal with few exceptions. The other predicates, however, display a greater variation in their distribution in Old Hungarian. I will turn to these data now and will show that when we look at neutral sentences, the other predicates are sometimes pre-verbal (as expected on the basis of our present intuition and of the behavior of particles), but almost as frequently they are post-verbal. We find variation in the order of the same kind of element within one text. For example, nominal and locative predicates in copular clauses show variation in () and ().7 () a. mert halazoc valanac because fisher.PL were ‘for they were fishers’ b. ọ etke ke(dig) vala saska es vad mėz he food.POSS.SG CONJ was locust and wild honey ‘and his meat was locusts and wild honey’ (Munich Codex rb; va) () a. kic a· hazban vadnac who.PL the house.INE are ‘that are in the house’ b. ki vagon men˙ n˙ ècbèn who is heaven.PL.INE ‘which is in heaven’
(Munich Codex ra)
There is also variation within different translations of the same text, which can be seen, for example, when we compare Bible translations. The example in () contains a post-verbal locative predicate next to the copula, while the same element is preverbal in (), which is the same line from a later translation. In () and (), however, it is the later text that uses the post-verbal order. () es vala ot Herodes megholtiglan and was there Herod PRT.death.until ‘and was there until the death of Herod’ () es ot len herodesnek halalayglan and there is Herod.DAT death.until ‘and was there until the death of Herod’
(Munich Codex ra)
(Jordánszky Codex )
7 The line in (b) is often used in Biblical texts to refer to God, and since it is used frequently, it is almost like an idiomatic expression with a set word order.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
() es a· tọmlọćbè èrèsztessèl and the prison.ILL send.CAUSE.SG ‘and thou be cast into prison’
(Munich Codex rb)
() es vettessel az temleczben and cast.CAUSE.SG the prison.INE ‘and thou be cast into prison’
(Jordánszky Codex )
I have counted the first occurrences of VMs (i.e. predicative elements according to the classification from section .) in St Matthew’s gospel in the Munich Codex. Out of the elements, are particles and are other predicative elements. Their distribution is summarized in Table .. T . Particles and other predicative elements Order Pre-verbal expected
Particle
Other
Pre-verbal unexpected
Post-verbal expected
Post-verbal unexpected Together
I call ‘unexpectedly pre-verbal’ those elements that occur pre-verbally in nonneutral sentences. Of the seventeen unexpectedly pre-verbal particles, ten are with negation (while we also find eleven clauses with regular NEG-V-PRT negative order in the ‘post-verbal expected’ group) and seven are in imperatives (and there are also twenty-one regular V-PRT imperatives). The unexpectedly post-verbal particles occur in sentences that may be interpreted as progressive, and they all involve the particle el ‘away’, which has a clear directional meaning in many of its uses, so it may also be a directional predicative complement to motion verbs. One of the examples—which actually occurs twice with the same order—is given in (). () es zallot le and flew down ‘and [the rain] descended’
(Munich Codex vb)
When we turn to the other VMs in this small corpus, we find much more variation. Only four examples exhibit unexpected pre-verbal predicative elements, and all of those may involve focus or contrastive topic interpretation of the VM and, therefore, appear at the beginning of the clause for that reason. However, eighty-six of the predicative elements are unexpectedly post-verbal, that is, they are post-verbal in seemingly neutral sentences. There does not seem to be a pattern behind the variation, so we have to conclude that they did not have to move to the pre-verbal predicative position obligatorily at this point in the history of the language.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Particle-verb order in Old Hungarian and complex predicates
It needs further work to account for the difference between particles and other VMs. However, the already mentioned lightness and sometimes the lack of lexical meaning of particles may provide a clue. The fact that particles were never shifted to the right (both because they are light and because they are not good candidates for topicalization) resulted in their different behaviour, and they provided a different cue as compared to heavier or more conventional arguments (e.g. bare objects). When the VO order was generalized to all internal arguments—including predicative ones—particles may have been the first cue to reanalyse the pre-verbal position as a predicative one. Predicate movement and complex predicate formation may first have been generalized to particles and it only became completely general at the next stage.
. Conclusions We can conclude that despite the change in the Hungarian word order from SOV to SVO by the beginning of the written period of the language, the PRT - V order did not change because the previous pre-verbal argument position was reanalysed as a pre-verbal predicative position where complex predicates are formed. I adopted the proposals for Modern Hungarian which assume that this position is PredP, the specifier of which is the host of predicative phrases, while the verb moves into the Pred head. I further assumed that this is the site of complex predicate formation, which takes place in overt syntax in Hungarian, and that this position was already present in Old Hungarian. Old Hungarian particles, therefore, form a complex predicate with the verb and are moved to the pre-verbal position because of their predicative nature. This movement seems to be obligatory as the slight variation that we find in their ordering can be attributed to different independent factors. Predicate movement into the pre-verbal position and thereby complex predicate formation in overt syntax does not seem to be obligatory with the other predicative elements. We find variation in their word order properties that cannot be put down to independent factors, and thus seem to call for an explanation in terms of optionality in predicate movement. I proposed that after the reanalysis of the word order to SVO, internal arguments could appear post-verbally and it was the movement of particles that triggered the generalization that all predicates should be preverbal but this involved a period when the rule was not generalized to all predicates yet, and this allowed for word order variation. The exact formulation of the trigger for predicate movement and how it allows for variation, as well as the extent of variation at later stages, needs further investigation.
Acknowledgements The research presented here has been funded by the Hungarian Scientific Research Fund (OTKA No. and No. ); their support is gratefully acknowledged. I would also like to thank the audience of the th Diachronic Generative Syntax Conference and two anonymous reviewers for their useful comments.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Veronika Hegedűs
Old Hungarian sources Jókai Codex: After / around . P. Balázs, János (ed.), Jókai-kódex. (Transcription of the original record, the corresponding Latin text, with introduction and notes) Budapest: Akadémiai Kiadó, . Jordánszky Codex: –. Toldy, Ferenc and György Volf (eds), A Jordánszky-kódex bibliafordítása. Buda, . Munich Codex: . Nyíri, Antal (ed.), Müncheni kódex. (Critical edition with the corresponding Latin text) Budapest: Akadémiai Kiadó, .
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
7 An effect of residual T-to-C movement in varieties of English JUDY B. BERNSTEIN
. Introduction In this chapter I examine discontinuous subjects in varieties of English. In particular, I consider examples that display an expletive pronoun linked to a quantificational expression and an intervening modal or finite auxiliary. Examples of the phenomenon in Older Scots are provided in () (examples from DSL: Dictionary of the Scots Language ): Older Scots: () a. Thar suld na man hyd fra the face of God. (DSL, s.v. hide, v.) there should no man hide from the face of God b. . . . thar may na power do thaim dammage . . . there may no power do them damage
(DSL, s.v. power, n.)
The expletive pronoun (thar) is linked to a quantificational expression (na man, na power) and a modal or finite auxiliary (suld, may) intervenes. Recognized as a minority language by the European Charter since , Scots descended from the Old Northumbrian dialect of Old English (Macafee ). The label ‘Older Scots’ is applied starting from the mid-s (the period of the earliest records) through , after which the label ‘Modern Scots’ is applied.1 Older Scots is an ancestor variety to Appalachian English (Montgomery , ), a group of varieties spoken in the region of the Appalachian Mountains in the southeastern US. Montgomery details two migrations, the first from the Scottish Lowlands to Ulster, in Ireland, and the second from Ulster to America beginning in the late s. By the time of the American Revolution in , this migration had The subdivisions of periods of Older Scots include pre- to literary Scots (to ), Early Scots (to ) and Middle Scots (–; Early Middle Scots: –; Late Middle Scots: –). See Macafee (), incorporating material by the late A. J. Aitken). 1
Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Judy B. Bernstein . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
brought approximately , Scotch-Irish immigrants to America, many of them settling in the area now known as Appalachia (Montgomery : ).2 As in Older Scots, Appalachian English discontinuous subjects display an expletive pronoun (there) linked to a quantificational expression (nobody, anyone), as well as an intervening modal or finite auxiliary: Appalachian English: () a. ‘He’s stubborn and wild,’ Ephraim said, ‘There can’t nobody ride him.’ ‘Nobody can ride him.’ (Montgomery and Hall : ) b. They asked four men, but there didn’t anyone want to leave their church ‘but nobody wanted to leave their church’ (Feagin : ) So Appalachian English and Older Scots discontinuous subjects share the following properties: (a) an expletive pronoun appears in clause-initial position; (b) a quantificational expression (indefinite, overwhelmingly negative) is the thematic subject; and (c) a modal or finite auxiliary is required. Building on prior work with Raffaella Zanuttini on Appalachian English, I make the following observations and proposals: A. Older Scots and also Appalachian English discontinuous subjects are instances of transitive expletives. B. Older Scots displays evidence of a verb second (V) grammar, one in which the finite verb raises to C and a Topic phrase raises to the specifier of CP (a CP-V grammar). This is consistent with the northern Middle English CP-V grammar described in Kroch et al. () and Kroch and Taylor (). C. Older Scots transitive expletives are also an instance of CP-V: the modal or finite auxiliary raises to C and the expletive pronoun occupies the specifier of CP and functions as an expletive Topic. D. Appalachian English (as well as other varieties of North American English) displays so-called negative auxiliary inversion, equivalent to () and () except that no overt pronoun is expressed. I follow Green’s () proposal for African American English that the modal or finite auxiliary raises to C, which hosts a Focus feature. I extend her analysis to Appalachian English transitive expletives. The chapter is organized as follows: In section . I examine discontinuous subjects in Older Scots and compare them to the transitive expletives found across the Germanic languages. I also briefly examine discontinuous subjects in Appalachian English. I conclude that in both languages, discontinuous subjects are instances of transitive expletives. In section . I show that Older Scots displays classic properties of a V grammar, specifically one in which the finite verb raises to C and a Topic to 2 This area includes southwestern Pennsylvania and Virginia, all of West Virginia, eastern Tennessee and Kentucky, western North Carolina, southeastern Ohio, and northern Georgia and Alabama (see Algeo ).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
the specifier of CP. In this way it patterns with other CP-V languages like Mainland Scandinavian, German, Dutch, and northern Middle English. I also propose that Older Scots transitive expletives are an instance of V. In section . I study so-called negative auxiliary inversion in Appalachian English and argue that it is equivalent to transitive expletives without an overt expletive. I follow Green’s () analysis of African American English negative auxiliary inversion and argue that both sentence types involve verb raising to Focus, a functional projection within Rizzi’s () articulated CP. Unlike the general V found across Germanic languages, which involve Topic not Focus, the cases of residual V in Appalachian English involve Focus. Some concluding remarks are made in section ..
. Discontinuous subjects in Older Scots and Appalachian English In this section I examine the salient properties of discontinuous subjects in Older Scots and compare their distribution with that of transitive expletives, a construction type that has been studied extensively in the literature of the Germanic languages. I will propose that the discontinuous subjects found in Older Scots are an instance of transitive expletives. Most of the properties observed to occur in Germanic languages with transitive expletives are also found in Older Scots. I further propose that discontinuous subjects in Appalachian English are instances of transitive expletives, a sentence type absent in general American English. .. Older Scots discontinuous subjects as transitive expletives Transitive expletives are a sentence type familiar from work on the Germanic languages (see Bures ; Vikner ; Bobaljik and Jonas ; Jonas ; Holmberg and Platzack ; Richards , among others). I illustrate the general pattern in Icelandic (example from Bobaljik and Jonas ): () ϸað hafa margir jólasveinar borðað búðing. there have many Christmas.trolls eaten pudding ‘Many Christmas trolls have eaten pudding.’ Transitive expletives have been characterized as involving ‘multiple subjects’ since the expletive (ϸað) and a quantificational expression (margir jólasveinar), although contributing a single interpretation, occupy different positions in the structure. In this Icelandic example, the finite auxiliary occupies a position between the expletive pronoun and a quantificational subject, the associate. The same general pattern is seen in Older Scots:3 () Thar myght na molet mak me moy na hald my mouth in there might no scold’s.bridle make me meek nor restrain my mouth PRT (DSL, s.v. thar(e), adv.)
3
All Older Scots examples are from Dictionary of the Scots Language unless otherwise noted.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
In this example as well, a finite verb in the form of a modal (myght) occurs between an expletive (thar) and a quantificational expression (na molet), which together form the sentential subject. The lexical verb (mak) is transitive. Another similarity between Icelandic and Older Scots is that transitive expletives are not restricted to transitive verbs. Jonas shows that expletive constructions with lexical predicates in Icelandic appear with various types of verbs (Jonas : ): () a. ϸað hafa nokkrar kökur verið bakaðar fyrir veisluna. there have some cakes been baked for the.party (Jonas , (b–d)) b. ϸað hlupu fimm troll yfir brúna. there ran five trolls over the.bridge c. ϸað komu margir stúdentar í partíið. there came many students to the.party This resembles the range of expletive constructions with lexical predicates found in Older Scots: () a. thare sal na dene of gilde be chosin for this yere (DSL, s.v. thar(e), adv.) there shall no dean of guild be chosen for this year b. They’ll something wamble in your wame there.will something wriggle in your belly c. thar had few Inglis men past away; there had few English men passed away
(DSL, s.v. there, adv.) (DSL, s.v. partie, n.)
So at least in these two respects, Older Scots transitive expletives pattern with the Icelandic ones: a) a modal or finite auxiliary appears between an expletive and a quantificational expression; b) the lexical verb is often transitive, but does not have to be. There are differences between the languages as well. One difference between Icelandic and Older Scots transitive expletives is that we find no instances of transitive expletives with finite lexical verbs in Older Scots, but only with modals and finite auxiliaries.4,5 Icelandic transitive expletives readily appear with finite lexical verbs (example from Bobaljik and Jonas ): () ϸað borðuðu sennilega margir jólasveinar bjúgun. there ate probably many Christmas.trolls the.sausages ‘Many Christmas trolls probably ate the sausages.’
4
The following illustrates an Older Scots expletive construction with a finite auxiliary (‘had’):
(i) thar had few Inglis men past away there had few English men passed away
(DSL, s.v. partie, n.)
In this example, the lexical verb is unaccusative. 5 The absence of finite lexical verbs in Older Scots transitive expletives resembles those described for Late Middle English/Early Modern English in Jonas (), Ingham (, ), and Tanaka ().
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
Another difference between Icelandic and Older Scots is in the form of the expletive pronoun. Besides the expected examples with expletive there, Older Scots displays the same sentence type with the pronoun they: () a. Thay will na there will no betuix thir between these
man of jugement or learning mak difference man of judgement or learning make difference wordis (DSL, s.v. thar(e), adv.) words
b. They’ll something wamble in your wame. there.will something wriggle in your belly
(DSL, s.v. there, adv.)
Although we can straightforwardly take there to be an expletive pronoun in Older Scots, they is potentially ambiguous between an expletive pronoun and a personal pronoun. Montgomery () studies the historical trajectory of Older Scots expletive there and they, the likely source of the identical Appalachian English forms (see section ..). Although potentially ambiguous, I take Older Scots they to be an expletive pronoun in the cases under consideration. This is supported by the fact that we do not find these multiple subject constructions, which I will treat as transitive expletives, with other personal pronouns such as you and we (but see note ). A second reason for taking they to be an expletive pronoun in () has to do with number agreement or lack thereof. In (a,b), they is paired with a singular quantificational associate (na man, something). Taken as an expletive, no number mismatch occurs, but as a personal pronoun a plural co-indexed form would be expected. A plural co-indexed form is exactly what we find in the following Older Scots examples where the possessive their matches referential they in person and number: () a. Gyf thai oyse nocht their offycys wyth their awyn handis if they use not their offices with their own hands (DSL, s.v. hand, n.) b. Thai suld weill hawe prys That in thar tyme war wycht and wys they should well have value that in their time were strong and wise (DSL, s.v. thair(e), poss. adj.) Although perhaps less obvious, the following Older Scots example also displays the referential pronoun (thai), not the expletive pronoun: () Thai sulde hyr hous bryn And all that ewyre thai fand thare in. they should her house burn and all that ever they found there in (therein) (DSL, s.v. ever, adv.) If glossed over too quickly, thai in () might be interpreted erroneously as an expletive and paired with the lexical subject hyr hous. However, in that case an expletive would be paired with a definite expression, hyr hous, rather than a quantificational one. Instead, thai in () should be interpreted as the agent of the lexical verb bryn (‘burn’) and linked to referential thai in the conjoined sentence. Yet a third way to disambiguate the expletive pronouns from the referential ones is verbal agreement: the verbal form co-occurring with the expletive (easily observed in
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
existential constructions) is overwhelmingly the -s (singular) verbal form while that co-occurring with the referential pronoun is the bare (plural) verbal form (see Montgomery for details). These three arguments support treating Older Scots they as an ambiguous form, an existential or a referential pronoun. Returning to the accounting of differences between Icelandic and Older Scots transitive expletives, I have observed that the Older Scots cases generally involve a negative quantificational expression. No such pattern of negative quantification has been described for Icelandic, but the Older Scots pattern has been observed for Late Middle English and Early Modern English transitive expletives (see Ingham , ). The pattern also characterizes Appalachian English, as we saw in section ... As a reviewer notes, the overwhelmingly negative quantificational associates in the English varieties under discussion here constitute an intriguing subset of the sorts of associates permitted in Icelandic transitive expletive constructions.6 So although the Older Scots expletive constructions illustrated above strongly resemble those of Icelandic, the properties are not identical. I have identified three differences between Icelandic and Older Scots transitive expletives: ) an absence of transitive expletives with finite lexical verbs in Older Scots; ) an apparent second expletive pronoun (they) in Older Scots; and ) a strong tendency in Older Scots for the quantificational expressions to be negative. Besides the specific differences between Icelandic and Older Scots, the literature on transitive expletives in Germanic languages has identified a correlation between their presence and other properties of the grammar. Holmberg and Platzack (, ) discuss a correlation between the presence of transitive expletives and properties such as rich verbal morphology, V-to-T raising, object shift, stylistic fronting, V, and the possibility of a null expletive. Older Scots displays several of these properties, such as V-to-T raising, object shift, and stylistic fronting, as illustrated in the following examples): V-to-T movement (embedded contexts): () a. That he gaif never his sister the chaines that he gave never his sister the chains b. That that c. That that
he he he he
chape nocht buy not risis takes.up.arms
(DSL, s.v. never, adv.)
micht (DSL, s.v. never, adv.) power nocht with the Kingis officeris not with the King’s officers (DSL, s.v. nocht, n., adv.)
Object shift: () a. Thai ar commandit to revele it nocht they are commanded to reveal it not
(DSL, s.v. nocht, n., adv.)
6 See Vangsnes () for a detailed analysis of the types of associates allowed in Icelandic transitive expletives (occurring in what he labels ‘the intermediate position’). This contrasts with the more restrictive types of associates allowed post-verbally, those subject to the ‘definiteness effect’ described for Mainland Scandinavian and English.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
b. All lyke hir nocht all like her not.
(DSL, s.v. nocht, n., adv.)
c. He coft it nocht he bought it not
(DSL, s.v. nocht, n., adv.)
Stylistic fronting: () a. Wark beares witnes quha weill dois work bears witness who well does
(DSL, s.v. wele, adv., interj.)
b. Quha best did than, he had the heast stait who best did then, he had the highest state (DSL, s.v. quha, interrog., rel. pron.) c. Al men that euir has bene (DSL, s.v. ever, adv.) all men that ever has been In addition to the three properties illustrated in ()–(), Older Scots exhibits V, which I examine in some detail in section .. So although Older Scots displays many of the properties found in other transitive expletive languages, it seems to lack two, namely rich verbal morphology and the possibility of a null expletive.7 As a reviewer points out, it is not clear what role, if any, verbal morphology plays. Other Germanic languages lacking rich verbal morphology, such as Late Middle/Early Modern English (Tanaka ) and Afrikaans (Conradie ), have nevertheless been reported to display transitive expletives. Overall, when we consider together the empirical similarities between Older Scots and Icelandic expletive constructions and the syntactic properties in Older Scots that correlate with Germanic transitive expletive constructions, we find that the Older Scots discontinuous subjects display a substantial subset of the expected patterns. I conclude therefore that they are indeed instances of transitive expletive constructions. After examining Older Scots V in section ., I will propose that Older Scots transitive expletives are an instance of V. But first I briefly compare transitive expletives in Older Scots to what is found in its descendant, Appalachian English. .. Discontinuous subjects in Appalachian English Discontinuous subjects in Appalachian English resemble the transitive expletives of Older Scots. We find the presence of expletive there, followed by a modal or finite auxiliary, followed by a quantificational expression: () a. He’s a ole feller an’ has been stillin’ for ten years I know of, an’ there don’t never nobody bother his still. (Montgomery and Hall : ) ‘nobody ever bothers his distilling apparatus’
7 Although Older Scots verbal agreement cannot be characterized as rich, it does differ from that of the southern varieties of Middle English. The agreement pattern falls within what has been labeled the ‘Northern Subject Rule’ (see Bernstein and Zanuttini ). As far as the possibility of a null expletive (in non clause-initial positions, relevant here), I have not discovered any.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein (Montgomery and Hall : )
b. There can’t nobody ride him. ‘Nobody can ride him.’
As we saw for Older Scots, in addition to expletive there we find examples with they (), which in these contexts I also take to be a form of the expletive (see Montgomery , Tortora on expletive they in Appalachian English):8 () a. they didn’t nothing go right ‘nothing went right’ b. They can’t many people say that. ‘not many people can say that’
(Feagin : ) (Dante Oral History, Shearer –) (Feagin : )
c. They won’t nobody know you’re gone. ‘nobody will know you’re gone’
As seen above for Older Scots, we can disambiguate the expletives and the referential pronouns. In Appalachian English, referential they in () requires a plural form of the verb; expletive they in () does not: () They are learning so much here now. (Dante Oral History, Shearer –) () They is something bad wrong with her.
(Montgomery and Hall : lxii)
As we saw above for Older Scots and Icelandic transitive expletives, Appalachian English expletive constructions are not limited to transitive verbs: ()
a. They didn’ nobody live up there. (Feagin: : ) b. there wouldn’ nothin’ go down through there (Feagin: : ) c. There’d somebody come around with a truck once in a while (Montgomery and Hall : )
And although modals are found in the majority of the examples, finite auxiliaries have been found in a few, as in (a) above and also (). () There didn’ no girl go to the door with a boy.
(Feagin: : )
Finally, the Appalachian English discontinuous subjects, like their Older Scots counterparts, are overwhelmingly negative. In the Appalachian English examples, negation is usually expressed in both the quantificational expression and in the modal or finite auxiliary (via the morpheme n’t), a case of negative concord. I return to this issue in section . when I examine negative auxiliary inversion in Appalachian English. So in many respects Appalachian English discontinuous subjects resemble their Older Scots counterparts. Since I take the Older Scots cases to be instances of
8 Consider this one isolated example with we in Appalachian English (from Montgomery and Hall : ): We don’t nobody know how long we have. Native speakers report that this example is only possible on a non-referential interpretation of the pronoun we. The interpretation of the sentence would be a statement about the human condition rather than about a specific set of people that includes the speaker.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
transitive expletives, I will extend the analysis to the Appalachian English examples.9 It is important to keep in mind that, unlike present-day Germanic languages and also Older Scots, Appalachian English does not display the clustering of properties that correlate with transitive expletives. There is no rich verbal morphology, no V-to-T movement with lexical verbs, and no object shift. There is also no productive V. Nevertheless, I propose that these Appalachian English examples involve transitive expletives as well: the modal or finite auxiliary raises over the thematic subject to C and the expletive occupies the specifier of CP. The proposal that the expletive in Appalachian English transitive expletives does not occupy Spec,TP, the canonical position for expletive subjects in contemporary varieties of English, would be nicely supported by absence of inversion in questions, assuming that such inversion involves V-to-C. Unfortunately, the judgements are not consistent. Of three Appalachian English speakers I consulted, two accepted inversion in examples like (b) and one rejected it. The status of inversion with expletive subjects in existential constructions is more straightforward: examples like () are easily found. (Montgomery and Hall : )
()
a. There can’t nobody ride him. b. (*)Can’t there nobody ride him?
()
And they’s been these plants come in like that, different ones all along, ain’t they? (Mountain City, TN; author’s et al. fieldwork)10
What emerges from this discussion of Appalachian English transitive expletives is the observation that, like its Older Scots ancestor and also Late Middle and Early Modern English, the robustness of the Icelandic phenomenon is not to be found in varieties of English, which only display a subset of properties. Recall that a similar observation was made about Afrikaans in Conradie (). In the next section, I examine the Topic function of V in Older Scots and Germanic more generally and consider a V structure as an approach to Older Scots transitive expletives. Although V topicalization is no longer productive in present-day varieties of English, I’ll follow the idea in section . that the verb raises to C in declarative clauses in some varieties of English.11
9
Henry and Cottell () examine transitive expletives in present-day Belfast English and consider the pattern to be an innovation, since they have observed them mainly among younger speakers. Some of the properties they describe for Belfast English do not pattern with Appalachian English or Older Scots. Furthermore, what I have taken to be transitive expletives in Appalachian English cannot be seen as an innovation since they are found among older speakers (as well as younger speakers). I will not further discuss Belfast English here, since my goal is to examine the Older Scots pattern and make the connection to Appalachian English. 10 The example in () was collected in during NSF-funded fieldwork conducted by Judy Bernstein, Marcel Den Dikken, Christina Tortora, and Raffaella Zanuttini, and graduate students of Den Dikken, Tortora, and Zanuttini. 11 As is well known, a Topic can be expressed clause initially in some present-day varieties of English, such as that of the author; however, the finite verb cannot raise over the subject: (i) A millionaire she’s not. (cf. *A millionaire is she not.) (ii) A million dollars I don’t have. (cf. *A million dollars don’t I have.)
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
. Verb second in Older Scots In this section I examine the data illustrating that Older Scots has a V grammar. Specifically, the properties match those described for the northern varieties of Middle English, which pattern like Mainland Scandinavian, German, and Dutch (Kroch et al. ; Kroch and Taylor ). In these languages, labelled ‘CP-V’, the evidence supports the hypothesis that the verb raises to C. The southern varieties of Middle English as well as Icelandic and Yiddish are also V; however, in these languages the verb raises only as far as I (Pintzuk ). In this section I explore the nature of the evidence for V, specifically CP-V, in Older Scots. Kroch et al. () and Kroch and Taylor () distinguish V in northern and southern dialects of Middle English. The structure of V in northern Middle English varieties, CP-V, involves movement of the verb to C and a phrasal constituent, a Topic, to the specifier of CP (Kroch et al. : ). In contrast, in southern Middle English varieties V involves movement of the verb to I and a phrasal constituent, the Topic, to the specifier of IP (Kroch et al. : ). Kroch et al. argue that imperfect language learning by the Scandinavian invaders, and not direct borrowing (since the native language was probably IP-V), led to the development of CP-V in northern varieties of English (Kroch et al. : ). The idea is that the imperfect learning of English led to a collapse in the verbal agreement markers and IP ceased to be available as a landing site for the tensed verb, which shifted to CP (Kroch et al. : ).12 According to Kroch and Taylor (: –, –) there are at least two patterns consistent with a CP-V grammar. The first is that inversion is displayed with both lexical and pronominal subjects. They reason that this cannot be an effect of the syntactic status of the pronouns, which they hypothesize to be the same as in the southern varieties. Instead, they take the near categorical inversion with pronouns in V contexts to result from the movement of the verb to a higher position, namely C, in the northern English varieties. A second property of CP-V languages is an absence of V in most complement clauses except with the complement of bridge verbs (verbs like think, say, believe; Kroch et al. : ), since the structure of the complements of the bridge verbs is like that of matrix clauses. In regular complement clauses, the complementizer itself would occupy the C position, rendering the position unavailable for a raised verb. Unlike the CP-V languages, IP-V languages like Yiddish and Icelandic display inversion in V contexts almost exclusively with lexical subjects. Furthermore, V is displayed in a wider range of complement clauses. Since the verb raises only as high as I (now T) in languages like Icelandic, Yiddish, and southern Middle English, there is no competition for the C position occupied by the complementizer in embedded clauses.
12 This would be consistent with the Older Scots pattern in matrix clauses; however, I note that Older Scots exhibits V-to-T raising in embedded contexts (recall ()), which is not predicted by Kroch et al. () for northern Middle English. Perhaps V-to-T was retained in Older Scots embedded clauses and lost in matrix clauses, but it’s not obvious why that would be.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
Following the pattern expected of a CP-V language, Older Scots displays inversion with lexical and pronominal subjects, as illustrated in () and ():13 V with lexical subjects: () a. And sa remanis the ile allanerlie occupijt with ws Scottis in and so remains the island only occupied by us Scots in Scotland (DSL, s.v. sa, adv., conj.) Scotland b. Than then help help
was the schiref of the kingis danger quite Be the greit was the sheriff of the king’s displeasure repaid by the great of this lady sa quhite (DSL, s.v. quite, adj.) of this lady so freed
c. And sue is the compter superexpendit vj li. xvj s. vj d. and so is the accountant overspent £.. (DSL, s.v. sa, adv., conj.) V with pronominal subjects: () a. To Scotland went he than in hy (DSL, s.v. than, adv.) to Scotland went he then in haste b. Tharfor departis ws in thre All sall nocht sa assailyt be & in therefore depart us in three all shall not thus assailed be and in thre partis hald our way (DSL, s.v. sa, adv., conj.) three parts hold our way c. Yhet has he nocht sa mekill fre, As to do That at hys hert yet has he not so much free, as to do that which his heart hym drawis to (DSL, s.v. fre, adj., n.) him draws to d. This cane thai sa sorofully this can they say sorrowfully
(DSL, s.v. say, v.)
e. Confes cleir can I nocht, nor kyth all the cas confess clear can I not, nor reveal all the case (DSL, s.v. nocht, n., adv.) f. With kirk-gude sal I neuer haue ado (DSL, s.v. never, adv.). with church-property shall I never have to.do g. Sa neid ye not to doute I uill aduance him (DSL, s.v. sa, adv., conj.) so need you not to doubt I will advance him
13 Since there is no parsed corpus of Older Scots, I was unable to determine frequencies of the patterns discussed. I chose the earliest Older Scots examples of the phenomenon in order to be parallel to the northern Benet text discussed in Kroch and Taylor () and Kroch et al. (), and to avoid as much as possible the mixture of V and non-V sentences described for the northern texts of around (Kroch and Taylor : ).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
Similarly and focusing on a single text, The Prose Works of Sir Gilbert Hay () from the Helsinki Corpus of Older Scots (HOCS), we can see robust evidence of V with non-subjects in matrix clauses (see ()), as well as inversion with pronominal subjects (see ()). V in matrix clauses (HOCS, Sir Gilbert Hay, ): () a. And thus suld the warld be gouernyt be common reugle of and thus should the world be governed by common rule of gouernaunce governance b. ffor than ar’ kingis ande princis caus’ of the abusioun of the ordre of for then are kings and princes cause of the misuse of the order of knighthood knychthede c. and thus ar’ the kingis dred for the knychtis and thus are the kings intimidating for the knights inversion with pronominal subjects in matrix clauses (HOCS, Sir Gilbert Hay, ): () a. Now lykis jt him to speke of the office yat folowis the said ordre. now likes it him to speak of the office that follows the said order. b. And thar’fore ordanyt he jn this warld mony of knychtis to be . . . and therefore ordained he in this world many of knights to be . . . c. And thar’for’ wald he nocht . . . and therefore would he not . . . d. ffor rycht as a juge has powar’ by for just as a judge has power by sentence rycht sa has he power’ opinion just so has he power
his his to to
office office kepe keep
to juge and geue a to judge and give an his jugementis his judgments
The second property of CP-V languages, an absence of V in embedded clauses, is harder to establish definitively for two reasons: a) it is difficult to learn whether any cases of subject + finite verb involve V; b) I am unable to examine every embedded clause in the Dictionary of the Scots Language or in the unparsed Helsinki Corpus of Older Scots. Nevertheless, my preliminary analysis of the texts yielded no examples of embedded V with non-subject Topics. Having provided some evidence of a CP-V grammar in Older Scots, I return now to the sentences with transitive expletives examined in section ... I repeat the examples from () here: () a. Thar suld na man hyd fra the face of God (DSL, s.v. hide, v.) there should no man hide from the face of God b. thar may na power do thaim damage there may no power do them damage
(DSL, s.v. power, n.)
I observe that these examples resemble those in ()–(). That is, in () as well, the finite verb, in the form of a modal or finite auxiliary, occupies the second position in
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
the clause and precedes the thematic subject. The only difference between the earlier sentences and the transitive expletives in () is that we find an expletive (thar) filling what should be the clause-initial Topic position rather than an adverbial, prepositional phrase, NP complement, or lexical predicate. Can Older Scots expletive thar be treated as a Topic? I think so, and this position is in line with that of Rögnvaldsson and Thráinsson (: –, discussed in Iatridou and Kroch : ) for Icelandic. Rögnvaldsson and Thráinsson claim that the expletive pronoun in Icelandic occupies the same specifier position as a lexical Topic phrase.14 Jonas (: –) discusses the idea (of Thráinsson’s and others) that the Icelandic expletive ϸað might be inserted to satisfy the V constraint. So I extend this idea to Older Scots and suggest that the expletive in examples like () occupies the position of a lexical Topic phrase, specifier of CP. I will refine this idea in section . and suggest that the expletive fills the specifier of TopicP, a functional projection in the articulated CP structure of Rizzi (). The articulated CP structure will also allow a way to distinguish V involving Topic from V involving Focus, the latter relevant in wh-movement contexts in the languages under consideration. So aligning the Older Scots transitive expletive examples in () with the more familiar cases of V we saw above in ()–(), I take the finite verb (e.g. modal suld) to have raised over the thematic subject to the C position. The Older Scots expletive pronoun (there or they), as an expletive Topic, fills Spec,CP and the quantificational thematic subject occupies a lower subject position, presumably Spec,TP. A sketch of the internal structures I attribute to regular V and transitive expletives as V are illustrated in (a) and (b), respectively. () a. [CP thus [C thus b. [CP thar [C there
suldi [TP should suldi [TP should
the warld [T ti . . . verb second the world . . . na man [T ti . . . transitive expletives as verb second no man . . .
I mention here two interesting differences between regular V and transitive expletive V in Older Scots: ) whereas regular V is found with finite lexical verbs and modals or finite auxiliaries, transitive expletive V is only found with modals or finite auxiliaries; ) whereas regular V is found in affirmative and negative contexts, transitive expletive V is found almost exclusively with negative thematic subjects. I have no real explanation for these differences between Older Scots regular V and transitive expletive V. While I don’t yet know about the history of transitive expletives in Older Scots, it is interesting to speculate based on the information available. For example, transitive expletives appeared in (southern) Late Middle/ Early Modern English in the fourteenth century and disappeared during the sixteenth century (Tanaka : –).15 Let’s suppose that in parallel fashion they also 14 The relevant specifier position would not be CP in Icelandic, since this language is an IP-V language. See Thráinsson () for extensive discussion of the position of the expletive (specifier of IP vs specifier of CP) and whether it should be treated as a topic. 15 A reviewer points out that southern English probably lost V around and that this date does not neatly line up with the appearance of transitive expletives, which would have outlived the more general V. Although I’m not sure what to make of this, the facts might support aligning the Older Scots transitive
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
entered Older Scots as an innovation, a new subtype of V. Then perhaps their distinct properties (absence of finite lexical verbs, presence of negative thematic subjects) are tied to the pragmatics of the innovation as well as the gradual loss of lexical verb raising in Older Scots. I return to some of these issues in the next section, after a discussion of Appalachian English negative auxiliary inversion.
. Negative auxiliary inversion in Appalachian English So far I have examined transitive expletives in Older Scots and proposed that they are an instance of a more general V grammar, one in which the finite verb raises to C and an expletive Topic fills the Spec,CP position. I have also explored the idea that Appalachian English transitive expletives resemble the Older Scots forms in several respects, and yet the language does not display any general evidence of a V grammar and also lacks the other properties shared among the contemporary Germanic languages with transitive expletives. So is there any independent evidence for thinking that Appalachian English transitive expletives involve verb raising to C? I will return to this question after I briefly examine so-called negative auxiliary inversion (NAI) in Appalachian English. Negative auxiliary inversion (labelled as such in Green ) is a sentence type found in several varieties of North American English. It has been studied extensively in the literature on African American English (in Green , ; Labov et al. ; Martin ; Sells et al. ; Parrott ; and White-Sustaíta ). It has also been studied in Appalachian English (in Wolfram and Christian ; Feagin ; Montgomery and Hall ) and West Texas English (in Foreman ). The following Appalachian English examples illustrate the pattern: () a. Didn’t nobody get hurt or nothin’. (Wolfram and Christian : ) ‘Nobody got hurt.’ (= it is not the case that someone got hurt) b. Can’t but two people sit in back. ‘Only two people can sit in back.’
(Feagin : )
NAI displays a negated modal or finite auxiliary followed by an indefinite quantificational subject, which is overwhelmingly negative. Although a sentence-initial modal or finite auxiliary preceding a thematic subject is a word order also characterizing English yes/no questions, NAI can only be used to make an assertion (see Foreman for a discussion of the differences between NAI and yes/no questions). I would like to tie this sentence type to the Appalachian English transitive expletives from section ... I repeat the examples from () here: () a. He’s a ole feller an’ has been stillin’ for ten years I know of, an’ there don’t never nobody bother his still. (Montgomery and Hall : ) ‘ . . . nobody ever bothers his still’ expletives with those of Appalachian English. I argue in the next section that these involve Focus, not Topic, and I suggest this might be a result of language change.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English b. There can’t nobody ride him. ‘Nobody can ride him.’
(Montgomery and Hall : )
We can observe that the only surface difference between transitive expletives and NAI is the presence of an expletive subject (there or they) in transitive expletives in () and its absence in NAI in (). As we saw for the transitive expletives earlier, Appalachian English displays NAI with transitive verbs, in (), as well as with other verb types, in (): Negative auxiliary inversion with transitive verbs: () a. Don’t nobody like him. (Wolfram and Christian : ) b. Hain’t nobody hardly believed it. (Wolfram and Christian : ) c. Didn’t anybody have enough sense to teach us up here. (Montgomery and Hall : ) d. And won’t anybody know who you are until I tell ’em. (Feagin : ) e. Won’t nobody help her. (Feagin : ) Negative auxiliary inversion with other verb types: () a. There’s an old house up there but don’t nobody live in it. (Montgomery and Hall : ) b. It had this room that wouldn’t nobody stay in. (Wolfram and Christian : ) A final comparison of the two sentence types illustrates that they may be found within the grammar of single speaker.16 The example in () was produced by one speaker and illustrates an instance of NAI immediately followed by an equivalent transitive expletive. ()
Back in them days, didn’ nobody live there but just a little while, did they? I mean, though, back in them days, they didn’ nobody live up there. (Feagin : )
I would like to align Appalachian English NAI and transitive expletives syntactically. To do this, I begin by generalizing Green’s () analysis of NAI in African American English to the Appalachian English data. In agreement with observations made by Labov et al. (), Green maintains that sentences with NAI are more emphatic than their counterparts without inversion. Like African American English, Appalachian English displays the sentences without inversion: No inversion:17 () a. Nobody didn’t have it that I know of, but her? (Dante Oral History, Shearer –)
16 Appalachian English speakers who allow transitive expletives also allow negative auxiliary inversion, but the reverse is not the case. 17 The examples in (b,c) were collected in during NSF-funded fieldwork conducted by Judy Bernstein, Marcel Den Dikken, Christina Tortora, and Raffaella Zanuttini, and graduate students of theirs.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
b. Nothing didn’t never bother me yet. (Mountain City, TN; author’s et al. fieldwork) c. And you couldn’t, nobody couldn’t do nothing with that bull . . . (Mountain City, TN; author’s et al. fieldwork) d. And neither of the boys can’t play a lick of it. (Feagin : ) Green () characterizes the emphatic interpretation of NAI as ‘absolute negation’.18 She argues that African American English examples equivalent to ()–() involve movement of the negative modal or finite auxiliary to C, which hosts a negative Focus feature. From this position, the modal or finite auxiliary takes scope over the indefinite quantifier. To do this, she appeals to the articulated CP structure of Rizzi (: , and later work): ()
[ForceP . . . [TopicP . . . [FocusP . . . [FinitnessP . . . [IP . . .
Although Green () develops an analysis whereby African American English NAI involves movement of a negated modal or finite auxiliary to Focus within an articulated CP structure, she does not extend the analysis to cases of transitive expletives, which are not part of contemporary African American English. Nevertheless, they were attested in older stages of the language, such as in ex-slave narratives: () a. Dey don’ nobody know whar, suh. they don’t nobody know where, sir (literary source from , cited in Green : ) b. But they’d give me a note so there wouldn’ nobody interfere with me. so there wouldn’t nobody interfere with me (ex-slave narrator born in , cited in Bailey, Maynor, and Cukor-Avila : ) c. dey didn’ nobody hab ter stan’ over ’em they didn’t nobody have to stand over them (literary, written at the turn of the twentieth century, Chesnutt and Sollors , cited in White-Sustaíta : ) So although I don’t know why African American English no longer allows overt expletives, I propose that both NAI and the (now archaic) transitive expletives in African American English reflect raising of the negated modal or finite auxiliary to Focus in the CP domain. In the case of the transitive expletives, the expletive fills the specifier of FocusP. Returning to Appalachian English, I would like to carry over the idea that transitive expletives and NAI are two instances of the same phenomenon. So in both sentence types, a negated auxiliary or modal raises over the quantificational subject to C, understood as Focus. In the transitive expletive cases, an expletive
18 For an example like (a), Don’t nobody like him, the ‘absolute negation’ interpretation would be ‘not even one person likes him’.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
pronoun occupies the specifier of FocusP; with NAI, there is no overt material in the specifier position. I speculate that some grammatical change in the languages with NAI allowed the expletive to remain unexpressed.19 I observe that the expression of the expletive is not required in Appalachian English and is apparently unavailable in present-day African American English. In retaining the possibility of expressing an expletive pronoun, Appalachian English appears to represent the more conservative variety. I return briefly to Older Scots in order to reconsider some issues that arise in light of our discussion and analysis of Appalachian English transitive expletives and NAI. If transitive expletives and NAI are related to each other, why does Older Scots display transitive expletives and not NAI? I am not sure, but I note that NAI is obviously a more recent development, found only in varieties of North American English (White-Sustaíta ), but not in standard American English. NAI is not found among other Germanic languages with transitive expletives either. So it is plausible to imagine that some grammatical change paved the way for NAI. This was perhaps a reorganization of the features encoded in the expanded CP domain.20 Recall that in section . I analysed Older Scots transitive expletives as an instance of V, specifically CP-V. As I mentioned earlier, Appalachian English does not exhibit V of the sort found in Older Scots or the Germanic languages more generally. And yet I am claiming that a modal or finite auxiliary moves to C (Focus) in cases of NAI and transitive expletives in Appalachian English. So how do I distinguish Appalachian English verb raising to C from Older Scots verb raising to C? Appealing again to the articulated CP of Rizzi (; recall ()), I take Older Scots V to involve movement of the finite verb to Topic and the topicalized constituent to its specifier. In parallel fashion, the transitive expletives in Older Scots display an expletive Topic in the specifier of TopicP. So what accounts more generally for the difference in V between the two languages, Older Scots and Appalachian English? It seems to me that the difference reduces to the features that can be encoded in C in the two languages. In Older Scots, Topic is encoded; in Appalachian English, Focus is encoded. So generalized V in Older Scots, including transitive expletives, involves Topic, and V in Appalachian English (transitive expletives and NAI) involves Focus.21 Since Focus would also be the relevant feature for wh-elements, it would be relevant in both languages, since both display V in wh- contexts.22 But only Older Scots would additionally display
19 NAI is licit in embedded contexts with or without an overt complementizer (Green ). The same facts are true for Appalachian English. Green reasons that the complementizer occupies a different part of the CP layer, either Finiteness or Force. It is unclear to me how this would work in Germanic CP-V languages where embedded V is not generally attested. In a highly articulated CP structure, what blocks embedded V? 20 The loss of this specifier may be consistent with Van Gelderen’s economy principles. 21 A reviewer observes that other focus-oriented XPs in present-day varieties of English also trigger V: negative inversion (Never have I seen . . . ); degree inversion (So much did it rain . . . ); exclamative inversion (Boy am I mad). 22 Thanks to a reviewer for this suggestion.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
V in Topic contexts, like other Germanic languages but unlike varieties of presentday English. Van Gelderen (: ff.) argues that a so-called split CP (such as the articulated CP of Rizzi and later work) is triggered via the economy principle of Late Merge.23 Under this approach, the grammaticalization of complementizers (for, till, that) involves a syntactic change whereby the elements are merged higher up in the left periphery, CP. Late Merge within the CP domain leads to the incorporation of topics in the left periphery, via a Topic projection. According to Van Gelderen, changes in the syntactic character of the complementizers (i.e. Late Merge) between Old English and Middle English led to the development of topics in the later variety. The proposal about Late Merge as an economy principle driving language change would suggest that the Older Scots examples involving V topics and transitive expletive pronouns (there, they) are in a lower CP projection than what I’ve proposed for Appalachian English and African American English transitive expletives and NAI. In other words, in these present-day varieties of English, the topics are merged even higher up than in the older varieties. But in that case the relative order of Topic and Focus adopted in () would not capture the path of change, since I would be led to assume Focus higher than Topic. Based on cross-linguistic evidence of multiple Topics but not Focus, Rizzi (p. ) in fact proposes recursion of Topic positions, above and below Focus, and this would resolve the issue for the present-day varieties of English. Van Gelderen’s hypotheses about language change prompts a related idea to consider. Could it be that general V was lost in varieties of English not only because of a breakdown in the verbal agreement system, but also because of a change in the features encoded in the CP layer of the grammar? Such an idea would be consistent with the proposal distinguishing movement into the CP layers of Older Scots, less restrictive, vs those of present-day varieties of English, more restrictive. It is also in the spirit of recent work by Biberauer and Roberts (). They propose that the class of features that trigger V-movement to C in English have changed, specifically that over time the triggers have become more specific until eventually they are no longer triggers.
. Conclusion In this chapter I have examined word order variation among varieties of English, predominantly Older Scots and its descendant, Appalachian English. The two major sentence types studied, transitive expletives and V, involve deviations from the canonical SVO order. I have suggested that the two sentence types are related and also that the patterns found across the varieties share some but not all features. Older Scots displays robust evidence of a CP-V grammar, patterning with other northern varieties of Middle English as well as Mainland Scandinavian, Dutch, and German. It also displays transitive expletives, but their distribution is more restricted
23
Van Gelderen’s work focuses mostly on embedded clauses.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
An effect of residual T-to-C movement in varieties of English
than in a language like Icelandic. I have tried to equate V and transitive expletives in Older Scots. For both cases, I have proposed that the finite verb raises over the thematic subject to C. Adopting Rizzi’s () articulated CP structure, I have refined the analysis and suggested that either a topicalized constituent or an expletive Topic occupies the specifier of TopicP. In applying these ideas to Appalachian English, we immediately observe the striking similarities to Older Scots, but also the differences. Although there is no evidence of a V grammar of the sort we saw for Older Scots, in these respects Appalachian English more closely resembles Older Scots than it does standard American English. I have attempted to connect transitive expletives in Older Scots and Appalachian English, taking CP to be involved in both. Despite a lack of evidence for a general V grammar in Appalachian English, I have explored the idea that negative auxiliary inversion, which is arguably related to transitive expletives, involves movement to the CP domain, following Green (). Instead of Topic, however, the feature involved is Focus. I have suggested that perhaps the loss of generalized V is tied in part to a shift in the inventory of triggering features. In some varieties of English such as Appalachian English and African American English, Topic triggers may have given way to Focus triggers. Building on the earlier work of Pintzuk (), Kroch and Taylor () and Kroch et al. (), I have distinguished V cross-linguistically in closely related varieties of English: Older Scots, Appalachian English, and African American English. The insight of structural differences in types of V, IP vs CP, is carried over here as TopicP vs FocusP. The current work also distinguishes pragmatic correlates of the structural differences, topic-oriented vs focus-oriented material. The area of overlap among V phenomena in these varieties of English is FocusP, which encodes the V associated with wh-elements in all three varieties. So Older Scots distinguishes between Topic, for regular V and transitive expletives, and Focus, for wh-elements. V phenomena in the present-day varieties, Appalachian English and African American English, have settled into the higher region of the articulated CP, Focus, a development that seems consistent with Van Gelderen’s () economy principles. It is in this sense that V in these languages is residual and more restricted. Many questions remain. For Older Scots, I would like to learn more about the loss of V and the development of transitive expletives, and whether their development follows what has been described for Late Middle English and Early Modern English. For Appalachian English, I would like to learn more about the development of negative auxiliary inversion, which appears to be more robust than transitive expletives. I am also curious about the status of transitive expletives and V in present-day Scots. Are any of these properties still found in the contemporary variety? I plan to find out.
Acknowledgements This work develops ideas that grew out of collaborative work with Raffaella Zanuttini, whose suggestions and guidance have been invaluable. The work has been presented at the Diachronic Generative Syntax conference held in Lisbon, Portugal in July .
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Judy B. Bernstein
I thank the audience members for their helpful and constructive feedback. I also benefited from feedback I received at the SYNC conference held at the CUNY Graduate Center. I thank Michael Montgomery for sharing his knowledge and ideas, and for his encouragement. I am indebted to Caroline Macafee for her many (gentle) corrections to my flawed Older Scots glosses. I thank Goldie Ann McQuaid for help in gathering Older Scots examples, and I am grateful to Goldie Ann McQuaid, Tiffany Williams, and Greg Johnson for their judgements on Appalachian English sentences. Two anonymous reviewers provided extremely valuable feedback, for which I’m grateful. The initial work on Appalachian English was funded by a – American Council of Learned Societies fellowship and a – National Science Foundation grant (BCS ).
Primary Sources DSL—Dictionary of the Scots Language (a publication of Scottish Language Dictionaries). . HOCS—The Helsinki Corpus of Older Scots, compiled by Anneli Meurman-Solin, Department of Modern Languages, University of Helsinki, Finland. Available as: Rissanen, Marti, and Anneli Meurman-Solin (). The Helsinki Corpus of Older Scots. [Electronic resource: –]. Oxford: University of Oxford Text Archive. Shearer, Kathy (–). Dante Oral History Project. Recorded oral history interviews collected in Dante, VA, housed in the Archives of Appalachia, East Tennessee State University, Johnson City.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
8 Word order and information structure in the Würzburg Glosses CARA M. DIGIROLAMO
. Introduction In John O’Donovan’s () Grammar of the Irish Language, sentences that contain arguments in pre-verbal position are described thus, In poetry, or poetical prose, the natural order of sentences is sometimes inverted, and the nominative case placed before the verb . . . When the noun is placed before the verb, it does not immediately connect with the verb, but rather stands in an absolute state . . . Such construction, though unquestionably faulty, is often adopted by the best Irish writers for the sake of emphasis.
Although this is in regard to Modern Irish, Osborn Bergin () asserts that the structure where the logical subject is at the head of the sentence is ‘equally common at all periods’ (Bergin : ). In contrast to this, certain modern scholars have claimed that Old Irish does not front for emphasis. ‘Old Irish does not allow fronting for emphasis or in wh-questions . . . The only way to emphasise a constituent is to use a cleft sentence’ (Newton : ). In view of these conflicting opinions, when we see a seemingly argument-initial sentence in Old Irish, what can we say it is doing? Old Irish is unquestionably a VSO language. A vast majority of the sentences in the available corpus are clearly VSO, as in example (). () V S O IO LOC dobert a máthair dí báe déc dó assint síd give.PAST.S his mother.NOM two cow ten to.SG.M from.the fairymound ‘His mother gave him twelve cows from the fairymound.’ (Taín Bó Fraích, ) But as we see in (), not every early Irish matrix clause is strictly verb-initial. () sech ni ro-chim fon ainim a tir doroacht while NEG PERF-see.SG under.the blemish the land to.PERF.come.SG do imchaisin, a tír oa tuchadh ni róás in fect sa to behold and the land from-REL go-PERF.SG NEG FUT.SG now ‘With this blemish I cannot behold the land I have come to, and the land I have left, I cannot return to it now.’ (Tochmarc Étaíne s.) Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Cara M. DiGirolamo . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
Here, the speaker is emphasizing the parallelism between a tir doroacht ‘the land I have come to,’ and a tír oa tuchadh ‘the land I have left’ by pronouncing the second element before the verb and negation (ni róás). It is clear that () is a fronting structure and not a cleft with a missing copula because the non-relativized negation is used. This and other examples suggest that there is indeed an available pre-verbal position that can serve to emphasize the fronted element. Sentences containing pre-verbal arguments are infrequent—appearing most often in dialogue, poetry, and explanatory glosses. They seem to appear unpredictably, and in places where the rules of formal writing are relaxed. Yet examples appear robustly enough that a significant number of other scholars have studied them. Proinsias MacCana (), in his effort to locate the Middle Welsh noun-initial ‘Abnormal Sentence’ historically, compared noun-initial sentences in Welsh and early Irish, and analysed Irish noun-initial sentences as stylistically or functionally marked variants of the norm. This, the nominativus pendens analysis, acknowledges that these structures must have both been synchronically productive and conveyed some sort of semantic effect. But MacCana labels this effect ‘rhetorical,’ and does not identify exactly what it was. In addition to the nominativus pendens, where the fronted element is in nominative case, regardless of its role in the sentence and is frequently resumed by a pronominal or a conjugated preposition, the Old Irish corpus occasionally shows evidence of pre-verbal non-nominative DP/NPs, prepositional phrases, and adverbs. Fronted Accusative: () cluiche n-aímin inmeldach, agtait fir is mná game ACC-pleasant delightful, play men and women ‘a pleasant and delightful game, men and women play’ (Immram Brain ) Fronted PP: () i=naimsir mordochei ro=comallad techt doib huili fo=recht nimdibi in=time Mordecai PERF=fulfil going to.PL all under=law circumcision ‘in the time of Mordecai it was fulfilled that they all went under the law of circumcision.’ (Ml fd) Adverb: () amae ni dúib sed omnibus gentibus. verily NEG to.SG but all.DAT men.DAT ‘Verily it is not to you but to all men.’ (Wzb fa) Stifter () describes two structures in Old Irish which may result in these sentence types. The first is a cleft with a dropped copula, identifiable by lack of resumption and relativization on the verb. The second is left dislocation, a sentence with a pre-verbal NP/DP/PP where the verb is not relativized and if the fronted element is an argument, it is resumed. Both types of structures are found in the corpus, as are ambiguous sentences where there is a clear lack of relativization on the verb, but no resumption, or where there is no way to tell if the verb relativized.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
As clefts cross-linguistically tend to be an indicator of focus, we would also expect clefts with dropped copulas to express contrastive/identificational focus. But the purpose of the second type of structure, nominativus pendens, and similar noun or PP initial sentences where the verb is not relativized, has only been described as ‘rhetorical’. The use of the word ‘rhetorical’ suggests that these structures should appear in poetic and high literary texts. But when all poetic texts are removed from consideration, noun and PP initial sentences still appear. And, in fact, the more ‘speech-like’ (informal and unedited) the text is, the more common these constructions become. If a construction is synchronically productive and common in speech, it is likely to have a clear and consistent function. And structures that involve the fronting of arguments and appear in speech-style contexts are found in many, if not all, natural languages. In this chapter, I argue that Old Irish fronting has a specific information structural purpose. Krifka () defines information structure (IS) as a method of information packaging that responds to the communicative needs of interlocutors, organizing the information in terms of salience, topics, and relevant alternatives. Fronting for discourse purposes is common enough to encourage some linguists to assert the universality of topic and focus projections, TopP and FocP, and argue for a crosslinguistic analysis of both structure and function. In addition, structures marked for discourse are more frequent in speech than in carefully edited written language (Birner ). Thus, it seems plausible that these Old Irish constructions are discourse structures that actively affect information structure. Using examples gleaned from a close examination of the Würzburg Glosses and the Milan Glosses, and examples from other Old Irish texts, I motivate a synchronic analysis of argument-initial sentences in Old Irish. In particular, I describe two uses of fronting in Old Irish: aboutness topic and contrastive topic, and discuss a possible syntactic model for the IS-centric CP layer. I then discuss how the uses of argument fronting in Old Irish compare to the ones available in Modern Irish.
. The Bergin kerfuffle The earliest and most frequent discussions of non-canonical word order in Old Irish have all centred on two peculiar constructions: Bergin’s Construction and Tmesis. Bergin (), though aware of the nominativus pendens-style construction, where the argument precedes the whole verbal complex and does not affect the stress or morphology on the verb, focused on a different—now eponymous—construction. This construction also has the verb following an argument, but a simple verb has the conjunct ending and a compound verb is prototonic, indicating that the fronted argument is in some way incorporated into the verbal complex. () Bergin’s Construction Adv O V ceso femmuin mbolgaig mbung although seaweed swollen reap-SG ‘although I reap blistered seaweed’ (Bodl. Sanas Cormaic s.v. prull)
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
Tmesis is a similar construction where a full argument appears in second position in the sentence. Unlike Bergin’s Construction, only the word order is non-standard. The morphology remains as it would if the verbal complex were not split. () Tmesis PV S V ad- cruth caín -cichither to form fair see.SG.FUT.PASS ‘fair form will be seen.’ (Serglige Con Culainn ) Bergin analysed both these constructions as archaisms. As one would expect of archaisms, they appear mainly in poetic texts. In contrast, as we can see in () and (), the constructions under discussion in this chapter neither split the verbal complex in two, nor do they trigger alternative stress or morphology on the verb. *-epil () intí for-a-tuit=som immurgu at-bail side the.one on-PRO-fall=EMP however PV-die.SG he ‘He who goes to it however, he dies.’ (Wzb fd) *-immolngai () ind=foisitiu in=giun imfolngi in=duine slán the=confession.NOM the=mouth.DAT produce.SG the=man.ACC healthy ‘The confession in the mouth produces the sound man.’ (Wzb fd) Eska () and Doherty () have endeavoured to integrate Bergin’s Construction and Tmesis with other alternative word-order sentences such as nominativus pendens. Doherty suggests that pre-verbal argument sentences are evidence for a verb second stage in early Celtic. Eska protests that they are not. But both Doherty and Eska assume that these sentence types are literary and high-register uses of preserved clausal configurations from earlier periods of the language. Although it is indeed possible that Bergin’s Construction, Tmesis, and nominativus pendens are related structures, the differences in their distribution and their morphology suggest that they should not be conflated without further investigation. Bergin’s Construction and Tmesis may indeed be archaisms or poetically transformed syntax, but I assert that the nominativus pendens-style sentences are examples of synchronically productive structures, and must, therefore, have a systematic meaning predicted by the discourse context.
. Methodology The Würzburg Glosses are a collection of marginalia on the Epistles of St Paul dated to around AD. They are not annotated, so this chapter relies on a by-hand search of the corpus. It is supplemented by evidence from the Milan Glosses from the first half of the ninth century, which were made into a searchable corpus by Aaron Griffith (), and contrasted with data from the Parsed Old and Middle Irish
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
Corpus (POMIC) (Lash ), which contains both poetic and prose texts from a broader historical period. The glosses, together, make up about per cent of the Old Irish Corpus—as collected in the Thesaurus Palaeohibernicus. Of course, only a small portion of the glosses contain complete sentences. But out of these sentences approximately per cent are clefts and per cent have a pre-verbal phrasal constituent (this includes clefts with a dropped copula, sentences with fronted arguments and sentences with fronted non-arguments). Compared with POMIC, where clefts appear about a third as frequently and left dislocations make up only . per cent of the data (note that this does not include clefts with dropped copulas as they have been annotated as clefts with a null COP), it is clear that glosses are an ideal source to look at these types of sentences. However, although marginalia contain high proportions of non-standard sentence types and are less edited and more speech-like than more careful texts, they have their own problems as linguistic data. The Irish commentary and the Latin text freely interact, and the glosses frequently code-switch. The full sentence glosses, however, are primarily commentaries on the Latin, not translations of the Latin. Moran () points out that although all readers of the glosses would be skilful in Latin, Old Irish is used even when a Latin explanation would be easily understood, indicating that Irish was respected as a medium of scholarly discourse. According to Bisagni (), code-switching in the glosses follows the same constraints as verbal code-switching by bilinguals, and rather than indicating deficiency in either language, different languages are used to realize different communicative effects. Based on these principles, we expect that the Old Irish content interacts with the Latin on a high level and is made up of primarily grammatical utterances. Identifying which type of context triggers which IS structure in corpus data is never easy. Sentences with topicalization are often seen as marginal, but certain contexts create perceptible infelicities as we see in () (Ward and Birner ). ()
a. No, we’re out of bagels. A bran muffin I can give you. b. You ate six bagels. #A bran muffin I can give you.
But as we lack native speakers of Old Irish, we do not have access to this kind of negative data. We can, however, assume that alternative sentence types are not bandied about freely. In poetry, where many syntactic constraints are relaxed (Fitzgerald ), they may occur without a triggering context. But in non-poetic texts they should only occur in felicitous contexts. These contexts, however, are difficult to identify. Götze et al. (: ) proposed one test for identifying Topic, the Aboutness Test. ()
The Aboutness Test: An NP X is the Aboutness Topic of a sentence S containing X if: a. S would be a natural continuation to the announcement ‘Let me tell you something about X.’ b. S would be a good answer to the question ‘What about X?’ c. S could be naturally transformed into the sentence ‘Concerning X, Si’ where Si differs from S only insofar as X has been replaced by a suitable pronoun.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
However, when this test was applied on real corpus data, it provided inconsistent results (Cook and Bildhauer ). Even native speakers cannot easily identify these kinds of discourse relations in corpora. How, then, can we hope to identify discourse contexts in Old Irish? Ward and Birner’s () catalogue of discourse structures in English suggests a possible methodology. By carefully matching syntactic structures with their contexts of use, they convincingly argued for a variety of distinct IS sentence types. The first step is to identify the structures we are examining. To do that, I examine argument initial sentences for specific lexical items, syntactic patterns, and other markers. If there are multiple sentences that follow a certain pattern, I evaluate the context—both Latin and Irish—to see if a discourse function can reasonably explain its appearance. The three structures I examine are: intí-Sentences (fronted relative clauses); Invisible Clefts; Fronting + Cleft. In attempting to identify particular discourse relations, I examine the contexts for signs of the IS features proposed by Krifka (): Givenness, Focus and Topic. I introduce these concepts briefly here. Prince () argued that Givenness is important to take account of from a processing standpoint. Maintaining coherent anaphora allows the listeners to keep track of the major players in a story, and it also allow the speaker to use briefer referring expressions. By promoting given expressions to the beginning of the sentence, and shifting new information to the end, IS structures can allow the hearer to prepare to evaluate new information. Rooth () proposed the now standard analysis of Focus as bringing into salience a Hamblin-style set of alternatives. When ‘John’ is focused, the set containing all the people who are not John suddenly becomes relevant, but the proposition is stated to be true of John, and not necessarily true of anyone else. Topics represent the changing structure of an organized conversation. If we assume Roberts’ () Question Under Discussion (QUD), where the goal of a conversation is represented as a question, we can see topics as ‘file cards’ (Reinhart ) each focusing around a certain subquestion of the main conversational question. As utterances are made, the file cards record the relevant propositions. When topics shift, new file cards are formed.
. The data .. Intí-sentences intí-sentences are sentences where a full relative clause, headed by intí is found at the beginning fo the sentence followed by a non-relativized verb, and occasionally by a clefted sentence. In the following two glosses, we have three examples of fronted intí clauses. ()
isbéss didu indliacc berir ilbeim friss et inti dothuit foir oboing a chnámi intí foratuitsom immurgu at-bail side. (Wzb fd) ‘it is the manner, then, of the stone, that many blows are given to it, and he who falls upon it breaks his bones: he, however, on whom it falls, he perishes.’
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses ()
ardoecmalla inmertrech cuicce pecthu indlína dodaaidlea combi óinchorp pectho asmberar et asberar corp dondlúim máirsin innapecthach intí immurgu téte adochumsi doecmallaside foir indlúim máir inchoirpsin et inphecthasin bís forsin mertrich et nísnderigsi dano act atballat diblínaib (Wzb fd) ‘for the harlot gathers unto her the sins of those that visit her, so that there is one body of sin, which it is called, and “body” is said of that great mass of sinners. He, however, who goes to her, gathers on him the great mass of that body and that sin which is on the harlot and she doesn’t strip it off then, but they both die.’
Morphologically, intí is a determiner combined with indeclinable pronoun (in(t) + í). Though í is indeclinable, it is usually labelled as nominative, making it an example of nominativus pendens. It also tends to be resumed by the subject of the main clause or the subject of a cleft preceding the main clause. Old Irish scholars have claimed that the use of intí is often emphatic (DIL ). It occasionally appears alone, but is usually followed by relative clauses, both in fronted position and in normal position. It is reasonably common, appearing not only in the Würzburg and Milan Glosses, but also in POMIC (Lash ). When we examine the sentences more closely we find other items that appear frequently with intí. (.[selected]) . . . inti do-thuit foir oboing achnámi . . . . . . the.one falls-REL on.it smash.SG bones . . . ‘ . . . he who falls upon it breaks his bones’ (Wzb fd) (.[selected]) . . . intí for-a-tuit-som immurgu at-bail side . . . the.one on-PRO-fall-EMP however PV-die.SG he ‘ . . . he on whom it falls, however, he perishes’ (Wzb fd) (.[selected]) . . . intí immurgu téte adochum-si doecmalla side . . . the.one however go.REL to.her-EMP gather.SG he foir indlúim máir . . . on.him the.mass great . . . ‘ . . . He, however, who goes to her, he gathers on him the great mass . . . ’ (Wzb fd) immurgu ‘however’ appears in two examples. According to its definition, immurgu can be used both for emphasizing force and for strengthening (also implied opposition) (DIL ). In (.) and (.), side also appears. side, a ‘deictic clitic’ and weak pronoun, emphasizes the third person of pronouns, possessives, verbs and conjugated prepositions (Griffith ). Since side appears in instances where the fronted item is the subject and not when it is the object, it seems that side is the resumptive subject. In other examples where the intí clause is the object of prepositions (Ml c, bc), it is usually followed by a conjugated preposition in the following clause, and thus seems to be resumed in those contexts as well. When we examine the data in its context, we find that our examples appear in sustained Old Irish text about a single subject. (.) and (.) appear amid a
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
discussion of ‘the stone’ (in ()) and each present a statement having to do with people who encounter that stone, one where they trip over it, and one where it falls on them. (.) appears in a discussion of ‘harlots: problems with’ (in ()) but instead of following up on what happens to the harlot, (.) presents a shift in focus, indicating that we have moved on to a related but distinct topic: ‘the person who visits harlots.’ This slight-topic-shifting that we see in all of the examples strongly resembles Krifka’s () definition of contrastive topic. Krifka () describes a contrastive topic as consisting ‘of an aboutness topic that contains a focus, which is doing what focus always does, namely indicating an alternative. In this case, it indicates alternative aboutness topics’. As discussed above, we take a topic to be a file-card representing a subquestion of the main conversational question. We assume that a focus is a set of alternatives. Therefore, a contrastive topic picks out a set of subquestions, with one marked out as relating to the proposition. Because a contrastive topic structure picks out a set, it reminds the listener of the existence of the other cells, as in (), where a question is answered incrementally, pointing out and then addressing two subquestions separately. ()
‘Do you like your siblings?’ ‘My sister I like. My brother I don’t.’
(.) and (.) are very much like () in that they are answering different subquestions of the same question. They are also clearly contrasting with each other. (.), however, has nothing to directly contrast with. But the use of immurgu and the subtle contrast with the previous topic suggests that there is a focus inside the fronted element. Since (.), (.), and (.) all involve a topic shift and most seem to contain some sign of an internal focused element, contrastive topic appears to be a reasonable analysis for these structures. .. Invisible clefts Many languages, including Old Irish, use clefting or intonational emphasis to mark focus, rather than left dislocation which patterns with topics. Old (as well as Modern) Irish is able to drop the copula in default situations, and the relative marking can be less than obvious. Since distinguishing sentences with null copulas and no relative marking from movement into a focus position is quite challenging, it seems plausible that these sentences may have been interpreted by some speakers as fronting for focus rather than as clefts. However, Nakamura () points out for Modern Irish that even when the relative marking is overt, nominalized sentences are syntactically and semantically distinct from clefts, showing that they are not simply sentences with unpronounced default copulas. This may, in fact, be the case for Old Irish as well. Nakamura () identifies the information structural content of nominalized sentences as including answers for wh-questions, informative focus, and no-focus introductory sentences. These contexts are all infelicitous for full clefts. This would indicate that we should not find invisible clefts bearing cleft-like focus meaning. Nevertheless, when we look at the Old Irish data, some instances of focused fronting do seem to appear. (), for example, contains nammá ‘only’, a focussensitive operator, in the fronted element, making it appear to be focused.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
() in dúib nammá ro=pridchad amae ni dúib sed omnibus gentibus. Q to.SG only PERF=preach verily NEG to.SG but all.DAT men.DAT ‘Has to you alone there been preaching? Verily it is not to you but to all men.’ (Wzb fa) Fronting nammá (‘only’) can indicate a contrastive topic rather than focus. Nevertheless, in this particular example, duib namma ‘to you alone’ is not functioning as a topic. First, the fact that it is a prepositional phrase rather than a DP, and one that cannot be interpreted as temporal or scene-setting makes it very unlikely to be topical. Second, it is embedded inside a question, indicating that it is part of the content being interrogated and not background. () is, however, very similar to the Latin that it is explaining, which may have influenced its word order. (.[context]) Corinthians : Án á vobís sermo Dei processit? Aut in vos Q from you.ABL word.NOM God.GEN proceed.PERF.SG or Q you.ACC solos pervenit? alone.ACC reach.PERF.SG ‘Was it from you that the word of God proceeded? Or was it you only that it reached?’ If () is indeed a cleft, the third person copula after the question particle in is only represented by nasal mutation. ‘In ndúib’ is expected, however, the distinction ̃ between mutated /ɪnnuiβ/ and partially assimilated /ɪnduiβ/ is nearly inaudible. As expected for Old Irish, the orthography is not always perfectly consistent. We do not expect any relative marking on the verb since the fronted element is a prepositional phrase. So () is a reasonable example of an invisible cleft. () also seems to be a focus. () dogailse fo=d-ruar (Ml fb) grief PV=SG-cause ‘(it is) grief which has caused it’ The translation suggests that this is also an invisible cleft, but there is no sign of a copula, nor of relativization on the verb. This gloss is taken from a commentary on Ephesians :. (.[Context]) Ephesians : IRASCEMINI ET NOLITE PECCARE . . . quia ista quæ sentitis quae dicitis non a ratione sed á merore uenerunt . . . ‘Be ye angry, and sin not.’ ‘ . . . because it is this you realize, that the things you say came, not from reason, but from merit’ This gloss seems to be responding to a confusion in the commentary about merore. To clarify what is meant in this statement, the annotator points out that the intended meaning of merore is dogailse ‘grief ’ rather than ‘merit’. This does not seem to be a topic construction, and instead points out a possible mistake and corrects it, which is a function we would expect of focus.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
Due to the particular morphology of Old Irish—droppable copulas and relative marking that often has no overt realization—we would expect a focus fronting-like construction to be available to speakers, though it might be less available in formal texts. Clefting, extensively discussed in MacCoisdealbha (), Nakamura (), and Sulger (), seems to be the preferred Irish focus structure. However, it seems that Old Irish has some evidence of a fronting for focus construction as well. .. Fronting + cleft Another structure that appears in the glosses is one where a fronted element is resumed in a copular cleft. In the previous section we suggested that in Old Irish, focus is expressed through clefted structures. If this is indeed the use of a cleft, what is the meaning of a cleft with a fronted element in front of it? In () and () the fronted element is a complex reformulation of an element obliquely referred to in the Latin. () méit donindnagar fornni fochith is=sí méit in=sin amount give to.PL suffering COP=sg amount DET=that donindnagar in=díthnad ni tabir dia fornni didiu fochith nád give DET=consolation NEG give god to.PL then suffering NREL fochomolsam cid ind=fochith follongam dober díthnad darahéssi bear which DET =suffering endure give consolation second.after ‘the amount which is given to us of suffering, it is this much which is given of consolation. God does not put on us then suffering which we cannot endure. Which suffering we endure, he gives consolation after it.’ (Wzb fb) (.[context]) Corinthians : Quoniam sicut abundant passiones Christi in nobís, ita et per Christum abundat consulatio nostra. ‘For just as the sufferings of Christ are abundant for us, so also our consolation is abundant through Christ.’ In (.) the Latin states that ‘the sufferings of Christ are abundant for us’. In the Irish, the same referent is reformulated as ‘the suffering given to us’, and then inserted into a comparative construction: ‘the amount of suffering given is equal to the amount of consolation given’. () aní dodesta dichomalnad caesta crist domsa is=occa that absent fulfilment suffering christ to.SG.EMP.SG COP=at.SG attóo is=hed didiu desda di=sudiu dul martre tar=farcenn=si SUBS.SG COP=SG then absent from=this go martyrdom across=sake=EMP.PL ‘that which was wanting unto me of fulfilment of Christ’s Passion, it is at that I am. This, then, is what was wanting: to go to martyrdom for your sake.’ (Wzb fd) (.[context]) Colossians : Qui nunc gaudeo in passionibus pro vobis, et adimpleo ea, que desunt passionum Christi, in carne mea pro corpore eius, quod est Ecclesia:
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
‘I now rejoice in my sufferings for you, and do fill up the things lacking of the tribulations of the Christ in my flesh for his body, which is the Church.’ In () another complex and oblique referent is followed up with a cleft. This cleft pronominalizes the referent as the object of a preposition. Here as well, the complexity of the concept may be triggering topic left dislocation to aid in the processing of the sentence. As the referent is not the direct subject of the main clause, the clefting may also serve to clarify the reference relationship between the fronted element and the pronominal. One of the uses of topicalization is to present a backgrounded or unclear referent as an entity that will subsequently be discussed. As the fronted element in both () and () is indeed the topic of not only the immediately following proposition, but also the proposition contained in the sentence after that, it is clearly indicating what the next portion of the discussion is to be ‘about.’ () and () are like the contrastive topics we saw in the intí-clauses in that they are heavy and fronted, but unlike them in that they do not seem to be creating a subtopic that contrasts with a set of alternatives. Instead, the fronted elements are simply what the following propositions are about. Therefore, it must be the case that () and () are examples of aboutness topics. But other examples of fronting + cleft, such as (), do appear to answer a subquestion like a contrastive topic (‘some . . . , others . . . ’). () ni
taibre grád for=nech causa a=pectha l. a=chaingníma arbitt give rank on=anyone causa SG=sin or SG=good.deed for-SUBS alaili and ro=finnatar apecthe ro=síu docói grád forru. alaili others there PERF=known SG-sin PERF=this went rank on-PL others is=iarum ro=finnatar berir dano frilaa brátha COP=then PERF =known carry also toward.day Doomsday ‘Thou shouldst not confer orders on any one causa of his sin or of his good deed: for there are some whose sins are known before their ordination, of others it is afterwards that [their sins] are known. He will be judged, therefore, on Doomsday.’ (Wzb fa) NEG
And unlike () and (), the cleft does not contain a pro-form to resume the fronted element. (), like the intí-sentences, appears to contain a contrastive topic. (), an instance of fronting an intí-clause above a cleft, also resembles a contrastive topic, in which the people who the Lord desires to save are contrasted with the ones that he does not desire to save. ()
Ceist cid natat sláin indhuli mátchobra quia omnia quaecumque uoluit dominus fecit nianse hóre nád comeicnigther nech fritoil l. is pars pro toto ar=ni=fil cencel nabelre is=in=biuth or COP pars pro all for=NEG=COP race NEG.language COP=DET=world dinadrícthe nech l. indíi adchobra=som dohícc ithé ronícc from.which anyone or DET.to.SG desire=EMP.SG to.atone COP.SG PERF.atone tantum (Wzb fb) tantum
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo ‘Question, why are not all saved if He desires it because everything that the Lord wishes he brings about, not hard (to say): because no one is constrained against (his) will.’ ‘Or it is a part for all, for there is neither race nor nation in the world of which some one shall not have been saved. Or those whom He desires to save are they whom He has saved so far.’
From this data it seems clear that both contrastive and aboutness topics can appear in front of clefts. When these topics also serve as an argument of the clefted sentence, the question of whether or not they truly resume remains open. () Is
hé in peccad rogéni anuile comaccobor SG DET sin PERF.make every concupiscence ‘It is sin which has wrought every concupiscence’ (Wzb c, Old Irish)
COP
() shows the basic structure for an argument cleft, where both a pronoun and the full argument appear above the main relativized verb. When the argument is fronted over the cleft, there is usually a pronoun immediately after the copula, but that pronoun would exist regardless of where the argument appeared. Unlike fronted intí clauses and other structures where resumption is necessary and the main sentence is syntactically grammatical without the fronted element, clefts with a fronted argument may be the result of a movement operation rather than a topic merged in the left periphery.
. Summary and syntax In the previous sections we discussed three Old Irish fronting structures. With fronted intí relative clauses we saw examples that looked like contrastive topics, creating an opposition between discourse referents without affecting the focus value of the sentence. When we looked for a fronted focus, what we found tended to look like clefts. The third structure we examined combines both a left dislocated element and a cleft. These, it seems, can be either contrastive or aboutness topics. Syntactically, if we assume that the main verb marks the top edge of IP, we would expect all of these fronted elements to appear in the left periphery. As they do not interact with the prototonic/deuterotonic stress alternations (as seen in () and ()), they are external to the prosodic phrase containing the verbal complex as well. The standard model for left peripheral syntax is usually Rizzi’s () cartography, which contains a projection for focus and two recursive topic positions surrounding it. ()
[ForceP [Top* [FocusP [Top* [FinP [IP]]]]]]
But we do not have any evidence that topics are recursive in Old Irish, and this model does not distinguish between the types of topics. Frascarelli and Hinterhölzl () propose a more detailed left peripheral structure, as in (). ()
[ShiftP +aboutness [ContrastiveP [FocusP [FamiliarP* [IP]]]]]
They point out that ShiftP is above wh, but they do not indicate where the Force/Fin heads in Rizzi end up. If we were to map a Rizzi-style CP to this detailed topic
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
structure, the most likely would be the one in Rizzi (), which also contains three topic projections. ()
Rizzi (): [ForceP [TopP* [IntP [TopP* [FocusP [TopP* [FinitenessP [IP]]]]]]]]
()
Combined Version: [ForceP [ShiftP [IntP [ContrastiveP [FocusP [FamiliarP* [FinitenessP [IP]]]]]]]]
Here IntP appears below ForceP and the topic shifting position, but above FocusP and the two other topic positions. ShiftP corresponds to aboutness topic in our analysis, and on this model we would predict that these appear above Irish sentence initial particles such as in, the question marker, and negation. Though we do not have any evidence for fronting to a position higher than the question particle, we do get fronting over negation, as in (). Of course, NEG is very likely to be nearly immediately above FinP, and therefore lower than all of the IS projections. () intí bīas ann, nī bīa aidche troiscthe. the.one reaches there NEG SUBS night fasting ‘[For] the person who shall be staying therein, there shall be no night of fasting.’ (lh.) .. Clefts and focus In the section on Invisible Clefts, I discussed three structures that seem to be related to clefts. The first, of course, is the cleft itself, which expresses focus. The second is the nominalized sentence, which, in Modern Irish at least, is distinct from the cleft in both syntax and semantics, has no copula but does have relativization on the verb. The third is the Invisible Cleft, which has no overt copula nor does it have overt relativization on the verb, but like clefts, has focus meaning. This third structure is the most likely to be focus fronting. In example (), the fronted element appears lower than the question particle, which is consistent with our combined Rizzi/Frascarelli and Hinterhölzl analysis. ([Repeated]) in dúib nammá ropridchad Q to.SG only PERF.preach ‘Has to you alone there been preaching?’ (Wzb fa) The question still remains, however, whether the position the focused element is in is the same in clefts, nominalized sentences, and invisible clefts. The internal structure of clefts alone in Old Irish is inherently complex. Stenson () argues that the basic structure of cleft sentences is identical to copular identification sentences, where the second NP is a headless relative clause. ()
Copula [NP]focus [NP – relative clause]subject
Similarly to other Celtic languages, the copula + predicate has likely moved up into FinP, leaving the subject in SpecT (Roberts ). But Stenson () seems to
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
indicate that in these types of sentences it does not stop in FinP, but moves up to FocP. This does not account for the difference between clefted arguments and clefted non-arguments, where arguments are doubled by a pronoun immediately after the verb. ([Repeated]) Is
hé in peccad rogéni anuile comaccobor SG DET sin PERF.make every concupiscence ‘It is sin which has wrought every concupiscence’ (Wzb c)
COP
() Is
i=ndia dorecatar ní i=ntiu fesin in=God hope NEG in=PL self ‘it is in God they hoped; it is not in themselves.’ (Ml fd)
COP
For non-arguments, Stifter () points out that there is never any sign of relativization on the verb, suggesting that these may not contain relative constructions at all. Instead the copula and focused element are directly merged in FocP, with the copula serving as the head of the focus projection or a focus introducing discourse particle. Nominalized sentences, which lack both the pronoun and the copula, but have overt relativization, may also make use of the left periphery. As one of the most common uses of nominalized sentences is to answer wh-questions, and Stenson (), Sulger (), and Nakamura () all make a point of how much nominalized sentences (or clefts with dropped copulas, depending on the terminology) resemble wh-questions. As the projection usually used by wh-questions in the linguistics literature tends to be focus, it seems that nominalized sentences in Modern Irish use FocP, and this may also be the case for Old Irish. If the relativizing particle is in FinP, it may be realized as relative even in a monoclausal structure if FocP is filled. True argument clefts seem to be biclausal. And yet, if we follow Stenson (), they are still making use of the focus projection. Stenson (), however, puts the argument DP and the doubling pronoun into a single constituent. There does not seem to be a reason for this. The doubling pronoun may in fact be the predicate of the copula and in FocP, while the focused constituent remains in subject position. One piece of support for this proposal is that when there is a subject fronted over a cleft, the doubling pronoun remains in situ. This suggests that the doubling pronoun and the subject are not a constituent. However, it also may simply be serving as a resumptive pronoun for the left dislocated element. The ordering in () is consistent with clefts being in FocP. Both aboutness (shifting) and contrastive topics can appear above clefts. Focused elements appear to surface lower than Int (in ‘Q’, as in ). We do not have available data to show where contrastive and aboutness topics surface in relationship to Int or Force, and, in fact, splitting them into two projections in Old Irish may be overcomplicating the situation. Vermeulen () suggests that surface-ambiguous IS structures in Japanese can be identified by inferring the correct interpretation from contextual cues. And as aboutness topics and contrastive topics are minimally distinct this may be a valid interpretation of Old Irish as well. The contrastive interpretation of
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
the contrastive topic could easily be inferred from prosodic information, lexical cues, or the expectations resulting from the conversational goals. Including a difference in the syntactic (not semantic) features of aboutness and contrastive topics entails a difference in syntactic distribution which, in Old Irish at least, we do not have evidence for. .. Final ordering In addition to aboutness and contrastive topics, the Frascarelli and Hinterhölzl () model also contains a projection for familiar topics. This is a problem for our analysis of Old Irish. In our investigation of the data, there was nothing that stood out clearly as a familiar topic, and this projection may not be active in Old Irish, as it is not a topic-prominent language. Instead, the lowest topic position may be the one used for adverbial discourse markers such as immurgu ‘however’ or trá ‘indeed’. As these elements are usually found between the fronted phrase and the verbal complex, this is indeed where we would expect them. () cosmulius tra dombeir-som híc aris bésad leusom in fid simile then PRT.SG-give.EMP.SG here for.COP like on.PL DET wood do-thóbu et fid aile do-esnid and insert and wood other insertion there ‘A simile then he gives here. For they have a custom to cut the wood and insert another wood therein.’ (Wzb fb) Or, perhaps, it may host the no-focus fronted constituent in that type of nominalized sentence. From this data, I will propose a left periphery for Old Irish. ()
[ForceP [AboutnessP [IntP [ContrastiveP [FocusP [Adv [FinP]]]]]]]
. Diachronic perspectives Proinsias MacCana, in his article on the ‘abnormal’ word order in Middle Welsh, argued that in Irish there ‘must have always been a significant incidence of nouninitial types used as stylistically or functionally marked variants of the norm (: )’. Compared to Modern Welsh, Modern Irish still contains noun-initial structures at a reasonably high frequency. Looking at the extensive examples MacCana provides, mostly from the literature of the s, we find fronted elements that seem like both aboutness and contrastive topics. One example, (), is embedded inside a paragraph discussing the year of the famine. () Fir a bhí mór láidir lá den tsaol bhí siad ag tuitim man PRT be big strong day from time be they at fall as a seasamh leis an ocras from PRT standing with DET hunger ‘Men who had once been big and strong were falling off their feet with hunger.’ (F. Mhac Cumhaill, Na Rosa go Bráthach : )
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
The previous sentence discussed how the provisions they had made to face the famine were running out and they were reaching their limits. In () the author is presenting an example of the extent to which they were out of time. The fronting is being used to point out that even the people who were expected to succumb to the famine last are suffering. It is not fronting for Focus, however, because it does not present ‘men who had once been big and strong’ in contrast to all the other people suffering from the famine. Instead it seems more like a contrastive topic, indicating that in addition to the weak who are dying, the big strong men are also suffering. () appears in a context where a character named Conall is presenting a stone to a crowd of people. He looks at each person before he makes the statement that appears in (). () ‘An fear a thógfas an leac’, arsa seisean, ‘biodh a bhfuil the man PRT lifts DET stone said he SUBS.IMP PRT SUBS.DEP faoithe aige’ under.SG with.SG ‘The man who lifts the stone, he said, let (him) have whatever is under it.’ (S. Mac Grianna, An Grá agus an Ghruaim : ) This seems like focus, since ‘The man who lifts the stone’ is the one who will receive the reward to the exclusion of all of the others. It is unlike contrastive topic because it is not in opposition to any other topic of discourse. And yet, as it is in an imperative construction, it seems also interpretable as a vocative topic. A clearer instance of fronting for focus is (), in which the topic is the previously mentioned ‘deed’, but the ‘substance’, or lack thereof, of the deed is being emphasized. () ‘Dar mo chlaíomh [ . . . ]’, arsa seisean, ‘[ . . . ] go mbéadh ar by.god SG sword said SG.M PRT SUBS on ngníomh ní b’éfeachtaí!’ do thing be’substance ‘Éifeacht níl ins an ghníomh sin’, arsa sise substance NEG.SUBS in.SG DET doing DEM said SG.F ‘(I swear) by my sword that our deed would be more effective, said he.’ ‘There is no substance in that deed, said she.’ (S. Mac Grianna, An Grá agus an Ghruaim : ) The ‘deed with substance’ ngníomh ní b’éfeachtaí, is the alternative to the ‘deed that has no substance in it’ Éifeacht níl ins an ghníomh, as we expect in a Focus construction. Linguists such as Kiss (), claim that Modern Irish is not a topic prominent language and does not have topic fronting. The data collected in MacCana (), however, shows that there is definitely an available position for fronted elements. And these structures are not only in the Irish of the s. The European Union Conradh Amstardam also includes sentences containing topicalized phrases. In this case the phrase seems to be an aboutness topic.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Word order and information structure in the Würzburg Glosses
() An rialú arna thabhairt ag an gCúirt Bhreithiúnais [ . . . ], ní bheidh DET ruling as to at DET court justice NEG will feidhm aige maidir le breithiúnais de chuid chúirteanna nó apply at.SG regards with judgment of part courts or bhinsí an mBallstát tribunals DET states ‘The ruling given by the Court of Justice [ . . . ] shall not apply to judgments of courts or tribunals of the Member States . . . ’ Regardless of the other changes Irish has undergone in the past millennium, it seems to have maintained not only VSO word order, but also an available topic position above the initial verb. Unlike in Welsh, where the basic word order transitioned from verb initial to V and then back to verb initial (Willis ), possibly due to encouragement from topics, the availability of this structure has not influenced the basic syntax. Clefting, in Modern Irish, has retained its focus function. Relative marking, however, has become more consistently overt. () Modern Irish Is í an bhean a bhí tinn COP SG.F the woman REL be.PAST sick ‘It’s the woman who was sick.’ In Old Irish the relative marker for this type of sentence would simply be lenition, which is invisible in many phonological contexts. In addition to clefting, however, the nominalized sentence as well as an unmarked focus fronting structure are available in Modern Irish. Although the traditional analysis of Old Irish does not include the nominalized sentence and simply suggests that we have copula drop, the contexts where we find nominalized sentences in Modern Irish seem similar to where we find instances of copula drop in Old Irish. However, although Old Irish has copula drop and frequently invisible relativization, it is still unclear whether or not there is a true fronted focus position. Modern Irish, however, as we saw in () appears to have a real non-clefted focus. At a glance, MacCana’s collection of examples contains about half left-dislocated focuses and half left-dislocated topics. It seems plausible to suppose that our descriptive category of invisible clefts may have been the impetus to develop/proliferate sentences with fronted focus. As a/a(r) became the primary relative markers, sentences with missing copulas and less overt relativization markers were perceived as simple fronting for focus. Thus in Modern Irish, sentences with fronted focus, clefts with overt copulas, and nominalized sentences are all available. In terms of information structure, Irish has been rather conservative, maintaining the general function and position of its topic and focus structures. The main change is that simple focus fronting has become a more productive structure over time.
. Conclusions In this chapter we have investigated the use of noun-initial sentences in Old Irish. I have argued that noun-initial sentences are not simply used for stylistic purposes,
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Cara M. DiGirolamo
but are, in fact, examples of fronting for information structure. In the Würzburg and Milan Glosses, we find these structures quite frequently, and even in POMIC, they are reasonably well attested. Looking specifically for three types of structures, intí-Sentences (fronted relative clauses), Invisible Clefts, and Fronting + Cleft, we found that Old Irish appears to have both aboutness and contrastive topic, and mostly uses clefting for focus, though the cleft morphology is not always visible. Modern Irish, however, has seemingly developed a productive fronted focus position. Both Old and Modern Irish also show signs of the nominalized sentence, which is relativized like a cleft, but has no copula. The proposed syntactic structure is still tentative, and further investigation is required to verify the ordering of the constituents, but it seems clear that we can use a slightly adapted version of the Frascarelli and Hinterhölzl () model to describe the Old Irish left periphery. Although this chapter was mainly concerned with non-relativized noun-initial sentences, it is clear that more work needs to be done on clefting in Irish. In particular, building on Nakamura’s () descriptive work, it would be a worthwhile next step to develop a formal syntactic analysis which accounts for the syntactic and semantic differences between clefts, nominalized sentences, and focus fronting, and incorporate it with this outline of the Irish left periphery.
Acknowledgements This chapter is an expansion and development of the poster ‘Discourse Constraints in Old Irish’ presented at the th Diachronic Generative Syntax conference in Lisbon, Portugal. I would like to thank Sarah Courtney for her assistance in gathering data, critiquing and theorizing, Michael Weiss for his advice and Latin translations, Elliott Lash, Aaron Griffith, John Whitman, and the anonymous reviewers for their comments and corrections, and members of the audience for their questions and discussion.
Sources Griffith, Aaron (b) Dictionary of the Old Irish Glosses in the Milan MS Ambr. C inf. . Lash, Elliott (). POMIC: The Parsed Old and Middle Irish Corpus, Version ., Online: Dublin Institute for Advanced Studies, School of Celtic Studies, –present. .
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Part III Verb movement into the left peripheries
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
9 Subject inversion in transitive sentences from Classical to Modern European Portuguese A corpus-based study C H A R L O T T E G A L V E S A N D AL B A G I B R A I L
. Introduction Several studies of the history of Portuguese argue that Old Portuguese (henceforth OP) and Classical Portuguese (henceforth ClP) had a V-like syntax (cf. Ribeiro for OP; Moraes , Galves , , Paixão de Sousa , Galves et al. , Galves and Paixão de Sousa , , Gibrail , and Antonelli for ClP). Other researchers have criticized this analysis on the basis of the fact that both OP and ClP display properties that are not compatible with a V syntax, namely high frequencies of V and V sentences (cf. Kaiser , Fieis , and Rinke for OP; Eide for ClP). In this chapter, we focus on ClP, and its change to Modern European Portuguese (henceforth EP), bringing to the debate new data concerning transitive sentences. The data are drawn from sixteen parsed texts available in the Tycho Brahe Parsed Corpus of Historical Portuguese, written by Portuguese authors born between and . The size of the parsed corpus is , sentences, which corresponds to approximately , words. In order to compare ClP with EP, we include in the analysis descriptions of word order in EP (Ambar ; Costa ; Duarte ). Based on the difference between the syntax and interpretation of the different possible word order in transitive sentences (SVO, VSO, VOS) in ClP and EP, we argue that ClP was a V-like language, i.e. a language in which the verb occupied a high position in the clause, which made a high position for post-verbal subjects available as well. The chapter is organized as follows. In section ., we present a quantitative analysis of the evolution of the position of subjects in transitive sentences in declarative main clauses which evidence a drastic change in texts written by authors Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Charlotte Galves and Alba Gibrail . First published by Oxford University
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
born from onwards. Section . explores the qualitative correlate of this change, showing that the interpretation of post-verbal subjects in ClP is fairly different from what is described for the modern language. In section ., we bring further evidence, both from the synchrony of ClP and from the diachronic changes observed, in favour of the hypothesis of a grammatical change from ClP to EP, against Eide’s () claim that the change affected not syntax but information structure. The last section briefly summarizes the main findings of the chapter.
. The quantitative change from ClP to EP Several recent works have shown that a change affecting the position of subjects took place in texts written by Portuguese authors born from onwards (cf. Paixão de Sousa , Eide , Gibrail , Antonelli , Galves and Paixão de Sousa , etc.). Figure ., from Galves and Paixão de Sousa () shows that the frequency of post-verbal subjects (black triangles) in matrix declarative sentences with lexical subjects was higher1 in texts written by authors born in the sixteenth and seventeenth centuries, than in texts by authors born in the two subsequent centuries. 1.00
0.77
0.75
0.70 0.63
0.61 0.57 0.50
0.46 0.36
0.40 0.34 0.29
0.27 0.23
0.25
0.21
0.17
0.00 1450
1500
1550
1600
1650
1700
1750
1800
0.32 0.21
1850
1900
Post-verbal/Pre-verbal subjects (i.e., (x)vs/(x)vs+sv)
F . The relative VS/SV frequency in matrix declarative contexts, by author (Galves and Paixão de Sousa ) 1 The computation here only takes into consideration the sentences in which the subject is lexically realized. When null subjects are included in the analysis, the average frequency of VS sentences is per cent and per cent in the sixteenth and seventeenth centuries, respectively, and per cent and per cent in the eighteenth and nineteenth centuries, respectively. It is worth noting that there is no substantial change in European Portuguese in the frequency of null subjects over time (cf. Galves and Paixão de Sousa ).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese 120% 100%
98%
94%
80% 60% 40%
78% 64%
65%
36%
48% 45% 47% 44% 39% 34% 35%
20% 0%
2% 0%
1%
14%
6% 4%
11% 1%
93%
92%
89%
3%
6% 2%
6%
47% 33% 20%
8%
89%
20% 2%
7%
16% 5% 1%
1%
G an da vo 1 Pi 502 nt o 1 Co 510 ut o 15 So 42 us a G alh 15 eg 56 os Vi 159 8 eir a1 60 8 Cé u 1 Ba 65 8 r Ca ros va 167 lei 5 ro 17 02 Ai M res 1 ar 70 q En ues 5 a tre m 175 ez 0 es 17 G ar re 57 tt Al 179 or na 9 18 Ca 0 m ilo 2 18 O r ti 25 gã o 18 36
0%
20%
49% 44%
89% 83%
VOS
VSO
SVO
F . The evolution of Subject-Verb-Object order in matrix declarative transitive sentences (by author) 100.0%
91.3%
87.3%
90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0%
52.5% 36.7%
46.3% 39.0%
20.0% 10.0%
14.7%
10.8% 1.9%
10.9%
0.0% 16th century
17th century SVO matrix
18th century VSO matrix
5.8% 2.8% 19th century
VOS matrix
F . The evolution of Subject-Verb-Object order in matrix declarative transitive sentences (mean value by century)
In this study,2 we focus on the position of lexical subjects in transitive sentences, contrasting the three possible orders for the subject, SVO, VSO and VOS. Figures . and . show that the evolution of word order in transitive sentences is entirely comparable 2
The Corpus that is at the basis of this study is composed of the following texts: sixteenth century: Pero Magalhães de Gândavo (b. ), Historia da Província de Santa Cruz vulgarmente chamada Brasil; Fernão
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
with the overall evolution of the position of subjects. Until , in spite of some variations among authors, post-verbal subjects are generally more frequent than pre-verbal subjects.3 This state of affairs abruptly changes among authors born at the beginning of the eighteenth century, which display per cent pre-verbal subjects. The expansion of the use of pre-verbal subjects continues in the nineteenth century at the expense of VSO, which thus falls below per cent. Before we discuss the nature of this change, we show in the next section that, in ClP, post-verbal subjects are not only much more frequent than in the modern language, but they can also receive a distinct interpretation. From a semantic point of view, their felicity conditions are different as well, being associated with interpretations that are no longer available today.
. The qualitative change Costa (: –) describes the interpretation associated with the three possible Subject-Verb-Object orders in EP in the following terms (where ‘focus’ means ‘new information’).4,5 (a) The whole sentence is focused: SVO; (b) The object is focused: SVO; (c) The subject and the object are focused: VSO; (d) Only the subject is focused: VOS. The important point is that in sixteenth- and seventeenth-century texts, VSO and VOS clauses can receive completely different interpretations: (i). Both orders can be interpreted as new information (out of the blue), against (a). (), from André Barros (b. ) and (), from Manuel Galhegos (b. ) illustrate respectively the cases of VOS and VSO.
Mendes Pinto (b. ), Peregrinação; Diogo do Couto (b. ) Décadas; Frei Luís de Sousa (b. ) A vida de Frei Bertolameu dos Mártires; seventeenth century: Manuel Galhegos (b. ) Gazeta; Pe Antonio Vieira (b. ) Sermões; Maria do Céu (b. ) Vida e morte de Madre Elena da Cruz; André de Barros (b. ) Vida do apostólico Pe Antonio Vieira; eighteenth century: Cavaleiro de Oliveira (b. ) Cartas; Matias Aires (b. ) Reflexões sobre a vaidade dos homens; Marquesa de Alorna (b. ) Cartas; J. D. Rodrigues da Costa (b. ) Entremezes de cordel; nineteenth century: Almeida Garrett (b. ) Teatro; Marquês de Fronteira e d’Alorna (b. ) Memórias; Camilo Castelo Branco (b. ) Maria Moisés; Ramalho Ortigão (b. ) Cartas a Emília. The authors born at the very end of their century, Galhegos () and Garrett () were grouped with the authors born in the next century. 3 A remarkable exception to this claim is Fernão Mendes Pinto (born in ), whose atypical per cent use of SVO yields a strong effect on the mean value of the sixteenth century. It is worth emphasizing, however, that this author is very similar to his contemporaries with respect to the high frequency of preverbal non-subjects, which is typical of V-to-C grammars (cf. Galves and Paixão de Sousa ). 4 Ambar (: –) describes subject-verb and subject-auxiliary inversion in the context of preposed constituents in EP. She shows that this inversion can be obligatory in certain contexts, namely when the verb is semantically weak. The important point for our purposes is that she concludes that: ‘from the point of view of the textual (informational) structure, the subject that, as a result of Subject-Verb inversion, appears in post-verbal position is interpreted as focus, while the preposed constituent is interpreted as topic’ (Ambar : ; our translation). 5 Duarte (: ) summarizes this set of constraints in the following manner: ‘[in Italian and] in Portuguese declarative sentences with transitive and ditransitive verbs, in neutral contexts (out of the blue), post-verbal subjects are legitimate only if the internal arguments of the verb are not full lexical NPs’ (our translation).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese
() Aos de Janeiro de desferiu as velas o Padre at.the of January of loosened.SG the sails the Father Veira para aquele maior teatro do mundo Católico. Vieira to that major theater of.the world catholic ‘On January , Father Antonio Vieira left for that major theatre of the Catholic world.’ () Em várias partes das fronteiras fizeram os castelhanos fumo in several parts of.the borders made the Castilians smoke ‘The Castilians made fire on several parts of the border.’ In both cases, the whole sentences provide new information. In Costa’s terms, this means that they are focused, requiring SVO in EP. In (), the subject is not interpreted as narrow focus, which would be required in EP to license VOS. We will come back later to its interpretation (cf. ..), but suffice it to say that ‘Padre Vieira’ is the main character of the text, which is in fact his own biography. In (), neither the subject nor the object receive narrow focus interpretation, i.e. answering the question ‘who did what?’, which would be required in EP to license VSO. Thus, the sentence does not mean that whoever made fire were the Castilians, but it simply asserts that the Castilians made fire in many places on the borders. It is a piece of information uttered in the first newspaper published in Lisbon: Manuel Galhego’s Gazeta. (ii) In VOS, the subject is not obligatorily new information, against (d). In (), from Antonio Vieira (b. ), the focus of the sentence is clearly on the PP ‘da parte austral’, since it is explicitly focalized by ‘somente’ only. ‘O texto’ refers to a text previously mentioned by the author. () faz menção o texto sómente da parte austral makes mention the text only of.the part austral ‘The text only mentions the austral part.’ (iii) VSO and VOS are felicitous when neither the subject nor the object are focused, against (c) and (d). The context for () and (), from Antonio Vieira (b. ), which is given in the excerpt between the brackets, shows that neither ‘Ezechiel’ nor ‘a parede’ (‘the wall’) is new information in the sentences ‘Rompeu a parede Ezechiel’ (‘Ezechiel broke the wall’) and ‘Passa Ezechiel a terceira parede’ (‘Ezechiel passes through the third wall’). On the contrary, those two terms are given in the immediate prior context. We also see that in ClP, VSO and VOS seem to be associated with the same felicity conditions, which is not the case in EP, according to Costa’s description. ()
[Cuidas tu, Ezechiel, diz Deus, que não há aqui mais que o que apparece? Ora, rompe essa parede, e verás.] ‘Do you think, Ezechiel, says God, that there isn’t more here than it seems? Then break this wall and you shall see.’ Rompeu a parede Ezechiel . . . broke.SG the wall Ezechiel ‘Ezechiel broke the wall’
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
() [Por deante, Ezechiel, diz Deus terceira vez] ‘Go ahead, Ezechiel, God says for the third time.’ Passa Ezechiel a terceira parede passes Ezechiel the third wall ‘Ezechiel passes through the third wall.’ (iv) When the object is focused, the order of the sentence is not obligatorily SVO, against (b). In the sentences () and () below, from Manuel Galhegos (b. ) and Luis de Sousa (b. ), respectively, the direct objects are the new information in their respective sentences, but the subjects of those sentences are not pre-verbal. In this case again, the referents of the subjects were previously introduced in the narrative as the protagonists of the story.6 () tomaram-lhe os nossos algumas armas, e munições took.them the ours some weapons, and bullets ‘Ours took some weapons and bullets from them.’ () Começou el-Rei a igreja de São Vicente. begun.SG the.King the church of São Vicente Fundaram os estrangeiros a de Nossa Senhora. founded.PL the foreigners the.one of Our Lady ‘The king began the church of São Vicente. The foreigners founded the Church of Our Lady.’ We now turn to the issue of the interpretation of the post-verbal subjects. (v) Post-verbal subjects can be interpreted as topics, against (c) and (d). The more frequent interpretation of post-verbal subjects in ClP is that of either ‘familiar topic’ or ‘continuing topic’ (cf. Frascarelli and Hinterholzl ). According to the genre and content of the texts, familiar topics are either proper names which refer to the main characters of the narratives, God and other religious entities (the Devil, the Holy Spirit), frequently quoted authors (the authors of the Gospels in the Sermons), or abstract or generic entities like ‘the enemies’, ‘humankind’, vices and virtues, etc. In the biography of Frei Bertalomeu dos Mártires (Luis Sousa, b. ), twenty occurrences out of a total of forty post-verbal subjects refer to the Frei, who is the main character of the work. We also find recently mentioned notions and entities, generally preceded by the demonstrative ‘este’ (this). Working with unaccusative sentences, Eide () labels these kinds of topics as ‘background’. This fact, already evidenced in the examples above, is illustrated again in (–), from Luis de Sousa (b. ), André Barros (b. ), and Manuel Galhegos (b. ), respectively: () Começou Frei Bertolameu seu noviciado desassombradamente began.SG Frei Bertolameu his novitiate boldly ‘Frei Bertolameu began his novitiate boldly.’
6 Possibly, sentence () may receive another interpretation, also available in EP, in which it would answer the question ‘who did what?’.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese ()
Estimulavam estes golpes o ânimo ao valor; stimulated.PL these blows the cheer to.the valour ‘These blows stimulated excitement for valour.’
() [Quase todo este mês ventou, choveu, e nevou,] ‘Almost all this month, it winded, rained, and snowed,’ e fez muito dano a tempestade and made.SG much harm the tempest ‘and the tempest did much damage.’ These examples nicely show the role played by post-verbal subjects from the informational structural point of view. In (), as in (), it refers to the main character of the work. As mentioned above, this is the case for half of the post-verbal subjects in the text. Other familiar topics are found in examples ()–(), either because their referents are frequently recurrent in the texts (the Castilians in the Gazeta, the quoted texts in the Sermons) or because they have just been introduced in the previous sentences. In this case, the notion of ‘familiar’ comes very close to the notion of anaphor. Anaphoric terms can be repetitions, as in the case of () and (), or summarizing terms as in (), where ‘tempestade’ anaphorically refers to ‘ventou’ was windy, ‘choveu’ rained, and ‘nevou’ snowed. Moreover, they also appear, as expected, with an anaphoric determiner, as in (). In the next section, we address the question of the nature of the change between ClP and EP. We argue that a syntactic change was at the origin of both the difference in interpretation of post-verbal subjects and the loss of VS as a dominant pattern of word order.
. The nature of the change The natural question is what produces the quantitative and qualitative difference between ClP and EP, and, correlatively, what is the nature of the change that happened at the beginning of the eighteenth century. Two hypotheses are available. The first hypothesis is that what happened was a grammatical change, which yielded a change in the interpretation of post-verbal subjects. ClP was a V-like language, as argued by the authors mentioned above, and at some point, the movement of the verb to a position in the C layer (Fin, according to Antonelli ) was lost. This entails that the high position for post-verbal subjects was lost too, leaving as a unique option a low position associated with focus interpretation (we will come back later to this point, cf. ...). From this point of view, post-verbal subjects can receive a topic interpretation in ClP because they occupy a high position in the clause. When the verb ceases to move to a position in the C layer, this position is no longer available. The second hypothesis is that what happened was a discursive change, more precisely, a restructuring of information structure. Eide (: ) claims that ‘The restructuring involved that only information that is new can occur post verbally, while background information is placed before the verb.’
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
From this point of view, ClP and EP would generate the same structures, but they are interpreted differently at the C-I interface (and possibly at the A-P interface as well).7 In what follows, we give independent empirical arguments in favour of the first hypothesis, by showing that both the verb and the post-verbal subject do occupy high positions in ClP.8 In order to make this point, we first focalize ClP as a synchronic grammar, capitalizing, on the one hand, on its differences with respect to EP, and on the other hand, on its similarities with V languages. We then bring further evidence from diachronic aspects related to subjects in the evolution from ClP to EP. .. There is a high position available for post-verbal subjects in ClP ... Two different positions for post-verbal subjects Working in the cartographic framework of Rizzi (), Antonelli (: –) argues that, in ClP, tensed verbs move to Fin and post-verbal subjects are either in Spec/TP or in a lower position since low adverbs like ‘bem’ (‘well’) can occur either after or before them, as illustrated by the following sentences, from Sousa (b. ). () E nos gasalhados e abraços mostraram os cardeais and in.the welcome.PL and greetings showed.SG the cardinals legados bem este contentamento; delegates well this satisfaction ‘In the welcome and greetings the cardinal delegates showed this satisfaction well.’ () E quadra-lhe bem o nome de Piemonte . . . and fits-it.DAT well the name of Piemonte ‘And the name of Piemonte fits it well’ Only the order illustrated in () is possible in EP (cf. Costa ). Antonelli () shows that in ClP, two different interpretations are available for post-verbal subjects. Depending on the structural context, they can be interpreted as new information (focus), or as given information (familiar topics, as we argued above). In some cases, this is ambiguous because no other element of the clause suggests what their position in the structure may be. In this case, they are compatible with either interpretation, depending on the textual context. But the presence of adverbs like ‘bem’ (‘well’) indicates the position of the subject, which in turn associates it with a new or given 7 In a more recent text (cf. Eide ), Eide adopts the view that there was a syntactic change provoked by a previous change in information structure and prosody. 8 It is outside the scope of this chapter to address the issue of the reason for the change. Note that the loss of V has been reported and studied in many languages. To continue with the case of Romance languages, the cause of the loss of V in French has been imputed by Yang () to the existence of prodrop. This cannot be the case for Portuguese since no significant change has occurred in the frequency of null subjects in this language (cf. Galves and Paixão de Sousa ). Other scholars have argued that the origin of the loss of V in the history of French is to be found in the loss of the Germanic prosody (cf. Adams , etc.). Since a prosodic change is likely to have occurred after the sixteenth century in Portuguese (cf. Teyssier ; Frota et al. ), this hypothesis is worth investigating (cf. Galves and Galves ; Eide ). Galves and Paixão de Sousa () rely on the co-occurrence in time of the changes in clitic-placement and in subject position to claim that they were both due to a change in the prosodic patterns associated with marked constructions with enclitic clitics, which could have led speakers to reanalyse the pre-verbal subject position as the unmarked position of subjects.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese
informational status. In the next section, we bring other pieces of evidence that tensed verbs, and not focalized post-verbal subjects, occupy a high position. ... Other pieces of evidence Belletti (: –) shows the following contrast between Italian and Icelandic and concludes that ‘The contrast between ()–() [here (a and b)] and ()–() [here (a and b)] suggests that a further, higher position is available for subjects in Icelandic, but not in Italian.’ () a. *Capirà/spiegherà Maria completamente (al direttore). will.understand/explain Maria completely (to.the director) b. Capirà/spiegherà completamente Maria (al direttore). will.understand/explain completely Maria (to.the director) ‘Maria will understand/explain it completely (to the director).’ () a. það luku sennilega einhverjir studentar alveg verkefninu. there finished probably some students completely the.assignment b. *það luku sennilega alveg einhverjir studentar verkefninu. there finished probably completely some students the.assignment ‘Some students probably finished the assignment completely.’ (Bobaljik and Jonas (: ) Many examples representative of the Icelandic order are available in the classical period of the CTB; see, for instance, () above and ()–() from André Barros (b. ), where the subject follows the verb but precedes a manner adverb: () e sendo este negócio dos maiores, que podia ter uma and being this deal of.the greatest, that could have a Monarquia, o fiou o animoso Rei inteiramente da Monarchy it.ACC entrusted the brave king entirely to.the intelligência e indústria de um só homem; intelligence and industry of a only man ‘and being among the greatest of deals, which could have a monarchy, the brave king entrusted it entirely to the intelligence and industry of just one man.’ () Referiu o depois em Carta sua o mesmo Padre Vieira referred.SG it after in letter his the same Padre Vieira formalmente assim: formally this.way ‘Afterwards Padre Vieira himself referred to it formally in a letter this way’ The position of adverbs also evidences that direct and indirect objects can be moved outside of the VP, leaving behind a focused phrase that follows the adverb, as can be seen in the following sentence from Maria do Céu (b. ): () Em muytas occasiões advirtio Deos à Madre Elena interiormente in many occasions advised God to.the Mother Elena internally o que convinha à sua honra the what suited to.the her honour ‘On many occasions God internally advised Mother Elena what suited her honour.’
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
It is interesting to note that in (), the two DPs that occur between the verb and the adverb are clearly familiar topics: ‘Deos’ (God) and ‘Madre Elena’ (Mother Elena), the main character of the text, since it is her biography. This suggests that there is more than one position available for familiar topics between the verb and the boundaries of vP. We come back to this issue below. The VS Floating Q order provides further evidence for the high position of the post-verbal subject, as illustrated in (), from Manuel Galhegos (b. ): () Estavam já os esquadrões ambos para marchar were already the squadrons both to march ‘The squadrons were both already about to march.’ Finally, one of the arguments presented by Costa (: ) in favour of a low position of post-verbal subjects, against the movement of the verb over a high subject, is the possibility of the Aux-Past Participle-Subj order.9 In the corpus, we find both Aux-Subj-Past Participle and Aux-Past Participle-Subj. The former is exemplified in () and (), from Diogo do Couto (b. ) and Maria do Céu (b. ), respectively: () A este tempo estava Dom Christovão curando-se em casa da at that time was Dom Cristóvão healing.REFL in house of.the Rainha; queen ‘At that time Dom Cristóvão was recovering at the queen’s house.’ () havia esta Religiosa professado de Fevereyro, e falecido a had this nun professed in February and died on dezoito do Abril seguinte eighteen of.the April following ‘This nun had professed in February and died on the eighteenth of the following April.’ As for the cases in which subjects follow the past participle, in many cases they are clearly foci, as can be seen in (), from Manuel Galhegos (b. ). This means that they occupy the low subject position identified by Antonelli (). However, there are cases in which post-participial subjects are not likely to be foci, as exemplified in (), from Maria do Céu (b. ), where the post-verbal subject is a familiar topic. () Está-se fabricando, por traça de Antônio Pessoa Campo, is.IMPERS building, by conspiracy of Antonio Pessoa Campo, uma carreta para o tiro de Dio; a chariot for the charge of Dio ‘They are building, by conspiracy of Antonio Pessoa Campo, a chariot for the charge of Dio.’
9 Costa does not comment on the possibility of Aux-Subj-PP, which he does not seem to consider to be excluded in EP.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese
() A esta Religiosa hauia previsto a Madre Elena a sua to this nun had forecast the Mother Elena the her morte pouco tempo antes de succedida; death little time before of succeeded ‘To this nun, Mother Elena had predicted her death shortly before it occurred.’ We shall assume, without further discussion, that the order exemplified in () is produced by an additional movement of the past participle to the periphery of TP. ... The derivation of VOS As for VOS, the fact that the subject in this order is not obligatorily interpreted as narrow focus indicates that it is in a high position too and that, accordingly, VO moves to a higher position than in modern languages. Adapting Belletti (, )’s analysis, we propose that VO moves to Spec/Fin, in contrast to EP in which, according to Costa (), VOS order is derived by scrambling of the object. VO movement is licensed in ClP since, as argued by Antonelli (), the φ-features of the verb are able to satisfy the edgefeatures of Fin, as they do with the edge-features of T in Modern pro-drop languages (cf. Alexiadou and Anagnostopolou ). Two facts support this analysis. (i) VO is a constituent that excludes the subject. This is shown by the following sentences from André Barros (b.) and Luis Sousa (b.), respectively: () de tal sorte lhes conciliou os ânimos, e humanou os in such sort them.DAT conciliated the spirits, and humanized the afectos a sinceridade desta Informação, affects the sincerity of.this information ‘The sincerity of this information reconciled the spirits and humanized the affects, in such a way that . . . ’ () Ornam a capela e acompanham a sepultura muitas estátuas de ornate the chapel and accompany the sepulture many statues of finos mármores fine marble ‘Many statues of fine marble decorate the chapel and accompany the sepulchre.’ (ii) In VOS, the semantic properties of the object are different in ClP and EP. Costa () proposes several arguments for scrambling of the object in VOS in EP. One of them is that ‘if the object is indefinite, the acceptability of the sentence is not as good, and it clearly degrades with non-specific indefinite DPs’. In the following sentences, from Manuel Galhegos (b. ), we observe VOS sentences in which objects are non-specific indefinite DPs (‘geral sentimento’—‘general sorrow’; ‘água’—‘water’; ‘missa’—‘mass’): () causou geral sentimento a sua morte pelo muito que caused general sorrow the his death for.the much that perderam os púlpitos de Portugal. lost the pulpits of Portugal ‘His death caused general sorrow for all that the pulpits of Portugal lost.’
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
() deu-lhe água às mãos o visconde de Vila Nova de gave-him.DAT water to.the hands the Viscount of Vila Nova da Serveira, Serveira ‘The Viscount of Vila Nova de Serveira put water on his hands.’ () Disse missa o bispo capelão-mor, said mass the bishop arch-chaplain ‘The bishop arch-chaplain said the Mass.’ In fact, the kind of sentences illustrated by ()–() come close to what Belletti claims to be the rule in Italian, which is that ‘the more “prototypical” the situation illustrated by the VO sequence is, the better the status of VOS becomes’ (Belletti : ). ‘Causar sentimento’ (‘to cause sorrow), ‘dar água’ (‘to give water’, in a religious ceremony), ‘dizer missa’ (‘to say Mass’) are indeed prototypical situations. This contrasts with Costa’s claim that objects in VOS in EP have to be specific, and could be understood as a semantic property of VO movement as opposed to scrambling of the object.10 In this section, we have presented synchronic evidence that in ClP V(X)S is syntactically different from V(X)S in EP. In the next section, we shall see that there is also diachronic evidence that a change affected the position of pre-verbal subjects, and consequently the position of tensed verbs. .. Evidence that the position of the subject, and consequently of the verb, is diachronically reanalysed ... The sharp increase in the frequency of SVO As shown by Figure ., until , in the texts of many authors, SVO is less frequent than VSO. After , the reverse happens with all the authors, and VSO becomes a universally marked construction. In the previous sections, we have argued that post-verbal subjects did not occupy the same position in ClP and EP. In the remainder of the chapter, we will consider other changes observed with the authors born after , which support the claim that not only has the position of post-verbal subjects changed, but also, preverbal subjects do not occupy the same position. The conjunction of these two facts strongly supports the hypothesis that verb-movement to a high position was lost. ... The emergence of the XSV-cl order Galves and Paixão de Sousa () observed that with the authors born before , both the V orders XSV and SXV appear with proclisis, but only SXV is attested with enclisis. They interpret this fact as a restriction on the position of subjects when they are in an external position (as shown by the position of the clitic). In this case, they can only be the topmost topic of the sentence. In the texts of the authors born in the first half of the eighteenth century, the 10 Belletti () insists on the fact that VOS in Italian is extremely marginal. A look at Figure . in section . shows that this is also true in ClP, and that, in this language, it seems much more prone to individual and, perhaps, stylistic variation than VSO. A larger corpus is needed to check whether the very low frequency of this order in the first half of the sixteenth century is due to the authors considered or whether it is a general property of the period.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese
formerly impossible XSV-cl order becomes as frequent as the other V orders with enclisis. This suggests that pre-verbal subjects cease to occupy a topic position. ... The high frequency of pronouns in VSO sentences It is worth noting the increasing frequency of the realization of post-verbal subjects as pronouns. The frequency of pronominal post-verbal subjects in VSO, very low with the authors from the sixteenth and seventeenth centuries (from per cent to per cent), is predominant with the authors from the nineteenth century, of occurrences (Almeida Garrett, b. ), / (Marquês de Fronteira e d’Alorna, b. ), / (Camilo Castelo Branco, b. ), and / (Ramalho Ortigão, b. ), respectively. This approximates the syntax of inversion found with the authors born after to the syntax of modern languages like Italian.11 Belletti (: ) notes that ‘a systematic class of exceptions to the general ban against VSO in Italian is provided by cases in which S corresponds to a personal pronoun’. She then makes the following claim: The contrast between the personal pronoun and the lexical noun phrase suggests that pronouns should avail themselves of a further position in the postverbal domain, which is excluded for lexical noun phrases. This further subject position should be higher than the one filled by the lexical noun phrase.
It is remarkable that, when post-verbal, the pronouns are always contiguous to the verb. Furthermore, no occurrence of VOS, S being a pronoun, appears in the corpus. In the following example from Marquês de Fronteira e d’Alorna (b. ), and drawn from a non-transitive clause, the position of the subject pronoun ele (‘it’) before the adverb sempre (‘always’) shows that post-verbal subject pronouns occupy a high position in the nineteenth century:12 () Constava ele sempre de seis grandes tabuleiros com peixe consisted it.NOM always of six big trays with fish cozinhado de várias formas, cooked in various manners ‘It always consisted of six big trays with fish cooked in several ways.’ ... The loss of the matrix-subordinate asymmetry Until now, we have only considered matrix contexts. In this section, we shall look at the position of subjects in embedded declarative contexts. Figure . shows the distribution of SVO, VSO, and VOS in both embedded and matrix contexts, over time.
11 In a nutshell, VSO is only possible in Italian when S is not focused, which is the case of pronouns. According to Belletti (), this is due to the fact that, in Italian, the position where Accusative case is checked is higher than the position occupied by post-verbal full subjects. Pronouns are legitimate in postverbal position because they occupy a higher position. This is different from EP, which has VSO, with O focus. 12 Cf. Rinke ()’s discussion on the position of the post-verbal subject in Old Portuguese.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
100.0%
93.1%
92.7% 91.3%
90.0% 80.0%
87.3%
80.8%
70.0%
68.2%
60.0% 52.5%
50.0%
46.3%
40.0%
20.0% 10.0% 0.0%
39.0%
36.7%
30.0%
19.0% 14.7% 14,7% 12.8%
14.9% 10.9% 4.3% 16th century SVO matrix
17th century SVO sub
VSO matrix
10.8% 4.81% 4.9% 4,9% 18th century VSO sub
5.8% 3.0% 2.0% 0.0% 19th century VOS matrix
VOS sub
F . SVO, VSO, and VOS in matrix and embedded clauses
Figure . evidences, with the authors born before , a clear difference in frequency between matrix and embedded clauses. It turns out that in the latter, SV order is clearly predominant. It must be noted that for all the authors, except Fernão Mendes Pinto (b. ) (cf. note ), the rate of post-verbal subjects is much higher in root than embedded contexts. This contrast disappears from Matias Aires (b. ) on. Antonelli () argues that in subordinate clauses that have a lexical complementizer, post-verbal subjects can only occupy a low position since the verb is prevented from moving to Fin, which is already filled by the complementizer. Evidence for this analysis comes from embedded sentences in which C is not lexically realized. In this case, lexical subjects are categorically post-verbal, which strongly suggests that the verb is in Fin. From this point of view, pre-verbal subjects do not occupy the same position in main and subordinate clauses. In the former, they occupy a position in the left-periphery of the sentence, as with other topics. In the latter, however, they are in Spec/T, or adjoined to T. This analysis nicely accounts for the difference observed in Figure .. It also supports the hypothesis that a grammatical change occurs after , in which pre-verbal subjects cease to be treated as topics in main clauses, and come to occupy a specific subject position in the specifier of TP. In the new grammar, the former asymmetry between main and embedded clauses therefore ceases to exist. This analysis encounters an apparent empirical problem. Although less frequent, the VSO/VOS constructions exist in embedded clauses at a higher rate than in the modern language. Additionally, all the possible structures found in matrix clauses are found in subordinate clauses, as illustrated in (), from André Barros. Furthermore, the interpretation of the post-verbal subject in this sentence (‘as terras do Maranhão’—‘the lands of Maranhão’) is clearly that of a familiar topic. () é bem, is good, espiritual, spiritual,
que saiba o Mundo, que that know.SG the world that como veremos senão também as will.see.PL but also
não só no not only in.the no temporal, devem in.the temporal, owe.PL
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Subject inversion from Classical to Modern Portuguese
as terras do Maranhão à Religião da Companhia de the lands of.the Maranhão to.the Religion of.the Company of Jesus a felicidade, que logram. Jesus the happiness, that reach.PL ‘it is good that the world knows that not only on the spiritual level, as we will see, but also on the temporal level, the lands of Maranhão owe to the Religion of the Company of Jesus the happiness, that they achieve.’ However, examples like (), where the main predicate is the epistemic verb ‘saber’ (‘to know’), and the subordinate verb is preceded by a fronted adjoined clause, are compatible with Antonelli’s analysis, since they are likely to be dependent on CP recursion. The other cases of VS, in which S is not focused, appear in the absence of a lexical complementizer, as in (), where the subject ‘ella’ (‘she’) immediately follows the verb ‘teria’ (‘would have’). In this case, as argued by Antonelli (), the verb is itself in C. () entendendo lhe teria ella encomendado esta deligencia, understanding him.DAT would.have she asked this diligence, por ser sua familiar, entraraõ em huma Tribuna da for being her familiar, entered.PL into a tribune of.the Enfermaria que cahe para a Igreja, Infirmary that falls onto the Church ‘understanding that she would have asked him to be diligent, since he is her relative, they will go into a tribune of the Infirmary close to the Church The frequency of post-verbal subjects in embedded clauses in the classical period is therefore accounted for. There are more cases than in the modern period due to the availability of a high position for the post-verbal subject, but there are fewer cases than in matrix clauses during the same period since this high position is licensed either by CP recursion, which is in turn dependent on the main verbs. As emphasized by Antonelli (), the other case of VS in embedded clauses is when the embedded verb is in C, instead of a lexical complementizer, which is no longer an available option today.13
. Concluding remarks In this chapter, we provided evidence, against Eide (), that the change in the frequency and interpretation of post-verbal subjects from Classical Portuguese to Modern European Portuguese is due to a syntactic change that affected the position of the verb. The diachronic study of the syntax of transitive clauses in European Portuguese therefore brings more pieces of evidence to the claim that Classical Portuguese was a V-to-C language. This in turn licensed a high position for postverbal subjects, in which they were interpreted as familiar topics. At the turn of 13 On null C in ClP, see Antonelli (). Note that his analysis correctly predicts that once V-to-C is lost, null C is also lost.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Charlotte Galves and Alba Gibrail
the eighteenth century, V-to-C was lost, probably the effect of a reanalysis of marked topic constructions as non-marked subject constructions, and the VS order became dramatically less frequent and more semantically restricted, on its way to the modern state, in which the only interpretation available for post-verbal subjects is that of focus.
Acknowledgements This research was conducted within the Research Project ‘Portuguese in Space and Time: Linguistic Contact, Grammars in Competition and Parametric Change’ funded by Fapesp (grant /-). It also received support from CNPq (grant /-).
Source Tycho Brahe Parsed Corpus of Historical Portuguese. .
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
10 Analyticization and the syntax of the synthetic residue C H R I S H . R E I N TG ES AN D S O N I A C Y RI N O
. Introduction Syntactic variation and change provides us with a window onto the dynamic aspects of the human language faculty. A productive way of exploring the unity and diversity of grammatical structure has been in terms of universal principles and finite discrete parameters, which set limits on the ways that languages can differ. The Government and Binding premise of directly encoding parametric variation into a richly structured Universal Grammar (Chomsky ) has proven difficult to integrate into current minimalist theorizing (Chomsky ; Berwick and Chomsky ), which rather seeks to reduce the relative weight of the linguistic genotype (Universal Grammar) to a minimum. While this would seem to leave little or no room for parameterized syntactic configurations, parameters as such continue to be used as efficient analytical devices both for discovering systematic patterns in crosslinguistic variation, and, at a more abstract level, for the modeling of underspecification in grammar. The challenge for any explanatorily adequate theory of syntactic comparison must therefore be to delimit the parameter space in a principled way. In Kayne’s (, ) theory, all parameters are microparameters, i.e. local points of variation with limited clustering effects. In being connected to the feature content of individual lexical and functional items, microparameters can only affect those structures that contain the relevant item. Baker has revived the conception of macroparameters that define in one fell swoop the core structure of typologically distinct languages. Such large-scale parameters are heuristically significant in that they ‘point to loci of variation in the grammar as opposed to the lexicon’ (: ). To reconcile the micro- and the macroparametric approach to comparative syntax, Roberts and Holmberg () and Roberts (b) advance an alternative proposal, which derives macroparametric variation from aggregates of microparameters
Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Chris H. Reintges and Sonia Cyrino . First published by Oxford University
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
acting in concert for markedness reasons. This aggregate behaviour resides in the interaction of an underspecified UG, the Primary Linguistic Data and third factor related learning strategies, whereby a given structural parse is generalized to comparable combinations of syntactic elements. Many questions surrounding micro- versus macroparametric syntax arise anew in the context of grammar change when the language system at large is being affected. In previous work (Reintges , Reintges and Cyrino ), we have defended the utility of the notion of a macroparameter for the study of syntactic change and presented the Sapirian () drift towards greater analyticity as a case in point for a diachronic macrochange. As such, it exhibits the typical clustering effects—a point that will become clear while we proceed. The concern of the present study is with the flipside of analytic drift, to wit, the retention and survival of synthetic morphological structure in a context of widespread change. In what follows we shall attempt to verify the following hypotheses about the relic syntax of synthetically inflected verb forms in an emergent analytic system: ()
The parametric profile of non/productive morphology (i) In transitional grammars innovative morphological patterns spreading to new domains already conform to the new parameter setting. Unproductive or obsolete morphology conforms to the previous parameter setting and must therefore be acquired on an item-by-item basis. This particular corner of grammar, which Mackenzie and Van der Wurff () call ‘relic syntax’, belongs to the domain of microparametric variation. (ii) The synthetic residue can be integrated into the emergent analytic system after substantial reanalysis has taken place. Syntactic reanalysis involves changes in the feature content of the items involved, which subsequently operate within a delimited functional domain for feature checking/valuation purposes. (iii) The analyticization of verbal forms is intimately related to the scope of verb fronting processes insofar as the presence of an auxiliary element in the topmost temporal projection bars verb movement to this position. The shrinking of the available movement space leads to the exclusion of word orders previously derived by long verb movement to the left periphery of the clause (‘Verb Second’). Another syntactic consequence is the diminished positional freedom of adverbial modifiers.
We shall examine the global effects of synthetic-to-analytic drift in regard to two diachronic scenarios—one in which the analyticization process has almost, though not entirely been completed, and another one in which the process is still under way. Due to the partial incompleteness of the change, parts of the anterior synthetic system remain unaffected, albeit under very restricted conditions, as formulated in ()(i)–(iii) above. The empirical domain of our investigation concerns the analytic restructuring of the temporal systems of Brazilian Portuguese (Ibero-Romance [Indo-Germanic]) and Coptic (Ancient Egyptian [Afro-Asiatic], around third to fourteenth centuries
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
AD). Coptic has gone very far in abandoning its former synthetic features and exhibits a high degree of analyticity. Like all the other Romance language, European Portuguese underwent a major typological shift from the synthetically oriented structures of Latin towards the characteristically analytic structures of Romance (for an up-to-date comprehensive overview see Ledgeway ). The analyticization process is in a more advanced state in Brazilian Portuguese, where synthetically inflected tenses exhibit a decreasing productivity and are gradually being replaced by corresponding auxiliary verb constructions in the spoken language.
. The formal syntax of analyticization According to the World Atlas of Languages Structures (Dryer ), languages out of a total of , languages have tense and aspect suffixes as opposed to languages with tense-aspect prefixes and languages without tense-aspect affixes and stem changes. The latter group is mainly represented by analytic or isolating languages, which typically encode temporal and aspectual categories by means of free functional morphemes, either by auxiliary verbs or by non-inflecting particles or clitics. The analytic schema is cross-linguistically based on and diachronically evolves from monoclausal auxiliary verb constructions, in which the auxiliary verb carries all of the morphosyntactic information related to the lexical predicate, such as tense, aspect, mood marking, and agreement. This information is generally not replicated on the main verb, which instead appears in a deverbalized participial or infinitival form. The acquisition of nominal features is a good indication for the categorial change that the main verb has undergone, conceding that nominal categories can generally not be endowed with tense and finiteness features. The auxiliary verb may further be grammaticalized into a verbal particle, which expresses a particular temporal, aspectual, and modal value but which crucially lacks person and number specifications. With reference to analytic languages, Greenberg (: , ) acknowledges a category of uninflected auxiliaries, but excludes them from his pioneering study on word order correlations. For Dryer (: ) and Cinque (: , note ), on the other hand, the presence of inflectional morphology on a verbal item presents a decisive factor to distinguish tense-aspect auxiliaries from tense-aspect particles. Although we shall not discuss this issue further here, we conjecture that the deflection of tense, aspect, mood, and evidentiality (TAME) related particles is a byproduct of the analyticization process taken one step further. Anderson () offers a more detailed picture of the cross-linguistic distribution of the inflected auxiliary verb-plus-participle/ infinitive and the uninflected TAME particle-plus-infinitive constructions than can be provided here. The more abstract syntactic property corresponding to analytic structures is a division of labor between a defective verbal category without tense and finiteness specifications and an auxiliary element in the extended verbal projection, roughly
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
in the sense of Grimshaw (). Since the lexical verb is syntactically and morphologically divorced from the exponent of tense and finiteness, the verb– tense relation must be accomplished in the syntax via a more abstract feature computation procedure—the syntactic relation AGREE (see Chomsky ; Bobaljik and Wurmbrand and much related research). The auxiliary verb or particle acts as a probe for the nominalized verb, but the resulting probe-goal relation does not trigger verb movement, as schematically represented in the following tree diagram. ()
The agree-based verb-tense relation in thoroughly analytic systems TP Subject
AGREE
TP Tense Auxiliary verb Verbal tame particle
… VP
VP
Verb[+N-features]
Object
The AGREE-based relation between tense and the verb distinguishes analytic systems from agglutinative and fusional ones, in which verb raising to Tense supports temporal affixes and inflections that would otherwise be left stranded (on this point, see Cinque and Julien ). In a somewhat different proposal from ours, Roberts and Holmberg (: ), Roberts (b: –), and Holmberg and Roberts (: –) correlate the rise of analytic structures with some degree of morphological simplification and erosion on the one hand and with the loss of Verb-to-Tense movement on the other hand. However, analyticization is not necessarily an ‘all-or-nothing’ phenomenon, as it may leave a niche for residual verb movement. In line with previous studies on the syntax of tense and aspect (Giorgi and Pianesi ; Demirdache and Uribe-Etxebarria ; Julien ; Pesetsky and Torrego ), we shall take as our departure point the view that the temporal skeleton of the clause minimally contains two temporal projections TP and TP. The higher Tense head marks absolute tense distinctions such as [PAST] and [FUTURE], while the lower Tense head indicates aspectual oppositions and relative tense. Semantically, aspect has narrower scope than tense, as it is concerned with the internal temporal structure of the event itself. This is reflected in the syntax in that the aspectual Tense head is positioned below the Tense head, which modifies the entire sentence.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue ()
The limited scope of verb movement in emergent analytic systems
TP1 (Absolute Tense) Subject
TP1
Tense1 Auxiliary verb Verbal tame Particle
TP2 (Aspect, Relative Tense) TP2 Tense2 Verb
VP
VP
Object
Illicit Verb-to-Tense1 movement Licit Verb-to-Tense2 movement As free functional morphemes, the tense-aspect auxiliary verb or the verbal particle prevent the deflected lexical verb from raising and head-adjoining to the Tense head. In this way, analyticization severely restricts the scope of verb movement operations. However, verb movement as a core syntactic process need not be entirely dispensed with, as lower functional heads are still available as landing sites. The structure in (), in which the higher tense head is no longer available as a landing site for verb movement while the lower head still is, provides the relevant configuration for tenseto-aspect reanalysis. Tense-to-aspect reanalysis is a semantically induced change, through which an erstwhile syncretic tense-aspect form loses its absolute tense (Tense) features, while retaining its aspectual (Tense) ones. This process of feature economization may but need not go together with morphological change. It has visible consequences in the syntax, though, as the reanalysed form raises as high as Tense but raises no further.
. Deriving verb movement options Much comparative work over the past decades has singled out ‘rich’ agreement, i.e. the discrete morphological representation of person and number features in finite verb paradigms as a key factor motivating both the omissibility of weak subject pronouns and the presence of verb movement to Tense in a given language (see Roberts ; Bobaljik ; Koeneman for three representative though diverging views on the Rich Agreement Hypothesis). Although it is tempting to connect the two parameters to a single morphological property, there is an empirical problem
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
T . The typology of null subject and verb raising languages Rich agreement
Rich tense
Null subjects
Verb-to-Tense movement
Yes
Yes
Yes
Yes
Italian, Spanish, European Portuguese
No
No
No
No
English, Mainland Scandinavian
No
Yes
No
Yes
French, Middle English
Yes
No
Yes
No
No clear examples
(Adapted from Biberauer and Roberts : )
with this approach, as Verb-to-Tense movement is also attested in non-null-subject languages, French being a well-known case in point (Pollock ). Against this background, Biberauer and Roberts () advance an alternative proposal, in which the verb movement asymmetries between Germanic and Romance languages are correlated with the number of paradigmatic oppositions within the temporal system. The Richness of Tense Parameter includes within its scope modal and aspectual categories, as in the case of the Romance subjunctives, futures, conditionals and imperfectives. The Rich Agreement Parameter, on the other hand, is correlated with the availability of null subjects. The disentangling of the two parameters gives rise to the following typology of null subject and verb raising languages. In consistent null subject languages with rich tense paradigms, the finite verb moves to Tense (and possibly higher, as in the case of modals) and null subjects are licit under the appropriate discourse conditions. In languages in which both agreement and tense are impoverished, null subjects are illicit and there is no Verb-to-Tense movement. Languages with poor agreement and rich tense inflection represent an intermediate case, in which the finite verb raises to Tense while null subjects are generally not permissible. There are no clear examples of the reverse case of rich agreement and poor tense inflection. However, Icelandic seems to be a particularly puzzling case in this regard, since rich agreement does not license null subjects and Verb-to-Tense movement applies in the absence of rich tense inflection (see Biberauer and Roberts : – for further discussion, including other problematic cases). Biberauer and Roberts’s typological model just outlined makes the correct predications for finite verb raising and pro-drop as found in the classical null subject languages and straightforwardly accommodates the exceptional case of French. Conversely, many problems of the Rich Agreement Hypothesis resurface. To rephrase the argument given by Bobaljik (: –), since the correct generalization is a one-way implication from rich tense morphology to Verb-to-Tense movement, there must be causes for verb movement other than the number of paradigmatic temporal, aspectual, and modal oppositions in the language’s inflectional system. If such factors can be shown to be operative in languages in which the tense inflection is rich, then appeal to this kind of inflection cannot be part of a principled explanation. In Biberauer and Roberts’s parametric analysis, richness of tense and agreement is understood exclusively in relation to inflection. On the other hand, the proposed
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
typology is flexible enough to include languages with characteristically analytic structures, with Niuean (Malayo-Polynesian [Austronesian]) being a case in point. It fares less well though with languages with mixed synthetic-analytic behaviour, as we are about to demonstrate in two case studies on the verbal morphosyntax of Coptic Egyptian and Brazilian Portuguese. The main issues involved are, first, the question of morphological productivity (Bauer ; Dressler ) and second, the pivotal role played by residue verb movement in preempting argument extraction.
. Verb movement and A(rgument)-syntax in a highly analytic system Coptic represents the most analytic stage of the Ancient Egyptian language, which thus marks the endpoint of long-term morphosyntactic change (see Reintges for the details of the historical process). Since the synthetic-to-analytic drift did not go all the way through, the language has retained a considerable residue of synthetically inflected infinitives and statives, which move to the highest functional head of the inflectional domain but move no further. Long verb movement into the left periphery is excluded on the grounds that this is the domain in which to merge a broad range of grammatical particles, which act as movement-blocking categories. .. Tense-rich analyticity and the cartography of the Coptic clause Within the parameters of Biberauer and Roberts’s typology, Coptic would qualify as a non-null subject language with a rich tense system. Due to the complete absence of person–number agreement, the language does not allow null subjects of any kind. Rather, referential and non-referential pronominal categories are always lexicalized by means of personal pronouns. Richness of Tense is manifested in an elaborate
T . The inventory of the most common grammatical particles in Coptic Tense/Aspect/Mood/Evidentiality particles Pre-subject particles Perfect
Relative particles
Pre-verbal particles a, are
Epistemic Future
e, ere na
et
Negative Perfect
əmpe
Deontic Future
e
ənt
Habitual
ʃa, ʃare
Conditional
ʃan
ən
Negative Habitual
me, mere
Negative Auxiliary
təm
Preterit
ne, nere
Clause-chaining particles ən, ənt, ənta
Inferred Evidential
tare
Optative
mare
Negative Future
ənne
Terminative
ʃante
Negative Perfective
əmpate
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
inventory of tense-aspect-mood-evidentiality (TAME) marking particles. These specialized auxiliaries can be further subdivided into two positional classes of pre-subject and pre-verbal particles. There is also a smaller set of relative complementizers and clause-chaining particles. As for their categorial status, one might envisage an analysis of Coptic TAME particles along the lines of Niuean particles, which Massam (: ) and Biberauer and Roberts (: ) treat as tensed complementizers. An analysis along these lines would seem to predict such particles to be mutually exclusive with embedding complementizers, conceding that there can only be a single Comp/Force projection per clause. This prediction is indeed borne out by Niuean. As Massam (: –) points out, pre-verbal particles must be dropped when a lexical complementizer is present. No such co-occurrence restriction applies to Coptic TAME particles, which combine with relative complementizers and other TAME particles to form complex clusters of up to four elements (Reintges a: ). ()
Internally complex particle cluster e ne nt a=i eire gar ən-u-tʃi ən-kʲons REL PRET REL PERF=SG do.ABS PCL PREP-INDEF.SG-take.ABS PREP-violence ‘For if I had committed an act of violence . . . ’ (Acts , )
In terms of word order typology, Coptic can be classified as an example of an SVO language, in which free functional morphemes carrying temporal, aspectual, modal, and evidential information are placed in front of the clausal subject (but see Baker for a claim to the contrary). The resulting surface order T/Aux SVO order can be identified as basic word order on the grounds that this is the linear order, which involves a minimal amount of phrasal structure, as compared with other word orders. ()
Basic T/Aux SVO order a tə-sophia ket u-ɛːï na=s PERF DEF.F.SG-wisdom build.NOM INDEF.SG-house for=F.SG ‘Wisdom has built a house for herself.’ (Proverbs , )
In order to reflect the rich tense system of auxiliary particles, we shall propose a complex functional superstructure for verbal tense and aspect. Pre-subject particles such as the Perfect particle a are located in an uplifted Tense(Past) node, which corresponds in structural height to the Fin(iteness) node of the Rizzian () cartography. The formation of compound tenses formed with the Preterit particle ne provides evidence for yet another past tense head Tense(Preterit), with the positive specification of the two tense heads giving rise to the Pluperfect ne a=f sɔːtәm ‘he had heard’ (cf. example above). Pre-verbal particles, on the other hand, are merged into a Tense(Future) head, whose specifier hosts the pre-verbal subject DP. In hosting the epistemic future particle na, the deontic future particle e, and the conditional particle ʃan, the Tense(Future) node may very well be analysed as a root modal head, as these particles are arguably related not only to future tense but also to realis modality. Be this as it may, the syntax of Coptic TAME particles lends further support for Julien’s () hypothesis that the Tense(Past) and the Tense(Future) nodes are universally contained within the temporal skeleton of the clause.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue ()
The sequence of temporal heads in Coptic clause structure
TP(Preterit) Tense (Preterit) TP(Past) Particle ne Tense [±Past] tame particle
TP(Future)
Subject
TP(Future)
Tense [±Future] TP(Aspect) tame particle Verb Tense [Aspect]
VP
VP Object
The merger of pre-subject TAME particles into the left periphery of the clauses creates a situation reminiscent to that of Modern Welsh, in which Verb-to-Comp movement is blocked in the presence of clause-initial particles (Roberts : –; Rouveret : –). Accordingly, Coptic Egyptian no longer displays the Verb Second effects observable in earlier language stages. As further discussed in Reintges (: –), Verb Second in Early Egyptian (– BCE) does not apply tout court, but rather is intimately related to specific inflectional patterns (imperfectives, ‘strict’ morphological passives). When these patterns and their formal feature specifications were lost, so was long verb movement to Comp and the absolute verb-initial word orders associated with it. .. Verb movement of synthetically inflected infinitives and statives For a thoroughly analytic language, Coptic has a considerable synthetic residue of inflected infinitives and statives. These alternating stem forms are morphologically derived by associating an underlying root with a particular templatic pattern. Depending on grammatical and semantic appropriateness, consonantal roots appear in four and at most five mutually exclusive templates. The different stem alternants are formally distinguished from one another by means of stem-internal modification, in particular, by means of vowel changes, syllabification as well as partial reduplication and gemination. The ɛ ~ ɔ alternating stems of the biliteral root √KT ‘to build’ are presented in Table . together with their event semantics and argument-structure. From the perspective of major syntactic categories, inflected infinitives are reduced in their verbiness as these forms can no longer house tense and finiteness features. Massam (: –) draws a similar connection for Niuean lexical
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
T . Alternating lexical verb stems (√kt ‘to build’) Stem pattern
Form
Event semantics
Argument structure
Absolute state
kɔt
Eventive
Transitive verb plus prepositional object
Nominal state
ket
Eventive
Transitive verb plus direct object DP
Pronominal state
kot=
Eventive
Transitive verb plus object clitic
Stative
kɛt
Stative
Intransitive verb with Affected Subject
predicates, which are not morphosyntactically verbs. Due to the presence of nominal features, pro/nominal and absolute state marked stems can freely occur in various types of event nominalizations (e.g. pə-kɔt ‘the act of building’). The stative, on the other hand, has been singled out as the only surviving exemplar of a finite verb form (Reintges b: ). Accordingly, it has no corresponding state nominalization (*pə-kɛt ‘the (well-)built state’). Impoverished though the inflection of infinitives and statives may be, it still provides the relevant trigger for verb movement to the highest inflectional head, which we have proposed to identify with the Tense(Future) node. Consider in this regard the negated sentence in (), in which the main verb tənneu ‘to send’, (which appears in its nominal state form), the subject DP pə-nuːte ‘God’ and the direct object DP pe=f-ʃɛːre ‘his son’ all precede the negation adverb an ‘not’. Given that negation heads its own functional projection, as amply documented in Zanuttini (), the order Particle > Subject > Verb > Object > Negation provides strong evidence that the layered vP has been vacated by its main constituents. ()
Verb movement, subject raising and object shift in the nominal state pattern ənt a pə-nuːte gar tənneu pe=f-ʃɛːre an REL PERF DEF.M.SG-god PCL send.NOM DEF.M.SG=M.SG.POSS-child not e-pə-kosmos tʃe e=f e krine to-DEF.M.SG-world COMP REL=M.SG FUT.DEON judge.ABS əm-pə-kosmos PREP-DEF.M.SG-world ‘For God has not sent his son to the world to judge the world . . . ’ (John :)
The process of argument voiding at work here (Alexiadou and Anagnostopoulou ) shows fairly clearly that the internal and external argument cannot remain in their initial merge position but must rather undergo A-movement to the specifier position of, respectively, the lower and the higher Tense(Aspect) and Tense(Future) heads. A crucial piece of evidence for the Case-driven nature of subject raising comes from transitive-based statives, which, due to the elimination of the external (agent) theta-role, behave like adjectival passives. Accordingly, such statives coerce unaccusative movement of the internal argument to the canonical subject position. This analysis correctly predicts both the obligatorily intransitive character of stativeinflected verbs and the Subject > Stative > Negation order under negation, as seen in () (Reintges b: –, : –).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue ()
Unaccusative object-to-subject raising in the stative pattern arεu əm pə-son tɛt ən-hɛt an e-ʃatʃe perhaps NEG DEF.M.SG-brother persuade.STAT LINK-heart not to-talk.ABS nəmma=n to=PL ‘Perhaps the brother does not agree (lit. is persuaded of heart) to talk to us.’ (Apophthegmata Patrum, ed. Chaîne, no. , , )
Matters are different with regard to the absolute state pattern. In the negated transitive clause in (), the absolute state marked verb areske ‘to please’ and the subject DP pei-hɔβ ‘this thing’ move past the negation an, leaving behind internal argument pə-rɔːme ‘the man’ stranded. To avoid a Case Filter violation, the stranded object must be Oblique case-marked by the semantically vacuous locative preposition əm- (glossed as PREP, but left without translation). () Verb movement, subject raising and object stranding in the absolute state pattern eʃtʃe pei-hɔβ kʲe areske an əm-pə-rɔːme ən-ʃuːʃo if DEM.M.SG-thing PCL please.ABS not PREP-DEF.M.SG-man LINK-boast.ABS ‘If this thing does not please the man who prides himself . . . ’ (Shenoute, ed. Amélineau I. , ) Far from being peripheral to the syntax, verb movement to Tense(Future) plays a pivotal role in argument extraction. The Minimal Link Condition of early minimalism provides a straightforward way to capture this relation: phrasal extraction is strictly local unless it occurs within the domain of verb movement (Chomsky : –). ()
The synchronization of residual verb movement and phrasal extraction
TP(Past) Tense[±Past] tame particle
TP(Future)
Subject
TP(Future)
Tense [±Future] Verb Object
TP(Aspect) TP(Aspect) Tense [Aspect]
VP
Illicit V-to-Comp movement Licit residual verb movement
VP
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
In response to an anonymous reviewer’s comments, we would like to highlight the inflectionally driven nature of Verb-to-Tense(Aspect) movement. The synthetic inflection of the pro/nominal and the absolute state formations realizes both object agreement and [telic] aspect. Yet, this inflection cannot directly be related to direct object shift as this operation is restricted to the nominal state pattern. The subsequent movement step T(Aspect) to T(Future)—and this is a crucial point—is not inflectionally driven, as inflected infinitives lack subject agreement to begin with. The situation is less clear with the stative, where the remnant agreement marking may provide the relevant movement trigger. In extending the domain for subject raising, residual movement of inflected infinitives and statives represents a case of Verb-toTense movement that is not directly correlated with rich tense inflection.
. Analyticization as an emergent property in Brazilian Portuguese We shall now turn to consider the analytic restructuring of the temporal system in Brazilian Portuguese—an ongoing language change manifested in the decreasing productivity of synthetically inflected tenses and their replacement by a family of auxiliary verb constructions in the spoken vernacular. The drift towards a maximally general pattern of analyticity has consequences for the verbal syntax insofar as the space available for verb movement is severely limited. Our findings are somewhat unexpected under Biberauer and Roberts’s proposal that languages with rich tense inflection—as Brazilian Portuguese would seem at first sight—have Verb-to-Tense movement. .. The temporal system of Brazilian Portuguese and the Richness of Tense question It is by no means a trivial matter to decide whether or not a given language has ‘rich’ tense inflection in the sense of Biberauer and Roberts’s typology. Factors other than the number of paradigmatic oppositions may come into play, providing some noise in the data. Such intervening factors are the varying degrees of morphological productivity as well as the commonness and frequency of usage of a particular pattern. In a context of increased syntactic variation due to widespread language change, some patterns may be retained in the more formal registers of the language, while they have actually fallen out of usage in the spoken vernacular and must, as a result, be acquired through schooling. This is pretty much the situation with presentday Brazilian Portuguese, which is shifting from synthetic structures (which are still fully productive in European Portuguese) to the periphrastic construction of tense, aspect, and mood. All simple and compound tenses listed in Table . are presented in their first-person singular indicative active form. (The regular verb falar ‘to speak’ has been chosen to illustrate a typical paradigm.) Giorgi and Pianesi (: –) draw attention to two main particularities of the Portuguese temporal system. The first peculiarity is the complete absence of the Past Perfect *eu tive falado ‘I had spoken’—a compound tense formed with the Simple Past of the auxiliary verb ter ‘to have’ and the past participle. The non-existence of the Past Perfect is all the more surprising, as the corresponding forms exist in other
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
T . The temporal system of Brazilian Portuguese Simple tenses
Compound tenses
Present
eu falo
eu estou falando
Present + participle
Imperfect
eu falava
eu estava falando
Imperfect + participle
Future
eu falarei
eu vou falar
Present + infinitive
Pluperfect
eu falara
eu tinha falado
Imperfect + participle
Simple Past
eu falei
Future Perfect
eu terei falado
Past Perfect
*eu tive falado
Simple Past + participle
Romance languages. The second marked property of the Portuguese temporal system is the presence of a synthetically inflected Pluperfect (Mais-que-perfeito simples) eu falara ‘I had spoken’ besides the compound tense eu tinha falado. This form is historically derived from the Latin Pluperfect amāveram ‘I had loved’, with the perfective aspect morpheme -ui being lost in the process, while the past tense morpheme -er has been retained in the peculiar desinence -ra ~ -re (Huber : –, §§–). Although not entirely synonymous, simple and compound tenses are semantically close enough to enter into competition with one another, with the competition being resolved in favour of analytic tense constructions. It has long been observed (e.g. by Teyssier : §.) that the synthetic Pluperfect belongs to the literary register of standard Portuguese and has been replaced by the completely synonymous compound tense in the spoken variety. Matters are different with regard to the synthetic Future eu falarei ‘I shall speak’, which is still used in spoken Brazilian Portuguese even though it is not very common. The more productive form is the Prospective Aspect construction eu vou falar ‘I will speak’, which conveys modal connotations of intension, potentiality and hypothesis, in addition to the expression of the current relevance of a forthcoming situation (on this point, see already Campos and Rodrigues : –; Longo and Campos : –). On the face of it, Brazilian Portuguese still looks very much like a tense-rich language, closely resembling the consistent null subject Romance languages with elaborate temporal systems, such as Italian, Spanish, and European Portuguese. However, it is a well-known fact that Brazilian Portuguese, unlike European Portuguese, falls under the rubric of partial null subject languages (see, recently, Holmberg et al. ). Considering the obsolete status of synthetically inflected tenses, one should ask whether Brazilian Portuguese falls into the rubric of a tense-rich language after all. .. Tense-to-Aspect reanalysis and the loss of Verb-to-Tense1 movement Competition between simple and compound tenses is not at issue with the Simple Past, which simply lacks an analytic counterpart. However, the Simple Past is an inherently hybrid tense form as it may convey a range of non-past tense values. In this respect it
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
behaves more like a perfective aspect, as indicated by the traditional terminology pretérito perfeito ‘Preterit Perfective’ (Castilho : §....). The Simple Past is compatible with the temporal location adverbs ontem ‘yesterday’ (a) and hoje de manhã ‘this morning’ (b). It also admits future time reference (c), where the corresponding English construction would require the Present Perfect. ()
Past-perfective interpretation of the Simple Past a. Ontem eu falei com a minha família. yesterday I speak.PAST.SG with the my family ‘Yesterday I spoke with my family.’ b. Hoje de manhã eu falei com o meu colega. today at morning I speak.PAST.SG with the my colleague ‘This morning I have spoken to my colleague.’ c. Outro exame mais, e terminaste o curso. other exam more and finish.PAST.SG the course ‘One more exam, and you will have finished the course.’
As noted by Giorgi and Pianesi (: , –), Schmitt (: –) and others, the Simple Past may convey, in the appropriate contexts, the meaning of a Present Perfect—a feature that distinguishes it from its cognates in most Romance languages. The modification of the Simple Past tense form comi ‘I ate’ by the temporal adverb agora ‘now’ in (a) provides one such context, where the resultant state of a past eating event is observable at the moment of speaking. By contrast, the Present Perfect tenho comido ‘I have eaten’ in (b) is unavailable with that interpretation. It rather refers to the adopted present habit of eating sufficiently enough and thus requires more than one eating event. ()
Modification of the Simple Past and the Present Perfect by the temporal adverb agora a. Agora já comi o suficiente. now already eat.PAST.SG the enough ‘Now I have eaten enough.’ b. Agora já tenho comido o suficiente. now already have.PRES.SG eat.PTCP.M.SG the enough ‘Now I took the habit of eating enough.’
Giorgi and Pianesi (: –, , footnote ) maintain that Portuguese Simple Past is actually a Present Perfect—the same claim they put forward in relation to the Latin Perfect amāvī ‘I have loved’. However, it is important to point out—as Bertinetto and Bianchi (: ) already have—that a sentence like (a) above only shows that the Portuguese Simple Past is compatible with Present Perfect meaning; it does not prove that this tense invariably behaves in this way. Even though the semantic analysis of this form needs to be fine-tuned, this much seems to be true: despite the presence of past tense morphology, the Portuguese Simple Past is not a real past tense, but rather an aspectual category and as such a manifestation of the lower Tense head.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
The non-existence of the Past Perfect *eu tive falado ‘I had spoken’ in Brazilian Portuguese is thus predicted. The Simple Past form of the auxiliary ter ‘to have’ and the past participle cannot occur in the same periphrastic construction, as they compete for the same aspectual Tense position. This blocking effect is not operative in the temporal system of those Romance languages, in which the Simple Past has not undergone tense-toaspect reanalysis. A question remains with respect to the projection of the higher Tense node in those contexts in which the Simple Past has a Present Perfect interpretation. Giorgi and Pianesi (: ) propose that preposed temporallocation adverbslike agora ‘now’ specify a value for the reference time (R), and because of this, such adverbs must ‘enter into some relationship with a higher projection—for example, that of the auxiliary, or of T, and that dislocation plays a major role in establishing such a relationship’. Viewed from this perspective, the occurrence of the R-adverbial agora presents suggestive evidence that the relevant licensing head is contained within the functional superstructure of that clause. These and other considerations lead us to hypothesize with Sybesma () that the Tense head is universally projected and covert Tense-agreement with temporal location adverbs gives rise to different temporal interpretations. The syntactic configuration of the sample sentence Agora já comi o suficiente would look like diagram () below. ()
The syntax of the Portuguese Simple Past
TP 1 R-adverbial Agora
TP1 (Absolute Tense)
Subject eu
Tense Agreement
TP1 TP2(Aspect, Relative Tense)
Tense1 øaux Adverb já
TP2 Tense2 Verb comi
VP
VP
Illicit Verb-to-Tense1 movement
Object o suficiente Licit Verb-to-Tense2 movement
At the present stage of research, it remains an open question whether the synthetically inflected tenses of the Brazilian Portuguese temporal system pose learning
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
difficulties, with decreasing morphological productivity correlating with lower token frequency in the primary linguistic data. The Simple Past, on the other hand, has no analytic competitor but is in the process of being reanalysed as a perfective aspect. Tense-to-aspect reanalysis gives rise to mismatches between the semantic interpretation of the involved form and its hybrid morphological marking. Consequently, the Brazilian Portuguese learning child has to figure out that the past tense morphology corresponds neither to a fixed temporal value nor to the expected Verb-to-Tense movement.
. A ‘hidden’ word order change: on the comparative syntax of the adverbial sempre in European and Brazilian Portuguese We shall finally consider the distributional behaviour and interpretation of the universal time adverb sempre ‘always’ in European and Brazilian Portuguese—a topic that has received a considerable amount of attention in the literature (inter alia: Gonzaga ; Lopes ; Brito ; Ambar et al. ; Fiéis ; Cyrino ; Amaral and Del Prete ). We generally agree with the conclusion reached by Ambar et al. (: –) that the observed differences are significant and point towards an ongoing parameter change in Brazilian Portuguese, but unlike these authors, we shall propose a connection with the limited scope of verb movement in this language. .. The aspectual and modal uses of sempre in European Portuguese In European Portuguese the adverb sempre can assume two meanings, an aspectual and an epistemic-modal one. When used as an aspectual adverb, it designates the multiple occurrence of an event (pluractionality) or an event pattern as a characteristic property of the subject referent (habituality). Under both these readings, sempre universally quantifies over situations, which renders it suitable for the formulation of generic statements. ()
Aspectual interpretation of the adverb sempre (VERB > SEMPRE) a. O João está sempre em casa. [todos os momentos, todos os dias . . . ] the João be.PRES.SG always at home [‘at all moments, at all days . . . ’] ‘João is/stays always at home (i.e. in every relevant situation).’ (Gonzaga : ) b. O João estava sempre em casa. the João be.IMPERF.SG always at home ‘João was/stayed always at home.’ (Gonzaga : ) c. O João diz sempre a verdade. the João say.PRES.SG always the truth ‘João always says the truth (i.e. in every relevant situation).’ d. O João dizia sempre a verdade. the João say.IMPERF.SG always the truth ‘João always said the truth.’ (Gonzaga : )
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
Under the confirmative reading, sempre asserts the persistence of the truth of a particular proposition, which has been under discussion by the interlocutors and is therefore part of the common ground shared by them. In this function, it comes very close to a modal adverb that universally quantifies over epistemic worlds (Lopes ; Amaral and Del Prete ). ()
Confirmative interpretation of the adverb sempre (SEMPRE > VERB) a. O João sempre está em casa. [afinal ele está em casa] the João always be.PRES.SG at home [‘he is at home after all’] ‘João is after all/indeed at home.’ (Gonzaga : ) b. O João sempre estava em casa. the João always be.IMPERF.SG. at home ‘João was after all/indeed at home.’ (Gonzaga : ) c. O João sempre diz a verdade. the João always say.PRES.SG the truth ‘João indeed says the truth.’ (Gonzaga : ) d. O João sempre dizia a verdade. the João always say.IMPERF.SG the truth ‘João indeed/after all said the truth.’ (Gonzaga : )
The point of interest here is the three-way correlation that exists between positioning of sempre vis-à-vis the associated verb, the morphological tense of that verb and the resulting aspectual or modal interpretation. As pointed out by Gonzaga (: –), Ambar et al. (: –), and Fiéis (: –), the adverb sempre is placed in post-verbal position under the aspectual and in pre-verbal position under the epistemic-modal reading. The picture is further complicated by the fact that the preverbal Sempre > Verb order may have both an aspectual interpretation and an assertive-confirmative reading in Simple Past tense sentences—an option that is no longer available in Present and Imperfective Past tense sentences. ()
Confirmative and aspectual interpretation of pre-verbal sempre in Past Tense contexts a. O João sempre foi a Paris de comboio. the João always go.PAST.SG to Paris by train ‘João went to Paris by train after all.’ (√confirmative reading) ‘João always went to Paris by train.’ (√aspectual reading) b. */?? O João sempre {vai / ia} a Paris de comboio. the João always {go.PRES.SG / go.IMPERF.SG} to Paris by train ‘João {goes / went} to Paris by train after all.’ (√confirmative reading) ‘João always {goes / went} to Paris by train.’ (*aspectual reading) (Ambar et al. : –)
Temporal location adverbs such as ontem ‘yesterday’ can appear in different positions, but may not disrupt the string adjacency between pre-verbal sempre and the verb modified by it. In this connection, Ambar et al. (: ) present the following paradigm.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
()
Chris H. Reintges and Sonia Cyrino The adjacency requirement between pre-verbal sempre and the associated verb a. *Ele sempre ontem foi a Paris. he always yesterday go.PAST.SG to Paris b. √ Ontem ele sempre yesterday he always
foi go.PAST.SG
a Paris. to Paris
c. √ Ele sempre foi ontem a Paris. he always go.PAST.SG yesterday to Paris d. √ Ele sempre foi a Paris ontem. he always go.PAST.SG to Paris yesterday ‘He went to Paris after all.’ As first noted by Gonzaga (: ) only clitic pronouns may intervene between pre-verbal sempre and the associated verb. ()
Intervening pronominal clitics Ele sempre lhe disse isso. he always him.CL.DAT.SG say.PAST.SG this ‘He indeed said this to him.’
The string adjacency requirement just outlined is a good indication that pre-verbal sempre and the verb are both contained within the same left-peripheral projection, which Ambar et al. (: ) analyse as an AssertiveP and which we propose to identify with the ModepistemicP of the Cinquean (: ) hierarchy. To be sure, this is not just a terminological question. Our analysis differs from theirs in that we localize Rizzi’s () INT(errogative)P on top of the ModepistemicP rather than the other way around. To lend credence to this claim, we shall briefly consider whquestions containing an instance of pre-verbal sempre in its canonical modal use. Consider the grammaticality contrast between the declarative and the corresponding interrogative sentence in (a–b). ()
Declarative-interrogative sentence pair containing pre-verbal sempre a. O João sempre vai a Paris. the João always go.PRES.SG to Paris ‘João goes to Paris after all.’ b. *Aonde vai o João sempre? where go.PRES.SG the João always ‘Where does João go to after all?’ (Ambar et al. : )
The interrogative sentence in (b) involves the movement of the locative wh-adverb aonde ‘where to’ to the specifier of the INTP and residual Verb Second movement to its head. Since the verb moves independently of the adverbial sempre, the two constituents are no longer adjacent to each other, thereby producing an ungrammatical output. This is in essence Ambar et al.’s (: ) analysis. Accordingly, the syntactic configuration underlying pre-verbal sempre in its confirmative reading must be derived in a two-step procedure: the first derivational step involves the
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
preposing of the adverbial sempre, which moves from its VP-adjoined base position to an adjoined position of the specifier of the ModepistemicP. By hypothesis, this quantifier raising operation is the syntactic correlate of the shift in perspective from quantification over situations to quantification over modal worlds. The second derivational step involves verb movement all the way up to the Modepistemic head. ()
The syntax of preverbal sempre in its modal-confirmative use
Int(errogative)P Aonde
IntP
Int(errogative)
ModepistemicP
sempre
ModepistemicP Modepistemic Verb vai
TP1 (Absolute Tense) TP1
Subject o João
Tense1
TP2(Aspect, Relative Tense) sempre
Quantifier Raising
TP2 Tense2
VP
Wh-movement When formulated in terms of Rizzi’s (, ) criterial conditions, the structure in () instantiates the relevant criterial configuration for the wh-movement of aonde as well as for the quantifier raising of the adverb sempre, while the association of the verb with an unvalued modal feature triggers the observed subject–verb inversion. The next question is why the inverted verb stays within the ModepistemicP projection, without moving further up to the next higher INT head in order to satisfy the whcriterion. It seems to us that the final movement step is excluded by the Principle of Criterial Freezing, according to which criterial positions terminate movement chains: an element meeting a given criterion is frozen in place, and its chain cannot be extended any further.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
.. The exclusively aspectual use of sempre in Brazilian Portuguese All three naturally occurring examples in (a–c), taken from the corpus of the Gramática do Português Falado (‘Grammar of Spoken Portuguese’), exemplify the exclusively aspectual reading of the adverb sempre in contemporary Brazilian Portuguese. Further note that this reading is obtained with both pre-verbal and post-verbal order. ()
Consistent aspectual interpretation of the adverb sempre a. A gente se encontra sempre todos os meses the people REFL meet.PRES.SG always all the months nesse jantar. (VERB > SEMPRE) at.this dinner ‘We (lit. the people) always meet each other every month at this dinner party.’ (POA–: , Ilari : ) b. Vinha muito pianista estrangeiro, então a gente come.IMPERF.SG much pianist foreign then the people não perdia, nós estávamos sempre lá. (VERB > SEMPRE) NEG lose.IMPERF.SG we be.IMPERF.SG always there ‘Many foreign pianists came, so we [lit. the people] didn’t lose the opportunity and we were always there.’ (POA–: , Ilari : ) c. Eu não leio muito negócio de esporte, eu sempre I NEG read.PRES.SG much matter of sport I always viro as folhas. (SEMPRE > VERB) turn.PRES.SG the pages ‘I don’t read much news about sports, I always turn the pages (of the sports section).’ (POA–: , Ilari : )
The systematic absence of the confirmative reading is evidence that the semantic contrast between the still available pre-verbal and post-verbal positioning of sempre has been neutralized in Brazilian Portuguese. In the reminder of this section, we shall first offer a syntactic explanation for this fact and then comment on the markedness of post-verbal Verb > Sempre order. It is a well-established fact that Brazilian Portuguese has lost subject–verb inversion (see Lopes Rossi for the details of the historical process and the statistical evidence). As a result, the subject may intervene between fronted wh-item and the verb, yielding Wh-adverb > Subject > Verb > Sempre order. ()
No adjacency requirement between the fronted wh-item and the verb a. Aonde o João vai sempre de carro? where the João go.PRES.SG always by car
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
b. *Aonde vai o João sempre de carro? where go.PRES.SG the João always by car c. *Aonde vai sempre o João de carro? where go.PRES.SG always the João by car ‘Where does João always go to by car?’ (Ambar et al. : ) Unlike to what happens in European Portuguese, the finite verb inflection in Brazilian Portuguese cannot be specified with an unvalued wh-feature, which drives residual Verb Second (see Mioto and Kato for further details). To this one might add that none of the simple past tense forms can be endowed with a criterial feature related to a left peripheral head. The impossibility of creating the relevant specifier-head configuration for hosting the quantifier-raised adverb sempre and the verb modified by it follows. This, we believe, is another aspect of the decreasing productivity of simple tenses. Ambar et al. (: –) argue that the pre-verbal Verb > Sempre order rise to the habitual reading, whereas the reverse Sempre > Verb order conveys a correlation of event reading. We checked the following sentence pair for several speakers from different regions of Brazil, in which a temporal clause introduced with quando ‘when’ provides the appropriate context for selecting the Imperfective Past in the following main clause. ()
Correlation of event interpretation with pre- and post-verbal ordering of sempre a. Quando o João era criança, ele sempre ficava when the João be-IMPERF.SG child he always stay.IMPERF.SG em casa. (SEMPRE > VERB) at home b. Quando o João era criança, ele ficava sempre when the João be.IMPERF.SG child he stay.IMPERF.SG always em casa. (VERB > SEMPRE) at home ‘When João was a child, he always stayed at home.’
The picture that emerges is one of considerable variation. For one group of speakers the two orders can be used interchangeably without any detectable difference in meaning, while another group of speakers finds the pre-verbal order slightly more natural in an out-of-the-blue context. Yet, a third group of speakers accepts the post-verbal order as a marked alternative—a judgement also reported in the literature (Ambar et al. : ; Fiéis : footnote ; Cyrino : ). Here we propose to correlate the variation in regard to the pre- and post-verbal order to the availability of Verb-to-Tense movement in the grammar of individual speakers.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Chris H. Reintges and Sonia Cyrino
()
The syntax of preverbal sempre in its modal-confirmative use
Int(errogative)P Aonde
IntP
Int(errogative)
ModepistemicP
sempre
ModepistemicP Modepistemic Verb vai
TP1 (Absolute Tense) TP1
Subject o João
Tense1
TP2(Aspect, Relative Tense) sempre Tense2
Quantifier Raising Illicit Residual Verb Second
TP2
Marked V-to-T1 movement
VP
Wh-movement Delimiting the scope of verb fronting in the course of analytic drift proceeds in a topdown fashion, eliminating first verb movement to the clausal left periphery and then from there Verb-to-Tense movement. The former process has been completed while the latter is still underway.
. Concluding remarks Historical grammar change can be described in formal terms as the relation between one grammatical system and another, with the structural differences between the two systems being derivable from the resetting of at least one parametric value. Naturally, under this approach, the focus of research is on the change itself, while less attention is paid to linguistic continuity. In this comparative study, we have looked at analytic drift in two diachronic settings—one in which the change has only partially been completed and another one in which it is under way. In both these contexts, analyticization is an emergent grammatical property, which leaves parts of the anterior synthetic system intact. The relic syntax of synthetically inflected verb
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Analyticization and the syntax of the synthetic residue
forms preserves verb movement options, atypical of thoroughly analyticized systems. Furthermore, such forms may develop into hybrid forms, in which the original morphology no longer correlates both with a fixed temporal value and with a particular movement path. The loss of verb movement may also have effects in other domains such as adverbial syntax, as shown by our comparative study on the ‘hidden’ word order change of the temporal adverb sempre in Brazilian Portuguese. The diachronic facts reviewed here suggest a modification of Biberauer and Roberts’s () model for deriving verb movement options. First, to account for the Brazilian Portuguese case, residual movement must be allowed to target the next higher functional head but extending no further. In this way, it covers the same configuration space as nonfinite participles and infinitives. Second, to accommodate the Coptic case, the model must incorporate a notion of morphological complexity different from that of paradigmatically organized tense inflection. Synthetic inflection is related to the event-state contrast, a/telic aspect and object agreement. Yet it drives the movement of infinitive and statives to the highest temporal head of the inflectional domain, thereby extending the configurational space required for subject raising.
Sources Acts: Thompson, Herbert. . The Coptic Version of the Acts of the Apostles, and the Pauline Epistles in the Sahidic Dialect. Cambridge: Cambridge University Press. Proverbs: Worrel, William H. . The Proverbs of Solomon in Sahidic Coptic: According to the Chicago Manuscript. The University of Chicago Oriental Institute Publications . Chicago, Illinois: The University of Chicago Press. John: Horner, George W. . The Coptic Version of the New Testament in the Southern Dialect Otherwise Called Sahidic and Thebaic. vols. Vol. III: The Gospel of S. John. Register of Fragments etc. Facsimiles. Oxford: Clarendon Press. Apophthegmata Patrum, ed. Chaîne: Chaîne, Marius. . Le manuscrit de la version copte en dialecte sahidique de ‘Apophthegmata Patrum’ (Bibliothèque d’études coptes, ). Cairo: Imprimerie de l’Institut Français d’Archéologie Orientale. Shenoute, ed. Amélineau: Amélineau, Emile. . Œuvres de Schenoudi. Texte copte et traduction française. vol. Vol. I. Paris : Ernest Leroux.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
11 Loss of laten-support in embedded infinitivals in fifteenth-century Low Saxon G E R T J A N P OS TM A
. Introduction The rise of do-support in the history of English has a prominent place in diachronic studies (Ellegård ; Visser ; Denison ; Kroch ). In Kroch’s study, dosupport is one of the four diachronic changes that illustrate his Constant Rate Hypothesis (CRH). The CRH allows us to identify two diachronic processes with distinct actuation times but equal rates, as being realizations of one underlying parametric change. Only if two changes display equal process rates in the E-language, can they be related to one process in the I-language. In the case of do-support, the CRH allows Kroch to assume that do-support in questions (V-to-C) and in negative declaratives (V-to-T), as one and the same grammatical process, despite the fact that the change occurs later in negative declaratives. Kroch links both changes to a shared change in the properties of VP, which has become opaque for verbal extraction (Roberts ; Pollock ). In a phase-type analysis, we can tie the two processes to properties of the vP-edge. It crucially depends on the status of vP as a phase. If, on the other hand, the two processes can be shown to be distinct, the two types of do-support are probably tied to changes in some intermediate projection or to the landing site, CP or TP, respectively. In this chapter we study a periphrastic construction (‘do-support’) in a fifteenthcentury Low Saxon dialect in the Netherlands. The periphrasis shows up quite differently because of the basic OV nature and the strict V character of the dialect. It enables us to view the locus of verb movement from a different perspective. This fifteenth-century Low Saxon variant exhibits a change in the verbal system that is similar to what happens in the history of English but its direction is reverse: Low Saxon dialects in the Netherlands lost auxiliary support and lost a separate auxiliary class. It turns out that not so much changes in VP/vP but changes in the properties of TP and CP are at the basis of changes in laten-support. Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Gertjan Postma . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Loss of laten-support in embedded infinitivals in Low Saxon
. Infinitival verb fronting Coastal continental Germanic has a scarcely documented and poorly understood construction where an embedded clause-final infinitival verb undergoes fronting (henceforth: ‘Infinitival V or IV’) to a clause initial position.1 It is a subordinate construction where the subordinate clause is introduced by a coordinator rather than a subordinator, of the type try-and-do in modern English, illustrated in (a). In Continental Germanic, the coordinative construction shows the second conjunct in VO order, whereas the first conjunct shows the standard OV order, as represented in (b). () a. CP
b. &P
SU AUX
VP1 try
&P
SU &’ VP2
John will
CP
& help Peter and
C
&’
VP1
VP2
John will
O + Vinf
&
Vinf + O
and
This curious construction can be found in Old Frisian (), Middle Dutch (), and some modern coastal Dutch dialects still had it in the nineteenth century (). It is not found in Standard modern Dutch. In the next examples, we list the main verb in bold and its complement in italics. Conjuncts after ende ‘and’ systematically show VO order. () Soe scellet hya dae nesta riocht dei deer ney weer [toe riocht so shall they the next law day there after again to court gaen ende nyment als nw scold wessen habba] go.INF and take.INF.it as now should be have ‘Hence they shall go to court the next law day after it and (they shall) take it as if it now had taken place.’ (Old Frisian. Oosterhout , apud De Haan ) () Ende men sal [al die vorseide dingen pulveren] ende and one shall all those before.mentioned things grind.INF and [mingen metten zape van ruten] ende [maken pillen] mix.INF with soap from rue and make.INF pills ‘And one shall grind all the above ingredients and mix them with ruta [a plant: ruta graveolens] soap and make pills.’ (Middle Dutch. Jan Yperman, Cyrurgie, apud Overdiep )
1
Our IV-terminology is theoretically driven and presumes that the moved infinitival occupies a headposition. It shows up as a V-effect observationally, just as we see in imperatives, yes/no questions, conditionals, etc., but we assume that the target is the structural second position in the clause (C).
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Gertjan Postma
() We zelle [te waa gaen] en make de vleet op order we shall to net go.INF and make.INF the drift.net in order ‘We shall go to the nets and make the drift nets in order.’ (Dutch dialect of Katwijk, apud Overdiep ) Notice that the clitic object ’t ‘it’ appears to the right of the infinitive nymen ‘take’ in the Frisian construction in (), despite the fact that standard embedded clauses realize OV order. In these dialects, clitic objects scramble beyond the position of the full NP subject. In (), the infinitival verb nymen ‘take’ appears to the left of this scrambled position. The infinitive, therefore, must move quite high up in the syntactic tree. Similar considerations hold for Middle Dutch and dialectal Dutch.2
. Infinitival fronting in fifteenth-century Low Saxon Recently, a corpus has been made available of North-Eastern Dutch fifteenth-century verdicts.3 Interestingly, the IV construction occurs abundantly, probably because of the modal nature of most verdicts. In ()–() two typical instances are listed. () Item tusschen den Konen ende den smyt to Legele, wyl de Kone item between the Krones and the smith in Legele: will the Krone erffnysse hebn, so moet hy en rechte ansetten ende laten hem heritage have, then must he a lawsuit start and let-INF him vertugen. (Etstoel, verdict ) defend.through.witnesses ‘Similarly between the Krone family and the smith in Legele: if Krone wants to inherit, he must start a lawsuit and defend himself through witnesses.’ () Mach Claes van synen magen, [ . . . ], an syn hant nemen ende may Claes of his family [ . . . ] in his had take and maken hem der saeke onsculdich, dair hy voir angesproken was, make him in.the case unguilty, which he for called was, so . . . (Etstoel, verdict ) then . . . ‘If Claes can take in his hand twelve of his family members and (if he can) prove himself not guilty in the case that he was called for, then . . . ’ Semantically, the modal verb, moet ‘must’ in () and mach ‘may’ in (), has scope over both conjuncts. The coordinative structure of these verdicts can be analysed as VP coordinations in the scope of this auxiliary (a) or, alternatively, as CPcoordinations (b).
2 De Haan ( []: ) mentions the construction in passing in his study of the imperativus pro infinitivo (IpI) construction in Modern Frisian. 3 First published in Keverling Buysman (). For a more detailed description in English, cf. Postma (). Cf. .
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Loss of laten-support in embedded infinitivals in Low Saxon ()
a. XP AUX SU [VP and VP] b. XP [[CP AUX SU VP] and [CP AUX pro VP]]
VP=OV, VP=VO VP=OV, VP=VO
If we analyse them as VP coordinations, the scope issue is straightforward, but we must assume two different base-generated VPs or we must assume verb fronting to the edge of the second VP. Both assumptions would be ad hoc. If we analyse them as CP-coordinations in (b), we must assume an empty modal in the second conjunct. The fronting can be identified as V-to-C. Both assumption are not ad hoc. The first ingredient is assumed in Kayne () for imperative constructions, which share the MUST modality of these verdicts. Second, such embedded AUX to COMP constructions has been reported independently in Romance (Rizzi ) and Germanic (Hoeksema ). We only have to assume that a fronted empty AUX, the empty modal, carries the main verb with it, i.e. the lexical verb must have incorporated into the empty AUX. For a full justification of the V-to-C analysis of these verdicts, I refer to Postma (a,b). In this context I would like to add an extra empirical argument, drawn from the infinitive-clitic order, as observed in (). In (b) we see that the scrambled object clitic t is higher than the lexical subject oir oem. If we assume a low position of this subject, say in specVP, the object clitic must be above VP. This implies that the second conjunct is certainly larger than VP. () a. dye sal en recht ansetten ende wynnen-t myt enen buyrtuge he.TOP shall a lawsuit start and win.INF-it with neighbor witness ende overluden and authorities (Etstoel, verdict ) ‘He shall start a court and gain it with witnesses and authorities.’ b. so mach-t oir oem nyet wedder sprecken (Etstoel, verdict ) so may-it her uncle not against speak ‘therefore, her uncle may not deny it’ We will not make a decision whether the embedded constituent is a projection of T or C. We will simply opt for the CP implementation. We found more than instances of this otherwise rare construction. Because of its frequency in this juridical corpus (>, words), the construction can be studied diachronically with quantitative means. In Table . we give the breakdown over the fifteenth century. In Figure . we have plotted the global values in a graphic and fitted it with a straight line. The incidence remains relatively constant over this century. If anything changes here at all, there is a slight increase (slope = .e- .e-/words.year). For our purposes, we may consider the incidence of IV constant over this century.
. Laten-support In the previous section, we discussed a movement of the infinitive to a clause-initial position (past the clitic position) after the coordinator ende ‘and’. It can be observed over a period of years. This movement was not without cost. There is evidence
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Gertjan Postma
T . Breakdown of occurrence of IV in fifteenth-century Drenthe Year
IV
#Words
IV per w
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2.0
Rate
1.5 IV2 per 1000w 1.0
Fit to straight line
0.5
0.0 1400
1420
1440
1460
1480
1500
Year
F . Occurrence of IV in fifteenth-century Drenthe per , words
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Loss of laten-support in embedded infinitivals in Low Saxon
that during this century, there is a style variant in which the infinitival verb could not move. Instead, a dummy verb laten is inserted and undergoes the fronting. An example is given in (a). ()
[It is sentenced by the judges . . . ] a. . . . , dat Herman een rochte ansetten sal [ . . . ] ende laten die overlude . . . that Herman a lawsuit start shall [ . . . ] and let.INF the authorities laeden ende wes hem die overluede myt den bueren invite, and whatever him.DAT the authorities with the neighbours toe tugen, dair sall mede hene gaen (Etstoel, verdict ) PRT sentence.INF, there shall (it) with PRT go ‘ . . . that Herman must start a lawsuit and must invite the authorities, and what the authorities with the neighbours sentence, that way it will proceed.’ b. . . . , dat Johan Staels een rochte ansetten sall ende laden de . . . that Johan Staels a lawsuit start.INF shall and invite the bruytluden ende wes hem die bruytlude toe tugen, dair bride.people and what him.DAT the bride.people PRT witness.PL, there salt mede henne gaen. (Etstoel, verdict ) shall (it) with PRT go ‘ . . . that Johan Staals must start a lawsuit and must invite the bride’s family and what the bride’s family decides for him, that way it will proceed.’
The causative/permissive verb laten ‘let/make’ does not contribute a lexical meaning in this construction with main verb laeden ‘invite’. In (b) we give a parallel verdict without auxiliary: the lexical verb laeden fronts itself. The two verdicts are unfortunately not two versions of the same verdict but they are clearly parallel. A first piece of evidence that laten is not an ordinary lexical verb in the IV construction is the fact that the incidence of laten in this construction is significantly higher (%) than the overall incidence of laten in all contexts (%) justifies. The cross table is given in Table .. The deviance from the expectation value is highly significant (p Aux > VP b. VP > Aux > Neg c. Neg > VP > Aux
base order (no roll-up) full roll-up partial roll-up
I will take it that the presence of ‘V-Neg-Aux’ sequences (which also occur with compound tenses involving a BE-auxiliary) in the PLD present the language acquirer with sufficient evidence not to postulate a roll-up derivation. Instead, what seems to be at work is a less local type of movement, which targets a functional projection which is higher than Neg and T. However, it has to be lower than C, as left peripheral XPs in Latin surface to the left rather than to the right of subordinating conjunctions (Danckaert ). Adopting a parameterized approach to EPP-checking (as initiated by Alexiadou and Anagnostopoulou (), and further elaborated in Biberauer () and Biberauer and Roberts ()), I will assume that the VP undergoes EPP-driven A-movement to a position in the high middle field, a hypothesis that was also put
13 Coordination facts (as in (i), which is most plausibly analysed as involving two coordinated headinitial verb phrases, simply labelled as ‘VP’ here) confirm that the correct bracketing of VOAux-clauses is as in (ii), with the string VO forming a constituent to the exclusion of the auxiliary.
(i)
quicquam] [&° aut [VP ne qui-s [&P [VP occul-ere lest anybody-NOM hide-PRS.INF something.ACC or segrega-re caus-am su-am]]] pos-si-t be.able-PRS.SBJV-SG separate-PRS.INF cause-ACC his-ACC ‘so that nobody may be able to hide anything or put aside any of his property’ (= Tac. Hist. .)
(ii) [FP [VP V DPOBJ]i [TP [T Aux ti ]]] 0
See Danckaert (: –) for additional discussion of the constituency of VOAux-clauses.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Lieven Danckaert
forward in Danckaert (, ) and Mackenzie and Van der Wurff ().14 I will simply call the target site of this operation ‘FP’, a functional projection where the clausal EPP-requirement is encoded, and which is distinct from Cardinaletti’s () SubjP (which I take to host subject DPs that act as the ‘subject of predication’). As indicated in the trees in (), I assume the relevant head to come with the feature specification [+V, +EPP^], which results in A-movement to SpecFP. One can understand this operation as involving pied-piping, where it is actually (the φ-features of) the highest VP-internal argument DP which is probed for. For reasons of space, I cannot here provide further evidence of this particular analysis. Suffice it to say that there is reason to assume that Classical Latin VPAux-clauses (i) are derived by means of VP movement (cf. absence of HMC-effects when non-finite verbs cross negation) and (ii) that this movement operation is not strictly local. The reader is referred to Danckaert (, ) for additional evidence in favour of the claim that Classical Latin employs VP movement to satisfy the EPP. A summary of the analysis thus far is provided in the trees in (), which show a VPAux-clause in which roll-up movement has applied inside the displaced VP (yielding the order ‘SOV-Neg-Aux’) (a), and one in which the moved VP is head-initial (b), which corresponds to the surface order ‘SVO-Neg-Aux’. Circled phrases have undergone roll-up movement, boxed phrases have been A-moved. For the sake of simplicity, I assume that the modal sits in T°, and that it simply takes an extended verb phrase (VoiceP) as its complement: () a.
FP F’
VoiceP DPea
Voice’
F° [+V,+EPP^] vP
Voice° [+V]
Neg° [+V]
DPia √°
14
T°[+V]
vP
√P t√°
NegP
v° [+V^]
TP tVoiceP
t√p
v°
On the possibility for VPs to satisfy the EPP-requirement, see also Travis ().
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
The decline of Latin VOAux
b.
FP F’
VoiceP DPea
Voice’ Voice° [+V]
F° [+V,+EPP^] vP
v° [+V] √°
v°
Neg° [+V] T° [+V]
√P t√°
NegP TP tVoiceP
DPia
The tree in (b) now allows us to see why—if the analysis pursued here is on the right track—Classical Latin allows for the linear order VOAux: given that VP displacement is triggered by a feature associated with an EPP-feature and not with a categorial feature, the relevant operation is not subject to FOFC. As we will see in the following section, corpus data show that the VOAux pattern is only very rarely attested in Late Latin, suggesting that the grammar of this language is crucially different from the one of Classical Latin.
. The diachrony of VOAux .. A corpus study The texts that I have investigated come from a period of six centuries, which naturally falls apart into two subperiods, separated by a period from which fewer (especially non-Christian) texts are preserved. We can distinguish an earlier period from c. BC (taken as an average date of composition of Cicero’s writings) until c. AD (Gaius) and a later period from c. (Palladius) until AD (Gregory of Tours). A full description of the corpus is given in Table .. As indicated in the last column, the consulted text editions were either taken from the annotated database Hyperbase (Brunet and Mellet n.d.) or from the non-annotated online database Brepolis (www.brepolis.net). From this corpus, , clauses were collected, all containing: (i) a form of the modals possum ‘be able’ or debeo ‘have to’ (cf. section ..); (ii) a transitive infinitive (V); and (iii) an NP (DP) direct object (O). I included objects with structural (accusative) as well as inherent (genitive, dative, or ablative) case. The quantitative results are summarized in Tables . and ..
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Lieven Danckaert
T . Description of a Latin prose corpus Author (work(s))
Date (approx.)
Source
.
Cicero (selection of speeches)
BC
Hyperbase
.
Caesar (De bello ciuili, De bello Gallico –)
BC
Hyperbase
.
Varro (Res rusticae; De lingua Latina)
BC
Brepolis.net
.
Hyginus (Astronomica)
BC
Brepolis.net
.
Vitruvius (De architectura)
AD
Brepolis.net
.
Livy (Ab urbe condita)
AD
Brepolis.net
.
Seneca (Epistulae, De clementia, De beneficiis, Dialogi, Apocolocyntosis)
AD
Hyperbase
.
Petronius (Satyrica)
AD
Hyperbase
.
Frontinus (Strategemata, De aquaeductu urbis Romae)
AD
Brepolis.net
. Tacitus (Historiae, Annales, opera minora)
AD
Brepolis.net
. Gaius (Institutiones)
AD
Brepolis.net
. Palladius (De ueterinaria medicina, De agricultura)
AD
Brepolis.net
. Augustine (Sermones –, –A, –E, –A)
AD
Brepolis.net
. Gesta Conlationis Carthaginiensis
AD
Brepolis.net
. Vegetius (Epitoma rei militaris, Mulomedicina)
AD
Brepolis.net
. Pompeius Maurus (Commentum artis Donati)
AD
Brepolis.net
. Caesarius of Arles (Sermones –)
AD
Brepolis.net
. Gregory of Tours (Historia Francorum)
AD
Brepolis.net
c.– AD: few attestations.
.. Discussion The diachronic evolution of the VOAux pattern is diagrammed in Figure .. The data taken into account here are only a subset of the corpus reported on in Tables . and ., namely VPAux-clauses with a VP-internal object (i.e. all AuxVP-clauses as well as VAuxO-orders (i.e. VPAux-clauses with an extraposed object) were excluded). What is plotted is the relative frequency of head-initial VPs in VPAux-clauses (as compared to the incidence of head-final VPs). As shown by the smoothed regression line, we observe an initial rise of head-initial VPs in VPAux-clauses, followed by a fairly abrupt decline of this pattern, apparently starting in the second half of the first century AD. When we compare the average frequency of the order VOAux in the periods before and after the time gap in the corpus, it appears that the order VOAux is attested in . per cent of the VPAux-clauses in the earlier period, as compared to . per cent in the later period. This difference is statistically significant (independent samples T-test, equal variances assumed after logarithmic transformation with added constant (+), p=.).
T . Word order in Latin clauses with a modal auxiliary (absolute figures) Date
Author
AuxVP
VPAux
AuxVO
AuxOV
VAuxO
VOAux
OAuxV
OVAux
OV
VO
BC
Cicero
BC
Caesar
BC
Varro
BC
Hyginus
Vitruvius
Livy
AD
Seneca
AD
Petronius
AD
Frontinus
AD
Tacitus
AD
Gaius
c.– AD: few attestations. AD AD
Palladius Augustine
AD
Gesta Conl. Carth.
AD
Vegetius
AD
Pompeius Maurus
AD
Caesarius of Arles
AD
Gregory of Tours
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
AD AD
Date
Author
AuxVP
VPAux
AuxVO
AuxOV
VAuxO
OAuxV
VOAux
OVAux
VO
OV
.
.
.
.
BC
Cicero
.
.
.
.
BC
Caesar
.
.
.
.
.
.
.
.
.
.
.
BC
Varro
.
.
.
.
.
.
.
.
.
.
BC
Hyginus
.
.
.
.
.
.
.
.
AD
Vitruvius
.
.
.
.
.
.
.
.
.
.
AD
Livy
.
.
.
.
.
.
.
.
.
.
AD
Seneca
.
.
.
.
.
.
.
.
.
.
AD
Petronius
.
.
.
.
.
.
.
.
.
AD
Frontinus
.
.
.
.
AD
Tacitus
.
.
.
.
.
AD
Gaius
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
c.– AD: few attestations. AD
Palladius
.
.
.
.
.
.
.
.
.
.
AD
Augustine
.
.
.
.
.
.
.
.
.
.
AD
Gesta Conl. Carth.
.
.
.
.
.
.
.
.
.
.
AD
Vegetius
.
.
.
.
.
.
.
.
.
AD
Pompeius Maurus
.
.
.
.
.
.
.
AD
Caesarius of Arles
.
.
.
.
.
.
.
.
.
.
AD
Gregory of Tours
.
.
.
.
.
.
.
.
.
.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
T . Word order in Latin clauses with a modal auxiliary (percentages)
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
The decline of Latin VOAux
100
% of VP-internal VO
80
60
10
40 6
7
8 9
20 1 3 2 0
5
12
11
13 14 17 15
4 0
100
200
300
400
18
16 500
600
Year
F . Frequency of VO in VPAux-clauses with a modal auxiliary, c. BC– AD. Case labels: = Cicero, = Caesar, = Varro, = Hyginus, = Vitruvius, = Livy, = Seneca, = Petronius, = Frontinus, = Tacitus, = Gaius, = Palladius, = Augustine, = Gesta Conlationis Carthaginiensis, = Vegetius, = Pompeius Maurus, = Caesarius of Arles, = Gregory of Tours
Before we conclude that this difference is indeed real, let me point out that the observed decline is not an accidental by-product of the decline of one of the ingredients of VOAux, namely, the order VPAux. As can be deduced from Tables . and ., VPAux is less frequently attested in the later period (. per cent in the earlier vs . per cent in the later period; this difference is statistically significant (independent samples T-test, equal variances assumed, p=.)).15
15 Note that, contrary to what is often claimed, the overall frequency of the order VO (i.e. the second ingredient of the VOAux pattern) remains almost perfectly constant through time, at . per cent in the early period, and . per cent in the later (values averaged over VPAux and AuxVP contexts). As expected, the difference between these two averages cannot be shown to be statistically significant (independent samples T-test, equal variances assumed, p=.). See Danckaert () for more detailed discussion.
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Lieven Danckaert
T . Observed vs. estimated frequency of VOAux (VP-internal objects only: extraposed (VAuxO) and scrambled (OVAux) objects excluded) Author (text)
Date (approx.)
Observed VOAux (frequency in %)
Predicted VOAux (frequency in %)
Prediction error
Cicero
BC
Caesar Varro Hyginus
BC
Vitruvius
AD
.
.
.
AD
.
.
–.
AD
.
.
–.
Livy Seneca
.
.
BC
.
.
.
BC
.
.
.
–.
Petronius
AD
.
.
–.
Frontinus
AD
.
.
–.
Tacitus
AD
.
.
–.
Gaius
AD
.
.
–.
c.– AD: few attestations. Palladius
AD
.
.
–.
Augustine
AD
.
.
.
Gesta Conl. Carth.
AD
.
.
.
Vegetius
AD
.
.
Pompeius Maurus
AD
.
.
Caesarius of Arles
AD
.
.
.
Gregory of Tours
AD
.
.
.
If we assume that ‘VOAux = VPAux + VO’, we predict the combined rates of VO and VPAux to provide an accurate estimate of the frequency of VOAux. The predicted and the observed frequencies of VOAux are given in Table ..16 The rightmost column shows the ‘prediction error’, i.e. the difference between the observed and the predicted values: the closer this prediction error to zero, the more accurate the estimate. Two interesting conclusions can be drawn from these figures. First, the prediction error for the six last data points (cf. the shaded cells in Table .) is consistently positive, implying that the actually observed values are lower than what we expect if the observed decline were only an effect of the reduced frequency of VPAux-clauses, assuming everything else to remain the same. We can conclude that the decline of VOAux is independent of the decline of VPAux. Second, the range of the variation is much wider in the first period (the difference between the 16
The predicted rate of VOAux is the product of the frequencies of VO and VPAux given in Table ..
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
The decline of Latin VOAux
lowest and highest error being . units in the early period, compared to . in the later). This suggests that word order variation in Classical Latin is not solely governed by syntax. Rather, I take it that the grammar could generate a number of orders, and that the eventual choice between those options was to a large extent governed by usage-based factors. In Late Latin on the other hand, such usage-based factors were to a large extent outranked by syntactic constraints, with a reduction of word order variability (and lower prediction errors) as a result. In what follows, I will propose that one such syntactic constraint is FOFC, which in Late but not in Classical Latin restricts word order in pre-Aux VPs. .. A hypothesis In section .. it was argued that in Classical Latin, negated clauses with an auxiliary and a non-finite lexical verb (like () above), which typically feature the order ‘VP-Neg-Aux’, provide the language learner with evidence that VPAux-clauses do not involve roll-up movement, as the negator non ‘not’ intervenes between the right boundary of the displaced VP (VoiceP) and the auxiliary. Moreover, the interaction with verb movement suggests that in Classical Latin, the pre-verbal negator is a syntactic head. The syntax that was proposed for this stage of the language correctly predicts the word order VOAux to be available. The hypothesis that I will explore is that during the transition from Classical to Late Latin, non ceases to count as a ‘boundary’ between a moved VP and an auxiliary. I will suggest that non incorporates into the verbal head in T through a process of procliticization, thus forming a complex head ‘Neg-T’ (‘Neg-Aux’). Once this evolution has taken place, the language learner is no longer confronted with any strong indications that a VPAux-clause does not involve roll-up movement. As I will argue in section ., this opens up the way for reanalysing an A-moved VP in SpecFP as a structure in which all complement-head sequences are derived through roll-up. First, I will elaborate on the evolution of sentential negation.
. The diachrony of Latin negation .. Jespersen’s cycle The evolution of the Latin pre-verbal negator non argued for in this section can plausibly be considered part of Jespersen’s cycle, i.e. the well-documented tendency for the evolution of negation markers to follow a fixed cyclic pattern (see Jespersen ² []).17 A detailed structure of Jespersen’s cycle is given in () (from Breitbarth and Haegeman : section .., their ()):
For recent discussion, see (among many others) Van Gelderen (a, : ch. ), Jäger () and the contributions in Larrivée and Ingham (); particularly on Latin and Romance, Schwegler (, ) and Fruyt (: –). 17
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
()
Lieven Danckaert Stages in Jespersen’s cycle: stage single (pre-verbal/clitic) negation marker stage single (pre-verbal/clitic) negation marker plus optional phrasal emphasizer stage bipartite or embracing negation stage bipartite or embracing negation with the original marker having become optional and the original emphasizer having become the neutral negator stage single (phrasal) negation marker grammaticalized from the original emphasizer (stage ’ this marker becomes weakened to a clitic (pre-verbal) marker again)
Latin and its daughter languages provide some nice illustration of (). (Classical) Latin nōn itself is a contraction of two elements which arguably at some point formed a ‘bipartite’ negator (stage in ()), namely, (i) ne, the negator inherited from the Indo-European mother language and (ii) the emphasizer oinom ‘one’.18 I take it for granted that at the point where these two elements were fused together, the resulting lexical item was a phrase (XP) rather than a head (X°), and that only later it weakened to become the syntactic head that we can identify in the historical records (cf. Van Gelderen’s (, a, ) economy-based ‘Head Preference Principle’). Furthermore, we also know that the descendants of Latin non in at least some of the present day Romance languages are pre-verbal clitics (on French, see Kayne ; Rowlett ; on Italian, see Zanuttini (modern Italian varieties), (Old Italian); Manzini and Savoia : ch. ). For instance, Italian non canonically appears left adjacent to the highest verb, and can only be separated from it by clitic pronouns (arguably also X°’s, attached to the verb at an earlier point in the derivation). The same holds for French ne, to the extent that this element is expressed at all. The following three stages can thus be reconstructed with sufficient confidence. First, Latin noenom/nōn set out as a strong (phonologically independent), phrasal marker of sentential negation (stage ). Second, nōn gradually weakened to become a syntactic head. As a result, it obtained its strict pre-verbal position (namely, by virtue of the HMC). Third, (at least some) modern Romance pre-verbal negators that derive from Latin non are proclitic to the highest verb in their clause. .. Neg-incorporation, and its consequences I would like to propose that Classical and Late Latin constitute two different substages of stage in (), namely one in which non is already an X° but not yet proclitic, and one in which it becomes even weaker and is incorporated into the highest verb of the clause.19 The intuitive rationale behind this is that weakening of the XP nōn (*noenom) to clitic non/ne was a gradual process, which did not take place overnight. A number of substages are represented in (). In (a), the negative cf. Ernout and Meillet (4: ); Fruyt (: –). One could conjecture that these two stages reflect an increase in ‘structural deficiency’ (in the sense of Cardinaletti and Starke ) of non. 18 19
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
The decline of Latin VOAux
head non and the verb in T are still separate. As we have seen earlier, right on top of NegP, there is FP, whose head is endowed with an EPP-feature which at the next stage of the derivation will give rise to the creation of an inner specifier to host a moved XP. ()
a.
FP F° [+V, +EPP^]
NegP
Neg° [+V]
TP T° [+V]
...
Next, Neg and T form one single complex head in (b). As indicated, I assume that the syntactic correlate of this process of (pro)cliticization is verb movement to Neg°. We can speculate that this step of the process also goes hand in hand with phonological weakening of the negator, but whether this is indeed the case is at this point not obvious to verify. ()
b.
FP F° [+V, +EPP^]
NegP
Neg° [+V] Neg°
T°
TP ...
tT° [+V]
Finally, and very importantly, procliticization of the negator also has consequences elsewhere in the grammar. Earlier, it was argued that the ban on the hierarchically highest verb of the clause appearing to the left of non can be accounted for by calling upon the HMC. In the new grammar, this restriction still holds, but it is an effect of Neg-procliticization rather than of a constraint on movement. On the contrary, there now no longer is any reason why the (complex) verbal head could not move beyond NegP. Imagine that it moves to F°, giving rise to the configuration in (c): ()
c.
FP F° [+V, +EPP^] Neg° Neg°
F° T°
NegP tNeg° [+V]
TP tT° [+V]
...
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
Lieven Danckaert
In Danckaert (: –), it is argued at length that this last step is not only available, but that it also becomes the norm: assuming the ^-feature to be able to trigger both phrasal and head movement, a verbal head endowed with φ-features is now not only the closest goal for the probing F-head, but it would also be able to move up to F°, which effectively makes it possible for the EPP-requirement to be satisfied by means of head movement (an option which is explicitly made available in the systems proposed in Alexiadou and Anagnostopoulou and Biberauer ). Importantly, this development makes EPP-driven movement of VoiceP to SpecFP redundant.
. Syntactic reanalysis, and the demise of VOAux We are now in a position to explain the loss of the word order VOAux. Recall from section . that the frequency of this pattern drops sharply from the second half of the first century AD onwards, despite the order VPAux remaining productive (as we have seen, the incidence of the latter decreases too, but less dramatically so). In other words, there is reason to assume that the order VPAux, despite not being required anymore for the purpose of EPP-checking, was still available in the new grammar (i.e. the grammar with (i) Neg-incorporation and (ii) the new EPPparameter setting). This state of affairs can naturally be accounted for if we assume that the language learner reanalysed the string ‘VP-Aux’ as a roll-up structure, in which a large portion of the extended projection of the verb is moved to an outer specifier of F. Using the same notational conventions as above, the basic configuration thus arrived at can be represented in (), which is minimally different from (c) in that also the categorial V-feature of F now has the movement diacritic (as do all lower functional heads in the same extended projection, given FOFC):
FP
()
F° [+V^, +EPP^] Neg° Neg°
F° T°
NegP
tNeg° [+V^]
TP tT° [+V^]
...
I take it that at the level of FP, [+EPP^] triggers head movement, and [+V^] roll-up. For instance, a (CS)OVAux-clause like () would have the structure in (): ()
nisi gemmat-a fibul-a us-us non est unless jewelled-ABL brooch-ABL used-NOM.M.SG not be.PRS.SG ‘unless he did not use a jewelled brooch’ (= Hist. Aug. Carus, Carinus, Numerianus, .)
OUP CORRECTED PROOF – FINAL, 11/5/2018, SPi
The decline of Latin VOAux ()
SubjP
Subj’ Subj° [V]
FP NegP
FP
TP VoiceP
TP VoiceP
vP
vP
√P t√°
DPIA
tNeg°[V^] tTP
Neg° F°
tT°[+V^] tVoiceP Neg°
Voice’
Voice° [+V^]
F° [+V^, tNegP +EPP^]
NegP
T°
tvP
v° [+V^] t√P √°
v°
What we see here is a(n admittedly somewhat baroque) ‘snowballing’ derivation featuring no fewer than five applications of roll-up movement (the effect of some of which is obscured through head movement). In any event, it we adopt FOFC as formulated in (), the present analysis correctly predicts that whenever the categorial feature of F is endowed with the movement triggering feature, so is a clausemate v, resulting in the (quasi-)systematic co-occurrence of the order VPAux with the order OV (whereas the converse obviously doesn’t hold). In other words, Late Latin pre-Aux VoicePs are now subject to FOFC, and the order VOAux is no longer productively available.
. Conclusion In this chapter, I have analysed the diachrony of the Latin word order pattern VOAux. I have proposed that in Classical Latin VPAux-clauses involve a derivation in which the VP undergoes EPP-driven A-movement, an operation which is not subject to the (revised) ‘Final-Over-Final Constraint’ (cf. ()). I proceeded to suggest that in Late Latin this operation was reanalysed as an instance of roll-up movement, which is subject to FOFC. The crucial factor making this reanalysis possible is incorporation of the proclitic pre-verbal negator non into the hierarchically highest verb of the clause. As a result, the pattern VOAux is available in the earlier but not in the later period.
Acknowledgements The research reported on in this chapter was funded by postdoctoral grants of the ‘Bijzonder Onderzoeksfonds’ (BOF) of Ghent University (grant No. BOF/PDO/ ) and of the ‘Fonds voor Wetenschappelijk Onderzoek’ (FWO) (grant No. FWO/PDO/).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
14 On the decline of edge-fronting from Latin to Romance A D AM LE DGE W AY
. Introduction As noted by Väänänen (: ), discontinuous structures produced by edge-fronting (traditionally termed hyperbaton) represent ‘one of the most distinctive features of Latin with regard to Romance’, a feature which, although surviving residually in some of the earliest Romance texts (Poletto ), already shows signs of an inexorable decline in subliterary and late Latin (Adams : f., b: f.; Herman : –, : ; Bauer : –). Consequently, contrasts such as (a–b) highlight an important change in word order patterns from Latin to Romance, inasmuch as the former freely allows discontinuous structures such as (a), the result of edge-fronting of the adjective celeris to the DP left periphery in apparent violation of the Left Branch Condition (LBC; Ross ), whereas subextraction of prompt out of the French NP in (b) proves ungrammatical. () a. [DP [Spec celeris] [NP spe celeris subsidii]] confirmata swift.GEN hope.ABL assistance.GEN encouraged (Lat., Caes. B.C. ..) b. assurée par [DP [Spec (*prompt)] l’ [NP espoir d’ un prompt secours]] (Fr.) assured by swift the hope of a swift assistance ‘encouraged by the hope of speedy assistance’ Traditional arguments in favour of the non-configurational status of Latin have often highlighted the existence of pragmatically driven word order and discontinuous structures like (a), both of which were considered by Hale () robust criteria in establishing non-configurationality (Ledgeway : §.). Rather than treating them as two unrelated phenomena, with discontinuity interpreted as a purely structural concomitant of the absence of any hierarchical phrasal organization, I interpret phrasal discontinuity as a reflex of pragmatically driven word order
Word Order Change. First edition. Ana Maria Martins and Adriana Cardoso (eds) This chapter © Adam Ledgeway . First published by Oxford University Press
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
which, as part of a focusing/topicalizing device, targets individual pieces of structure fronting them to the left edge of individual phasal domains (DP, PP, vP, and CP) where they receive a pragmatically salient interpretation. Superficially, the Latin–Romance contrast in (a–b) would appear to suggest that the loss of hyperbaton and hence edge-fronting in the passage from Latin to Romance is not arbitrary, but follows from the complementary distribution of fronting to the DP left periphery and the presence of articles (cf. Uriagereka : ). In particular, the pattern highlights how the accessibility of the DP periphery hinges directly on properties of the D position which, if inactive/deficient as in Latin which lacks articles, freely allows fronting to its associated specifier, but blocks any such fronting in languages with articles like Romance where D is very much active. Indeed, this empirical generalization finds further support in Poletto’s (: §.) analysis of early Tuscan, where she observes that in those very rare cases in which ‘an XP is preposed in front of the N, the N never has a definite determiner.’ There consequently emerges a complementary development in the gradual loss of edge-fronting and the progressive emergence of the article (Ledgeway : §..), apparently placing the cause of this change with properties of D. Although empirically correct, we shall demonstrate that the correlation between the loss of edge-fronting and the appearance of articles is epiphenomenal and ultimately to be understood as the reflex of a change in the head parameter.
. Deconstructing non-configurationality In contrast to traditional accounts of changes in sentence structure from Latin to Romance in terms of the progressive rise of configurationality and concomitant functional structure (see Ledgeway : chs , ), I argue that the same empirical generalizations can be captured by assuming the presence of both configurationality and functional structure already in Latin. In essence, the perceived non-configurationality of Latin can be broken down into two main ingredients: i) grammatically free word order resulting from an ongoing change in the head directionality parameter, which a priori allows dependents/complements to occur on either side of their head; and ii) pragmatically driven word order, often producing discontinuous structures, resulting from the greater accessibility of topic- and focus-fronting to positions situated in the left edge of individual functional projections. Interpreted in this manner, the apparent emergence of configurationality in Romance is to be understood as the surface effect of the rigidification of the directionality parameter and the restricted accessibility of edge-fronting to left-peripheral positions within the functional structure, formal changes which conspire to give the superficial impression of differences in configurationality and functional structure in Latin and Romance. .. Deconstructing grammatically free word order: roll-up The gradual rigidification of word order in the passage from Latin to Romance can be explained in terms of a progressive reversal of the directionality parameter (for in-depth qualitative and quantitative analyses, see Ledgeway : –). Assuming the ordering of heads (in bold) and complements (underlined) in the development
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway
from archaic Latin to Romance to have undergone a shift from one harmonic principle of linear organization to another (namely, head-last ) head-first)—witness the consistently typical head-final and head-initial positions of parisuma and pari illustrated for archaic Latin and Romance in (a–b), respectively—the greater freedom of word order traditionally recognized for Classical Latin can been seen as a result of its occupying an artificially sustained intermediate position in this change, resulting in mixed (dis)harmonic linearizations like (a–b):1 () a. quoius forma uirtutei parisuma fuit (archaic Lat., CIL 2.) whose beauty.NOM valour.DAT most.equal.NOM was b. la cui bellezza fu pari al valore (It.) the whose beauty was equal to.the valour ‘whose beauty was fully equal to his valour’ () a. constantibus hominibus par erat (Lat., Cic. Diu. .) resolute.ABL.PL men.ABL equal.NOM it.was ‘[our apprehension] was equal to that of men of strong character’ b. illa erat uita [ . . . ] libertate esse parem ceteris that.NOM was life.NOM freedom.ABL be.INF equal.ACC rest.DAT.PL ‘What he considered life [ . . . ] was the being equal to the rest of the citizens in freedom’ (Lat., Cic. Phil..) The oscillation between head-last and head-first structures in the history of Latin, an area of variation of a purely formal nature with no discernible effect on interpretation, can be captured along two axes of variation, the first in terms of diachronic variation (head-last (archaic Latin) ) head-initial (early/late Latin)) and the second in terms of diaphasic (and no doubt diastratic and diamesic) variation (head-final (formal, literary) vs head-initial (subliterary, colloquial)). These facts can now be captured in terms of a progressive reversal of the directionality parameter towards a head-initial setting (Adams a; Oniga : f.; Clackson and Horrocks : , ), with classical Latin (a–b) displaying an ambivalent behaviour on account of its non-uniform characterization in relation to these two dimensions of variation, namely, non-archaic (predominantly head-first), but formal and literary (predominantly head-final). Formally, this linear variation can be understood in terms of the application or otherwise of a L(inearization)-movement termed roll-up. Following Kayne (), I assume the Universal Base Hypothesis according to which all languages present the same underlying head-initial word order, with identical arrangement of external (subject) and internal (object) arguments across all other categories, namely SXO. Thus, when the complement of a verb, for example, surfaces to the left of its head, it must have moved leftwards, i.e. rolled up, across the verb from its basegenerated position to a derived (inner) specifier. Both options from the verbal and
1 The causes underlying this change in the linearization of head and complement are explored in terms of the Final-over-Final-Constraint (FOFC) and the principle of cross-categorial harmonization in Ledgeway (: –).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
nominal domains are exemplified with the following minimal pairs, together with their structural representations in (). () a. instruit aciem (Lat., Liv. ..; complement in situ) drew.up.SG battle.line.ACC b. aciem instruit aciem (Lat., Liv. ..; roll-up) battle.line.ACC drew.up.SG ‘(he/the Samnite army) drew up battle line’ () a. metum rerum nouarum (Lat., Cic. Att. ..; complement in situ) fear.ACC things.GEN new.GEN ‘fear of revolution’ b. urbanarum rerum metum urbanarum rerum (Lat., Cic. Att. ..; roll-up) urban.GEN things.GEN fear.ACC ‘because of concern of the situation in the city’ ()
a.
V/NP Spec V/N instruit metum
b.
V/N' DP
V/NP Spec aciem urbanarum rerum
aciem rerum nouarum
V/N' V/N DP instruit aciem metum urbanarum rerum
In this light, the distinct grammatical organization of Romance and, in particular, its rigid head-complement order, now finds a straightforward explanation in terms of the loss of roll-up from the grammar, as exemplified in the corresponding French structures: () a. ils (* en ligne) se rangèrent en ligne (Fr.) they in line selves arranged in line b. la (* des choses) crainte des choses (Fr.) the of.the things fear of.the things .. Deconstructing pragmatically free word order: edge-fronting Besides fluctuation at the syntactic level between a conservative head-final and an innovative head-initial order, pragmatics is also recognized to play a significant role in determining Latin word order. This aspect of Latin sentential organization, largely absent in Romance, can be captured by assuming the greater accessibility of topicand focus-fronting to left-peripheral positions situated in the left edge of individual functional projections.2 In Romance, functional structure is readily exploited and 2 Apart from specific stylistic formulae, discontinuities and structural reorganizations produced by leftedge fronting are not necessarily ‘either a formal rhetorical feature or a colloquial feature’ (Powell : ),
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway
made visible through the lexicalization of head positions with functional categories such as determiners and auxiliaries (a), as well as through operations such as N-/V-raising (b) to these same head positions (Belletti ; Longobardi ; Cinque ; Ledgeway : ch. ). By contrast, Latin lacks such functional categories and N-/V-raising, but displays ubiquitous evidence for the presence of functional structure through its extensive exploitation of topic- and focus-fronting to the left-edge specifier positions of these same functional projections (c): () a. [TP/DP [Spec Ø] [T0 /D0 XAux/Det [VP/NP V/N XP]]] (Romance) b. [TP/DP [Spec Ø] [T0 /D0 V/N [VP/NP V/N XP]]] (Romance) c. [TP/DP [Spec XPTopic/Focus] [T0 /D0 Ø [VP/NP V/N XP]]] (Latin) Informally, we can say that at the level of functional structure Latin privileges an XP-type syntax and Romance an X-type syntax, with generalized overt exploitation of left-peripheral specifier positions in the former and head positions in the latter. By way of illustration, consider the examples in (), where SpecDP is overtly signalled by the discontinuous focused adjective summo in the Latin example (a), whereas in its Spanish translation (b) it is the head of DP that is overtly spelt out by the presence of the indefinite article un while the adjectival modifier del más alto remains in situ. In short, there emerges a strict complementarity between pragmatic edgefronting and the availability of functional categories like determiners and auxiliaries. Consequently, I hypothesize that the pragmatically determined word order of Latin follows from the availability of fronting to left-peripheral positions within the left edge, an operation apparently blocked in Romance (c) by the overt lexicalization of the corresponding head position. () a. [DP [Spec summo] [D0 Ø [NP homo [summo highest.ABL man.NOM b. [DP [Spec Ø][D0 un [NP hombre [del a man of.the
ingenio]]]] talent.ABL (Lat., Cic. De or. .) más alto talento]]]] (Sp.) more high talent
c. *[DP [Spec del más alto] [D0 un [NP hombre [del más alto talento]]]] (Sp.) of.the more high a man talent ‘a man of the highest ability’ Once local fronting to left-peripheral positions becomes available, these same positions may function as escape hatches, feeding further fronting operations to more remote and pragmatically more salient left-peripheral positions within the nominal and clausal superstructure, witness the extraction of summam to clause-initial position in () violating the LBC. but simply represent the way in which discourse information is conveyed in Latin (Kessler : ; Spevak ), an observation which undoubtedly aligns left-edge movement with a pragmatic device employed to license focused and, to a lesser extent, topicalized readings (Devine and Stephens : –; Powell : f., f., ). This is independently supported by the widespread recognition that the words involved in edge-fronting belong to the typical categories of focus and topic modifiers, the same classes of ‘preferential words’ (Powell : ) that Adams independently identifies as focal hosts for clitic pronouns (a: –) and clitic forms of ESSE ‘be’ (b: –).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
() qui [CP [Spec summam] [C0 inter eos habet [DP [Spec summam] who-NOM highest.ACC among them.ACC has [NP [summam]auctoritatem]]]] (Lat., Caes. B.G. ..) power-ACC ‘who has the highest authority among them’
. The decline of edge-fronting .. Previous accounts It is tempting to think of the decline of edge-fronting in the passage from Latin to Romance as a concomitant of the decline of the case system, since pragmatically driven word orders and discontinuity in Latin appear to be licensed by the availability of a rich case and agreement system (Bauer : ). Yet, as convincingly argued by Herman (: ), the case and agreement system of Latin was essentially still very much intact at the time of the first signs of the decline in edge-fronting, forcing us to recognize that ‘the loss of structural disjunction was not a consequence of the loss of the case system.’ There must therefore be some other explanation for the loss of edge-fronting in the passage from Latin to Romance. Above we saw that edge-movement targets a piece of nominal substructure and fronts it under focus/topic to at least the left periphery of the immediately containing DP (cf. a), from where it can escape the confines of the containing DP and raise to higher and pragmatically more salient focus/topic positions in the left edge of PP, vP, and CP (cf. ). However, in the absence of raising to the DP left periphery, all other potential higher edgemovements are blocked by virtue of the Phase Impenetrability Condition (PIC), an observation which highlights how the essential difference between Latin and Romance crucially lies at the level of the DP. Indeed, Romance presents extensive evidence for left-peripheral topic/focus positions at the level of CP and vP (cf. Ledgeway : §.., §...), suggesting that what prevents these from being targeted by nominal substructures must be the result of some barrier effect of the Romance DP which blocks subextraction out of NP in accordance with Ross’s () Left Branch Condition.3 Consequently, alongside the post-nominal base position (a), Latin also allows extraction of adjectives out of the NP to the DP left periphery (a) or, from there to a higher left periphery (a), whereas Romance allows neither of these latter two options (b, b). As a consequence, in Romance the only legitimate way for such modifiers to receive a pragmatically marked reading and reach the lower vP or higher CP left periphery is for the whole DP containing them to undergo edge-fronting (a–b). () a. inter [DP [NP aues [ albas]]] uetuit consistere coruum among birds.ACC white.ACC.PL he.forbade stay.INF raven.ACC (Lat., Ov. Met..–)
3
On French split-DP structures involving combien ‘how much’, see Baunaz ().
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway b. entre [DP [D0 las [NP aves [blancas]]]] vedó al cuervo among the birds white he.forbade to.the raven asentarse (Sp.) settle.NF.REFL ‘he stopped the raven from living among the white birds’
() a. [DP [Spec cumque albis] [D0 Ø [NP ossa [[cum albis] with.and white.ABL.PL bones.ACC medullis]]]] [ . . . ] condebat in aluum (Lat., Ov. Met. .–) marrow.ABL.PL he.buried in belly.ACC b. sepultaba [DP (*[Spec blancas]) [D0 los [NP huesos [con he.buried white.F.PL the bones with médulas [blancas]]]]] (Sp.) marrow.F.PL white.F.PL ‘he filled his belly with bones full of white marrow’ () a. [PP [Spec albis] in [DP [Spec albis] [D0 Ø [NP turribus [albis]]]]] white.ABL.PL in towers.ABL (Lat., Ov. Met. .) b. [PP (*[Spec blancas]) en [DP ([Spec blancas]) [D las [NP torres white in the towers [blancas]]]]] (Sp.) white ‘among the white dovecotes’ () a. *[ blanca] escogieron [DP [Spec blanca] [D0 la [NP limusina [blanca] ]]] (Sp.) white they.chose the limousine b. [DP la limusina BLANCA] escogieron [DP la limusina [blanca] ] (Sp.) the limousine white they.chose ‘they chose the WHITE limousine’ As observed in §, Latin–Romance contrasts such as ()–() point to a complementary distribution between fronting to the DP left periphery and the presence of articles. Diachronically, we witness a complementary development in the gradual loss of edge-fronting and the emergence of the article (cf. Ledgeway : §..), an observation which suggests the cause of this change should be situated with properties of D. Formally, there are a number of ways of implementing this empirical generalization, several of which we briefly review here (for an overview, see Bošković , a,b, ). For example, the Latin–Romance contrast could potentially be assimilated to the traditional parametric variation observed in subjacency effects across languages (Rizzi ), where it is noted, for instance, that, besides NP, IP is a bounding node in English, but not in Italian where CP is a bounding node. This difference therefore accounts for the fact that movement of the relative, which can cross at most only one bounding node, is illicit in the English example (a) where it crosses two bounding nodes, but well-formed in the Italian (b) where it crosses only one bounding node.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
() a. *George’s new idea [CP [of which] [IP I can imagine [CP [what] [IP you think of which]]]], will soon become known to everybody b. la nuova idea di Giorgio [CP [Spec di cui] [IP immagino [CP [Spec the new idea of Giorgio of which I.imagine che cosa] [IP pensi di cui ]]]] diverrà presto di pubblico what thing you.think will.become soon of public dominio (It.) domain Along similar lines, we might hypothesize that, whereas in Romance both DP and NP are bounding nodes, in Latin only NP counts as a bounding node and therefore fails to inhibit extraction out of DP in cases as such (a, a). While empirically adequate, this approach is however nothing more than a stipulation that does little more than restate the original empirical observation. Another approach, vigorously defended by Bošković (, a,b, , ) and Bošković and Gajewski (), is to assume that, in contrast to Romance, Latin entirely lacks the DP layer such that all nominals are realized as NPs (for a critical appraisal of the evidence, see Ledgeway : §.). This explains why Romance has articles but Latin does not and, at the same time, why Latin is able to violate the LBC since it lacks the barrierhood/phasehood instantiated by the DP layer present in languages with articles. On this view, one way to interpret the Romance ban on leftbranch extraction is to appeal to some notion of antilocality which ascribes the illicit nature of edge-fronting to the DP left periphery to the fact that it produces movement which is considered ‘too’ local/short and hence uneconomical. There are many proposals in the literature that broadly appeal to such a concept of antilocality (cf. Bošković , , a; Ishii ; Saito and Murasugi ; Grohmann , ; Abels , ): what these all have in common is the idea that movement must result in raising outside of the immediate minimal domain or phase, thereby crossing at least one (phasal) XP boundary. Consequently, in an NP-language like Latin (a) subextraction of an adjective proves well-formed, since NP does not qualify as a minimal domain or phase such that the observed movement escapes both antilocality and the Phase Impenetrability Condition (PIC). By the same token, in a DP-language like Italian any subextraction is ruled out precisely because it violates antilocality (b), as is raising bypassing the left periphery (c) which would violate the PIC which requires movement out of the containing phase to transit through the left edge of the DP. () a. clarissimae [NP testis uictoriae clarissimae] very.distinguished.GEN witness.NOM victory.GEN (Lat., Cic. Off. .) b. *[DP [Spec gloriosa] [D0 il [NP testimone glorious the witness c. *gloriosa [DP [D0 il [NP testimone della glorious the witness of.the ‘(the) witness of his glorious victory’
della vittoria gloriosa]]] (It.) of.the victory vittoria gloriosa]]] (It.) victory
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway
Another possibility, still within this NP/DP approach to the observed Latin–Romance contrast, is to assimilate nominal edge-fronting structures to clausal wh-extraction structures (Corver ): in the same way that realization of the complementizer that in (a) is said to block the tail of the extracted wh-phrase who from connecting with its higher surface and intermediate positions, a so-called that-trace effect, the illformedness of (b) can be understood in terms of the intervening effect of the definite article the, which, on a par with that in (a), prevents the tail position of the adjective red from being properly bound by its higher intermediate and surface positions.4 Such an analysis, however, incorrectly predicts that extraction via the left periphery of all phases, and not just DPs, should prove ill-formed when the head of that phase is also simultaneously filled. Yet, Ledgeway (: §...) demonstrates that extraction via SpecCP is indeed possible when the head position is also filled by an overt complementizer. () a. [Who] do you think [CP [Spec who] [C0 (*that) [TP [who] spoke?]]] b. *[Red] he bought [DP [Spec red] [D0 the [NP [red] book]]] More generally, there are some serious problems with the idea that the relevant Latin–Romance contrast can be reduced to the NP/DP distinction. First, if Latin lacks DP structure, then we are unable to capture the systematic left-peripheral topic/focus positions targeted by modifiers under local edge-fronting observed above (for detailed discussion, see Ledgeway : §..). Second, recent work by Giusti and Oniga (, ), Giusti and Iovino (), and Iovino (, a,b) has shown that Latin nominal expressions project a DP whose specifier hosts demonstratives, and a further left-peripheral projection which can host one pragmatically marked constituent. Finally, the blanket ban on local movement assumed by antilocality to account for the relative well-formedness of left-branch extraction in Latin and Romance proves too powerful, since it also excludes the independently motivated operation of roll-up, namely complement-to-specifier movement, required to account for surface variation in the head parameter (see also Biberauer, Holmberg, and Roberts ). Thus, although it appears necessary to conclude that the basic empirical generalization regarding the observed blocking effect of the LBC, absent in Latin, is related to the presence/absence of articles in the two languages, ultimately this distinction cannot be framed in terms of the NP/DP distinction. Nor is it possible, in view of the predictions of the PIC and antilocality, to claim, all things being equal, that the Latin D head (which fails to license articles) is in some sense defective, thereby neutralizing the potential barrierhood/phasehood of DP and licensing edge-fronting, whereas in Romance this same D head (which does license articles) is fully active and therefore able to endow DP with the observed barrierhood. Below I shall develop an
4 As pointed out by an anonymous reviewer, the supposed parallel between that-trace effects and edgefronting is further weakened by the observation the former only applies to extracted subjects, whereas no such restrictions apply to the extracted phrase in edge-fronting structures.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
account which derives the relevant distinction and decline of edge-fronting from changes in the head parameter and their implications for antilocality which, in turn, correlate with the covert/overt lexicalization of functional head positions such as determiners. .. Proposal Contra Bošković, I assume that nominals in Latin and Romance both project to DP, and that locality of movement is a priori constrained by the PIC and antilocality. Within this scenario, two basic questions need to be answered: ()
i. Why is edge-fronting of adjectives and other adjuncts to SpecDP possible in Latin, but not in Romance? ii. Why was edge-fronting progressively lost in the passage from Latin to Romance?
Given that the availability of edge-fronting cannot be reduced to the absence/ presence of the DP layer and concomitant availability of articles, it must follow from some other major difference between Latin and Romance. This difference, I argue, is to be identified with the head parameter, the setting of which was shown in §. to diverge in the two varieties: whereas in Romance the parameter is unequivocally aligned with the head-initial setting, in Latin the parameter fluctuates between both settings as a result of its occupying an intermediate position in the gradual shift from head-finality to head-initiality. In turn, this difference in the head parameter is ultimately responsible for the observed variation in edge-fronting, inasmuch as its setting directly determines the application or otherwise of antilocality in constraining movement. In particular, I propose that antilocality is not a blanket constraint on all instances of local movement, but should be parametrized across languages. This is already independently forced upon us by our analysis in §. of grammatically free word order in Latin, where head-finality was interpreted as the output of a roll-up operation that raises the complement to the specifier to the left of its selecting head (cf. b, b, b). Antilocality, however, predicts that roll-up should be excluded as a movement which is a priori too local/short. The suspension of antilocality therefore appears to constitute a sine qua non for languages like Latin which exhibit head-final orders derived through roll-up movement in accordance with the proposals in Kayne (: f., f.) and recent work on the Final-Over-Final-Constraint (cf. Biberauer, Holmberg, and Roberts , , ; Biberauer, Newton, and Sheehan ; Biberauer and Sheehan ). In Romance, by contrast, the head parameter is consistently aligned with the head-initial setting such that roll-up, and hence antilocal movement, never arises in the grammars of these varieties. We thus naturally derive from the different settings of the head parameter a concomitant parametrization in the role of antilocality in constraining movement in the two languages ultimately to be understood in terms of economy and markedness. On the reasonable assumption that head-initiality represents the unmarked and least costly option, as made explicit in the structural analysis (a) where head-initiality instantiates the
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway
basic, underlying order,5 I take antilocality to apply in full in head-initial languages as the default option. By the same token, given the derivationally more complex nature of head-finality proposed in (b), I take the suspension of antilocality to represent the more marked option which arises just in those languages where the grammar (systematically or optionally) requires the marked option of linearization-related roll-up movement. In short, we predict that the application or otherwise of antilocality is parasitic on the head-initial vs head-final parametric distinction: once its potential effects are quashed by the positive evidence of head-finality required to motivate roll-up, it fails to apply across the board licensing other short/local movements such as edge-fronting in apparent violation of the LBC. Returning to question ((i)) above, we now have a principled account for Latin– Romance contrasts such as (a–b): antilocality only applies in head-initial languages like Portuguese where it rules out the extraction of adjectives like nociva in (b) to SpecDP, but fails to apply in (partially) head-final languages like Latin where potential constraints on local movement are overridden by the requirement of rollup movement (witness the pre-nominal complement rei publicae in (a)) to satisfy the language’s head-final setting. () a. [DP [Spec infestam] [NP [Spec rei publicae] [N0 infestam rei publicae dangerous.ACC thing.DAT public.DAT pestem]]] (Lat., Cic. Cat. .) plague.ACC b. [DP [Spec (* nociva)] [D 0 uma [NP praga nociva ao dangerous a plague dangerous to.the estado praga]]] (Pt.) state ‘a plague dangerous to the state’ We are now also in a position to provide a complementary and highly natural solution to question ((ii)). Put succinctly, the loss of edge-fronting in the passage from Latin to Romance turns out simply to be a side-effect of the gradual reversal in the head parameter, whose ultimate head-initial setting signals the loss of roll-up and, with it, the concomitant loss of all local movements in accordance with the predictions of antilocality. As is apparent, the availability of roll-up and that of edge-fronting are intimately linked, insofar as the presence of the former implies, all things being equal, the availability of edge-fronting. This is a strong prediction and requires further in-depth study to assess its validity. However, some preliminary support for this view comes from Bošković’s work (see, for instance, his b study) on cross-linguistic variation in the well-formedness of left-branch extraction and hence edge-fronting. He establishes a robust correlation between languages that permit left-branch extraction and scrambling, an optional movement operation which, following Bošković (c), is
5 Similarly, according to the ReCoS group (http://recos-dtal.mml.cam.ac.uk) head-initial is less marked than head-final since head-final requires ^ (= movement diacritc), an extra feature (Ian Roberts p.c.).
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
assumed to be devoid of any pragmatico-semantic import (hence not topicalization/ focalization) that applies ‘for no reason at all’ (pace the Fox–Reinhart intuition to optionality). It is this same type of ‘grammatically free’ L(inearization)-movement which was argued in §. to characterize Latin through roll-up. In short, Latin belongs to those languages such as Serbo-Croatian, Russian, Japanese, and Hindi which are said to exhibit grammatically free word order (here interpreted as roll-up), which Bošković also identifies, albeit for different reasons, as a pre-requisite for licensing left-branch extraction. ... Reanalysis: roll-up ) edge-fronting The interrelatedness between roll-up (scrambling) and edge-fronting (left-branch extraction) finds further support in the observation that the two operations prove, in some cases, impossible to tell apart, such that erstwhile cases of roll-up are amenable to reanalysis as edge-fronting. In particular, given the textual nature of the Latin evidence, it is not always possible to tell whether a given string represents a case of grammatically induced raising with no pragmatic import, namely roll-up (a), or a case of pragmatically induced raising under focalization/topicalization, namely edge-fronting (b), since the surface landing position of the modifier, be it the left-peripheral specifier of NP (roll-up) or DP (edge-movement), is invariably to the left of the nominal head.6 () a. [DP [NP [Spec familiaris funeris] excusatio familiaris funeris]] b. [DP [Spec familiaris funeris] [NP excusatio familiaris funeris]] domestic.GEN funeral.GEN excuse.NOM ‘the pretext of a domestic calamity’ (Lat., Cic. Rab. Perd. ) Furthermore, as noted in Ledgeway (: §...), in an archaic period, only partially accessible to us, adjectival and nominal modifiers must have stood in prenominal position under unmarked pragmatic readings, but had already largely shifted to the post-nominal position by the time of most of our earliest records, with the pre-nominal position surviving as the position that licenses marked focused/ topicalized readings (cf. also Langslow .).7 These facts find an immediate explanation in terms of our conclusions about the interaction of roll-up and edgemovement above. With the progressive loss of roll-up as an option in the history of Latin and in the passage to Romance, we can immediately understand how the erstwhile roll-up operation which, as a purely grammatical principle of linearization devoid of any pragmatic import, originally placed complements and modifiers to the left of their associated head (cf. a), was progressively reanalysed, albeit with some degree of ambiguity, as a marked pragmatic order, no longer derived through roll-up,
6 No ambiguity arises when other disambiguating elements occur within the nominal group as in (i), where the prepositional adjunct ex urbe must have been fronted under roll-up to SpecNP and not fronted under focus to SpecDP since it is preceded by the demonstrative hosce in SpecDP:
hosce [NP [Spec ex urbe] sicarios ex urbe]]] (Lat., Cic. S.Rosc. ) out.of city.ABL killers.ACC these.ACC ‘these city assassins’ See also the later stages of the loss of OV in middle English (Biberauer and Roberts , ).
(i) [DP [Spec 7
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway
but increasingly through edge-movement to the left periphery of the relevant containing phase (b). Indeed, this view finds independent support in Devine and Stephens’ (: chs , ) survey of the position of nominal arguments and modifiers within the nominal group in classical prose. They find that the derived pre-head position tends to correlate with marked pragmatic readings (topic/focus), with very few cases of pragmatically unmarked preposing. A similar picture is noted in Ledgeway (: §...) regarding the increasingly marked partial anteposition (so-called ‘splitting’) of the accusative with infinitive (AcI) in non-literary texts and authors, which Adams (: f., f.) argues to correlate quite systematically with a marked (topic/focus) interpretation of the anteposed material. Both of these findings are highly indicative of an ongoing reanalysis of obsolescent p(arameter)-ambiguous (Clark and Roberts ) roll-up orders as pragmatically driven edge-movements, namely, a shift from L-movement (triggered by ^) to A0 -movement (triggered by an edge feature).
. Conclusion It has repeatedly been noted in the literature (cf. Uriagereka : ; Bošković , c, a,b, , ) that languages which violate the LBC also lack articles, an empirical observation also shown to hold for the relevant Latin–Romance contrast. Although this contrast was argued above to follow from the head parameter, in turn responsible for the parametrization of antilocality contraints on movement, it remains to be explained what significance, if any, the availability of articles has on the licensing of edge-fronting, and how this empirical generalization can be integrated with the analysis of left-branch extraction advocated here. According to a traditional observation (Schwegler ), the principal typological difference between Latin and Romance lies in a distinction between morphology and syntax: while Latin predominantly makes recourse to synthetic structures, the morphologically poorer Romance varieties make greater use of analytic structures (cf. Ledgeway : chs , , ). This is readily observable in the gradual loss of the case system (a), synthetic verb forms (a), and implicit subordination structures such as the AcI (a) in favour of analytic structures based on functional categories like determiners and prepositions (b), auxiliaries (b), and complementizers (b): () a. canis nonne similis dog.NOM not.Q similar.NOM b. el gos és semblant al the dog is similar to.the ‘are dogs similar to wolves?’ () a.
PLU-ET
(Lat.) rain.FUT.SG
b. at a próere (Srd.) it.has to rain.INF ‘it will rain’
lup-o (Lat., Cic. Nat. D. .) wolf.DAT llop? (Cat.) wolf
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
On the decline of edge-fronting from Latin to Romance
() a. tacitum te dicere credo / ‘Non reddes’ (Lat., Mart. ..–) silent.ACC you.ACC say.INF I.believe not you.give.back b. criju ca ti sta diciennu ‘unn’ ’i ripaghi’ (Cal.) I.believe that yourself PROG.SG saying not them you.repay ‘I fancy you say to yourself: “You won’t repay them” ’ Following Ledgeway (: ch. , ), we take Romance analyticity to be an epiphenomenal development, the manifestation of a deeper change but not its cause. At bottom, the analytic patterns of Romance are nothing more than the partial reflex of a more deep-rooted structural change, here interpreted in terms of a rigidification of the head-directionality parameter (§.) and its consequences for antilocality leading to the loss of edge-fronting to the DP left periphery. It is these formal changes in the directionality parameter and the differential role of functional structure in the two varieties which give the impression of differences in configurationality and functional structure in Latin and Romance. In this regard, we have seen that whereas Latin typically spells out left-peripheral specifier positions through roll-up and edge-fronting (cf. c), Romance typically spells out the corresponding head positions with functional categories (cf. a) such as determiners (cf. b), auxiliaries (cf. b), and complementizers (cf. b), or through head-movement operations entirely absent from Latin (cf. b), giving rise to an important typological difference in the syntax of Latin and Romance which privilege an XP-type and X-type syntax, respectively. We thus conclude that the rise of articles is not in any way responsible for the decline of edge-fronting but only indirectly linked to it, in that both developments represent independent reflexes of the change from head-finality to head-initiality, which signals a move away from a Latin dependent-marking syntax in which various dependents are raised to lexical and functional left-peripheral specifier positions in apparent violation of antilocality towards a Romance head-marking syntax in which functional head positions are increasingly spelt out under direct lexicalization or movement. There thus emerges a correlation between head-finality and syntheticity on the one hand, and between head-initiality and analyticity on the other (Bauer : ; Dryer ; Ledgeway : –), inasmuch as ‘left-branching morphological structures tend to become synthetic forms, whereas right-branching elements remain separate’ (Bauer : f.). Significant in this regard is Ledgeway’s (: §..) observation that Latin displays a degree of competition across different areas of the grammar between inherited head-final structures and innovative head-initial structures, where the relevant functional head is systematically phonologically null and overt, respectively. This can be seen in the following examples which contrast the AcI (a), a head-final construction introduced by a null head-final complementizer (cf. also Cecchetto and Oniga ), with a finite complementation structure introduced by an overt head-initial complementizer QUOD (b), and bare case-marked nominals (a), here analysed as silent postpositional phrases headed by a head-final null adposition (cf. also Emonds : ), with head-initial prepositional structures (b): () a. [[ tacitum te dicere V-TP] Ø CP] credo (Lat., Mart. ..) silent.ACC you.ACC say.INF I.believe ‘I fancy you say to yourself ’
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Adam Ledgeway b. scis enim [CP quod [V-TP epulum dedi]] (Lat., Petr. Sat. .) you.know for that feast.ACC I.gave ‘for you remember that I gave a public banquet once’
() a. Pompeius [ . . . ]proficiscitur [[ Canusium] Ø PP] (Lat., Caes. B.C. ..) Pompey.NOM sets.out Canusium.ACC ‘Pompey . . . sets out for Canusium’ b. miles [PP ad [Capuam]] profectus sum (Lat., Cic. Sen. ) solider.NOM to Capua.ACC set.out I.am ‘I set out as a soldier for Capua’ We thus witness a parallel development in the sentential and nominal domains where an archaic head-final structure headed by a null complementizer (AcI) or null adposition (bare case-marked nominal) is progressively rivalled, and eventually replaced entirely, by an innovative head-initial structure headed by an overt complementizer and preposition, respectively. This is precisely what we see in the development of the DP, which passes from an original head-final arrangement with a null D head (Latin, without articles) to a head-initial arrangement with an overt D head (Romance, with articles).8
8 It is significant to note that recent typological research (cf. Dryer : f., ) has equally highlighted a strong correlation between head-finality/-initiality and the phonologically null/overt realization of functional categories such as complementizers, auxiliaries, and determiners.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References Abels, Klaus (). Successive Cyclicity, Anti-locality, and Adposition Stranding. PhD dissertation, University of Connecticut, Storrs. Abels, Klaus (). Phases: An Essay on Cyclicity in Syntax. Berlin: De Gruyter. Aboh, Enoch (). ‘Snowballing movement and generalized pied-piping’, in Anne Breitbarth and Henk van Riemsdijk (eds), Triggers. Berlin: De Gruyter, –. Ackema, Peter, and Ad Neeleman (). ‘Context-sensitive spell-out’. Natural Language and Linguistic Theory : –. Ackerman, Farrell, and Gert Webelhuth (). A Theory of Predicates. Stanford, CA: CSLI Publications. Adams, James Noel (). ‘A type of hyperbaton in Latin prose’, Proceedings of the Cambridge Philological Society : –. Adams, James Noel (a). ‘A typological approach to Latin word order’, Indogermanische Forschungen : –. Adams, James Noel (b). The Text and Language of a Vulgar Latin Chronicle (Anonymus Valesianus II), Bulletin Supplement (University of London, Institute of Classical Studies) no. . London: Institute of Classical Studies. Adams, James Noel (a). ‘Wackernagel’s Law and the position of unstressed pronouns in Classical Latin’, Transactions of the Philological Society : –. Adams, James Noel (b). Wackernagel’s Law and the Placement of the Copula Esse in Classical Latin. Cambridge: The Cambridge Philological Society. Adams James Noel (). ‘The accusative + infinitive and dependent quod-/quia-clauses: the evidence of non-literary Latin and Petronius’, in Sander Kiss, Luca Mondin, and Giampaolo Salvi (eds), Latin et langues romanes: études de linguistique offertes à József Herman à l’occasion de son ème anniversaire. Tübingen, Germany: Niemeyer, –. Adams, Marianne (). Old French, Null Subjects and Verb Second Phenomena. PhD dissertation, University of California, Los Angeles. Agbayani, Brian, and Chris Golston (). ‘Second-position is first-position: Wackernagel’s law and the role of clausal conjunction’, Indogermanische Forschungen: Zeitschrift für Indogermanistik und allgemeine Sprachwissenschaft : –. Alexiadou, Artemis, and Elena Anagnostopoulou (). ‘Parametrizing AGR: word order, V-movement, and EPP checking’. Natural Language and Linguistic Theory : –. Alexiadou, Artemis, and Elena Anagnostopoulou (). ‘The subject-in-situ generalization and the role of case in driving computations’. Linguistic Inquiry : –. Algeo, John (). ‘The origins of Southern American English’, in Stephen J. Nagle and Sara L. Sanders (eds), English in the Southern United States. Cambridge/New York: Cambridge University Press, –. Amaral, Patrícia, and Fabio Del Prete (). ‘On truth persistence: a comparison between European Portuguese and Italian in relation to sempre’, in Marie Hélène Côté and Eric Mathieu (eds), Romance Languages: Linguistics Symposium on Romance Languages (LSRL), Ottawa, – May . Amsterdam/Philadelphia: John Benjamins, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Ambar, Manuela (). Para uma Sintaxe da Inversão Sujeito-Verbo em Português. Lisboa: Colibri. Ambar, Manuela, Manuela Gonzaga, and Esmeralda Negrão (). ‘Tense, quantification and clause structure in EP and BP: evidence from a comparative study on sempre’. Paper presented at Going Romance , University of Groningen (– November ). . Ambar, Manuela, Manuela Gonzaga, and Esmeralda Negrão (). ‘Tense, quantification and clause structure in EP and BP: evidence from a comparative study on sempre’, in Reineke Bok-Bennema, Bart Hollebrandse, Brigitte Kampers-Manhe, and Petra Sleeman (eds), Romance Languages and Linguistic Theory : Selected Papers from ‘Going Romance’, University of Groningen (th–th November ). Amsterdam/Philadelphia: John Benjamins. –. Anagnostopoulou, Elena (). ‘Participles and Voice’, in Artemis Alexiadou, Monica Rathert, and Arnim von Stechow (eds), Perfect Explorations. Berlin: De Gruyter, –. Anderson, Gregory D. S. (). Auxiliary Verb Constructions. Oxford/New York: Oxford University Press. Androutsopoulou, Antonia (). ‘Reduced relatives in DPs: evidence from adjective extraction in modern Greek’, MIT Working Papers in Linguistics : –. Antonelli, André (). Sintaxe de posição do verbo e mudança gramatical na história do português europeu, PhD dissertation, University of Campinas, São Paulo, Brazil. Antonelli, André (). ‘Orações subordinadas sem o complementizador Que no português clássico’. D.E.L.T.A. : –. Bacskai-Atkari, Julia (). ‘A komparatív operátor esete a mondatbevezetővel: szintaktikai változások a magyar hasonlító mellékmondatokban’ [The comparative operator’s affair with the complementiser: syntactic changes in Hungarian comparative subclauses], in Katalin É. Kiss and Attila Hegedűs (eds), Nyelvelmélet és diakrónia [Language theory and diachrony]. Budapest/Piliscsaba: Szent István Társulat, –. Bacskai-Atkari, Julia (). The Syntax of Comparative Constructions: Operators, Ellipsis Phenomena and Functional Left Peripheries. Potsdam, Germany: Universitätsverlag Potsdam. Bailey, Natalie, Guy Maynor, and Patricia Cukor-Avila (). The Emergence of Black English: Text and Commentary. Amsterdam/Philadelphia: John Benjamins. Baker, Mark (). ‘Mirror theory and morphosyntactic explanation’, Linguistic Inquiry : –. Baker, Mark C. (). Incorporation: A Theory of Grammatical Function Changing. Chicago: University of Chicago Press. Baker, Mark C. (). ‘Building and merging, not checking: the non-existence of (Aux)-S-V-O languages’. Linguistic Inquiry : –. Baker, Mark C. (). Lexical Categories: Verbs, Nouns and Adjectives. Cambridge: Cambridge University Press. Baker, Mark C. (). ‘The macroparameter in a microparametric world’, in Theresa Biberauer (ed.), The Limits of Syntactic Variation. Amsterdam/Philadelphia: John Benjamins, –. Bakró-Nagy, Marianne (). ‘A magyar határozott névelő kialakulásának szintaktikai vonatkozásai’ [Syntactic aspects of the development of the Hungarian definite article], in
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Büky, László, and Tamás Forgács (eds), A nyelvtörténeti kutatások újabb eredményei I. Magyar és finnugor mondattörténet. Szeged, Hungary: JATE Magyar Nyelvészeti Tanszék, –. Bartos, Huba (). ‘Az inflexiós jelenségek szintaktikai háttere’ [The syntactic background of the inflectional phenomena], in Ferenc Kiefer (ed.), Strukturális magyar nyelvtan . Morfológia. Budapest: Akadémiai Kiadó, –. Bartos, Huba (). ‘Mutató névmási módosítók a magyar főnévi szerkezetben: egyezés vagy osztozás?’ [Demonstrative modifiers in the Hungarian noun phrase: agreement or share], in Marianne Bakró-Nagy, Zoltán Bánréti, and Katalin É. Kiss (eds), Újabb tanulmányok a strukturális magyar nyelvtan és a nyelvtörténet köréből. Budapest: Osiris, –. Bašić, Monika (). Nominal Subextractions and the Structure of NPs in Serbian and English. MA dissertation, University of Tromsø, Norway. Basilico, David (). ‘Head position and internally headed relative clauses’, Language : –. Batllori, Montse, Maria-Lluïsa Hernanz, Carme Picallo, and Francesc Roca, eds (). Grammaticalization and Parametric Variation. Oxford/New York: Oxford University Press. Battye, Adrian, and Ian Roberts (eds) (). Clause Structure and Language Change. Oxford/ New York: Oxford University Press. Bauer, Brigitte (). The Emergence and Development of SVO Patterning in Latin and French. Oxford/New York: Oxford University Press. Bauer, Brigitte (). ‘Word order’, in Philip Baldi and Pierluigi Cuzzolin (eds), New Perspectives on Historical Latin Syntax. Volume : Syntax of the Sentence. Berlin: De Gruyter, –. Bauer, Laurie (). Morphological Productivity. Cambridge: Cambridge University Press. Baunaz, Lena (). The Grammar of French Quantification. Dordrecht: Springer. Bayer, Joseph (). ‘Nominal Negative Quantifiers as Adjuncts’, The Journal of Comparative Germanic Linguistics : –. Beekes, Robert S. P. (). Etymological Dictionary of Greek, Leiden Indo-European Etymological Dictionary Series. Leiden: E. J. Brill. Belletti, Adriana (). Generalized Verb Movement. Turin: Rosenberg and Sellier. Belletti, Adriana (). ‘Inversion as focalization’, in Aafke Hulk and Jean-Yves Pollock (eds), Subject Inversion in Romance and the Theory of Universal Grammar. Oxford/New York: Oxford University Press, –. Belletti, Adriana (). ‘Aspects of the low IP area’, in Luigi Rizzi (ed.), The Structure of CP and IP: The Cartography of Syntactic Structures, Volume . Oxford/New York: Oxford University Press, –. Benincà, Paola, and Cecilia Poletto (). ‘Topic, focus, and V. Defining the CP sublayers’, in Luigi Rizzi (ed.), The Structure of CP and IP: The Cartography of Syntactic Structures, Volume . Oxford/New York: Oxford University Press, –. Benincà, Paola (). ‘A detailed map of the left periphery of medieval romance’, in Raffaella Zanuttini, Héctor Campos, Elena Herburger, and Paul Portner (eds), Negation, Tense and Clausal Architecture: Cross-linguistics Investigations. Washington, DC: Georgetown University Press, –. Bergin, Osborn (). ‘On the syntax of the verb in Old Irish,’ Ériu : –. Bergin, Osborn J., Richard Irvine Best, Kuno Meyer, and James George O’Keeffe (eds) (–). Anecdota from Irish Manuscripts. volumes. Halle: Max Niemeyer.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Bernstein, Judy B., and Raffaella Zanuttini (). ‘A Diachronic Shift in the Expression of Person’, in Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar (eds), Parameter Theory and Linguistic Change. Oxford/New York: Oxford University Press, –. Bertinetto, Pier Marco, and Valentina Bianchi (). ‘Tense, aspect, and syntax: a review of Giorgi and Pianesi ()’. Linguistics : –. Berwick, Robert C., and Noam Chomsky (). ‘The biolinguistic program: the current state of its development’, in Anna Maria Di Sciullo and Cedric Boeckx (eds) The Biolinguistic Enterprise: New Perspectives on the Evolution and Nature of the Human Language Faculty. Oxford/New York: Oxford University Press, –. Bianchi, Valentina (). Consequences of Antisymmetry: Headed Relative Clauses. Berlin: De Gruyter. Biberauer, Theresa (). Verb Second (V) in Afrikaans: A Minimalist Investigation of Word Order Variation. PhD dissertation, University of Cambridge. Biberauer, Theresa, Anders Holmberg, and Ian Robert (). ‘Structure and linearization in disharmonic word orders’, in Charles Chang, and Hannah Haynie (eds), Proceedings of the th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Proceedings Project, –. Biberauer, Theresa, Anders Holmberg, and Ian Roberts (). ‘Linearization and the architecture of grammar: a view from the final-over-final constraint’. StiL—Studies in Linguistics (Proceedings of XXXV Incontro di Grammatica Generativa) : –. Biberauer, Theresa, Anders Holmberg and Ian Roberts (). ‘A syntactic universal and its consequences’, Linguistic Inquiry : –. Biberauer, Theresa, Glenda Newton, and Michelle Sheehan (). ‘Limiting synchronic and diachronic variation and change: the final-over-final constraint’, Language and Linguistics : –. Biberauer, Theresa, and Ian Roberts (). ‘Changing EPP parameters in the history of English: accounting for variation and change’, in William van der Wurff (ed), Journal of English Language and Linguistics (Special Issue on Word Order Change) : –. Biberauer, Theresa, and Ian Roberts (). ‘Cascading parameter changes: internally-driven change in Middle and Early Modern English’, in Þorhallur Eyþórsson (ed.), Grammatical Change and Linguistic Theory. Amsterdam/Philadelphia: John Benjamins, –. Biberauer, Theresa, and Ian Roberts (). ‘Subjects, tense and verb movement’, in Theresa Biberauer, Anders Holmberg, Ian Roberts, and Michelle Sheehan (eds), Parametric Variation: Null Subjects in Minimalist Theory. Cambridge: Cambridge University Press, –. Biberauer, Theresa, and Ian Roberts (). ‘Conditional inversion and types of parametric change’, in Bettelou Los and Pieter de Haan (eds), Verb-Second Languages: Essays in Honour of Ans van Kemenade. Amsterdam/Philadelphia: John Benjamins, –. Biberauer, Theresa, and Michelle Sheehan (). ‘Disharmony, antisymmetry, and the finalover-final constraint’, in Valmala Elguea and Myriam Extebarria (eds), Ways of Structure Building. Cambridge: Cambridge University Press, –. Biberauer, Theresa, and George Walkden (eds) (). Syntax Over Time: Lexical, morphological and information-structural interactions. Oxford/New York: Oxford University Press. Birner, Betty J. (). ‘Recency effects in English inversion’. in Marilyn Walker, Aravind K. Joshi, and Ellen F. Prince (eds), Centering Theory in Discourse. Oxford/New York: Oxford University Press, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Bisagni, Jacopo (). ‘Prolegomena to the study of code-switching in the Old Irish Glosses’. Peritia – [–]: –. Blümel, Rudolf (). Einführung in die Syntax, Heidelberg: C. Winter. Bobaljik, Jonathan David, and Dianne Jonas (). ‘Subject positions and the roles of TP’, Linguistic Inquiry : –. Bobaljik, Jonathan D. (). ‘Realizing Germanic inflection: why morphology does not drive syntax’. Journal of Comparative Germanic Languages : –. Bobaljik, Jonathan D., and Susi Wurmbrand (). ‘The domain of agreement’. Natural Language and Linguistic Theory : –. Boisacq, Emile (). Dictionnaire étymologique de la langue grecque étudiée dans ses rapports avec les autres langues indo-européennes. Heidelberg: Carl Winter Universitätsverlag. Bošković, Željko (). ‘D-structure, θ-criterion, and movement into θ-positions’, Linguistic Analysis : –. Bošković, Željko (). The Syntax of Nonfinite Complementation: An Economy Approach. Cambridge, MA: MIT Press. Bošković, Željko (). On the Nature of the Syntax-Phonology Interface: Cliticization and Related Phenomena. Amsterdam: Elsevier Science. Bošković, Željko (). ‘On multiple wh-fronting’, Linguistic Inquiry : –. Bošković, Željko (). ‘On left branch extraction’, in Peter Kosta, Joanna Błaszczak, Jens Frasek, Ljudmila Geist, and Marzena Żygis (eds), Investigations into Formal Slavic Linguistics. Berlin: Peter Lang, –. Bošković, Željko (a). ‘On the clitic switch in Greek imperatives’, in Olga Tomić (ed.), Balkan Syntax and Semantics. Amsterdam/Philadelphia: John Benjamins, –. Bošković, Željko (b). ‘PF merger in stylistic fronting and object shift’, in Arthur Stepanov, Gisbert Fanselow, and Ralf Vogel (eds), Minimality Effects in Syntax. Berlin: De Gruyter, –. Bošković, Željko (c). ‘Topicalization, focalization, lexical insertion, and scrambling’, Linguistic Inquiry : –. Bošković, Željko (a). ‘On the locality of left branch extraction and the structure of NP’, Studia Linguistica : –. Bošković, Željko (b). ‘Left branch extraction, structure of NP, and scrambling’, in Joachim Sabel, and Mamoru Saito (eds), The Free Word Order Phenomenon: Its Syntactic Sources and Diversity. Berlin: De Gruyter, –. Bošković, Željko (). ‘What will you have, DP or NP?’, in Emily Elfner and Martin Walkow (eds), Proceedings of the Thirty-Ninth Annual North East Linguistic Society. University of Massachusetts, Amherst: GLSA, –. Bošković, Željko (). ‘More on the no-DP analysis of article-less languages’, Studia Linguistica : –. Bošković, Željko, and Jon Gajewski (). ‘Semantic correlates of the NP/DP parameter’, in Suzi Lima, Kevin Mullin, and Brian Smith (eds), Proceedings of the Thirty-Ninth Annual North East Linguistic Society. University of Massachusetts, Amherst: GLSA, –. Bošković, Željko, and Jairo Nunes (). ‘The copy theory of movement: a view from PF’, in Norbert Corver, and Jairo Nunes (eds), The Copy Theory of Movement. Amsterdam/ Philadelphia: John Benjamins, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Breitbarth, Anne, and Liliane Haegeman (). ‘Not continuity, but change: stable Stage II in Jespersen’s Cycle’. Manuscript. University of Cambridge; Ghent University, Belgium. Breitbarth, Anne, and Henk van Riemsdijk (eds) (). Triggers. Berlin: De Gruyter. Brito, Ana Maria (). ‘Clause structure, subject positions and verb movement: about the position of sempre in European and Brazilian Portuguese’, in Yves D’Hulst, Johan Rooryck, and Jan Schroten (eds), Romance Languages and Linguistic Theory : Selected Papers from ‘Going Romance’ , Leiden, – December . Amsterdam/Philadelphia: John Benjamins, –. Broekhuis, Hans, and Veronika Hegedűs (). ‘Predicate movement’. Lingua : –. Brunet, Étienne, and Sylvie Mellet (n.d.). ‘Hyperbase .: Logiciel hypertexte pour le traitement documentaire et statistique des corpus textuels’. Base de littérature latine. Bases, Corpus & Langage (Université de Nice—Sophia Antipolis) and LASLA (Université de Liège, Belgium). Bures, Anton (). ‘(Re)-cycling expletives and other sentences’. Generals papers, Massachusetts Institute of Technology, Cambridge, MA. Butler, Alastair, and Eric Mathieu (). The Syntax and Semantics of Split Constructions: A Comparative Study. Basingstoke: Palgrave Macmillan. Campos, Odette G. L. A. de Souza, and Ângela Rodrigues (). ‘A flexão verbal modotemporal no português culto do Brasil: sufixos e auxiliares’, in Ataliba Teixeira de Castilho, (ed.), Gramática do Português Falado. Volume : As Abordagens. Campinas, São Paulo: Editora Unicamp, –. Cardinaletti, Anna (). ‘A second thought on «emarginazione»: destressing vs. «right dislocation»’, in Guglielmo Cinque and Giampaolo Salvi (eds), Current Studies in Italian Syntax. Essays Offered to Lorenzo Renzi. Amsterdam: Elsevier, –. Cardinaletti, Anna (). ‘Towards a cartography of subject positions’, in Luigi Rizzi (ed.), The Structure of CP and IP: The Cartography of Syntactic Structures, Volume . Oxford/New York: Oxford University Press, –. Cardinaletti, Anna, and Michael Starke (). ‘The typology of structural deficiency: a case study of the three classes of pronouns’, in Henk van Riemsdijk (ed.), Clitics in the Languages of Europe. Berlin: De Gruyter, –. Cardoso, Adriana (). Variation and Change in the Syntax of Relative Clauses: New Evidence from Portuguese. PhD dissertation, University of Lisbon. Castilho, Ataliba Teixeira de (). Nova Gramática do Português Brasileiro. São Paulo, Brazil: Editora Contexto. Castillo, Juan Carlos, John E. Drury, and Kleanthes K. Grohmann (). ‘Merge over move and the extended projection principle: MOM and the EPP revisited’. Iberia: An International Journal of Theoretical Linguistics .: –. Cecchetto, Carlo, and Renato Oniga (). ‘Consequences of the analysis of Latin infinitival clauses for the theory of case and control’, Lingue e Linguaggio : –. Chesnutt, Charles, and Werner Sollors (). Charles W. Chesnutt: Stories, Novels and Essays. New York: Library of America. Chomsky, Noam (). ‘On wh-movement’, in Peter Culicover, Thomas Wasow, and Adrian Akmajian (eds), Formal Syntax. New York: Academic Press, –. Chomsky, Noam (). Lectures on Government and Binding. The Pisa Lectures. Dordrecht: Foris. Chomsky, Noam (). ‘A minimalist program for linguistic theory’, in Kenneth Hale, and Samuel J. Keyser (eds) The View from Building : Essays in Linguistics in Honor of Sylvain Bromberger. Cambridge, MA: MIT Press, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Chomsky, Noam (). The Minimalist Program. Cambridge, MA: MIT Press. Chomsky, Noam (). ‘Minimalist inquiries: the framework’, in Roger Martin, David Michaelis, and Juan Uriagereka (eds), Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik. Cambridge, MA: MIT Press, –. Chomsky, Noam (). ‘Derivation by phase’, in Michael Kenstowicz (ed.), Ken Hale: A Life in Linguistics. Cambridge, MA: MIT Press, –. Chomsky, Noam (). ‘Three factors in language design’. Linguistic Inquiry : –. Chomsky, Noam (). ‘On phases’, in Robert Friedin, Carlos P. Otero, and María Luisa Zubizarreta (eds), Foundational Issues in Linguistic Theory: Essays in Honour of Jean-Roger Vergnaud. Cambridge, MA: MIT Press, –. Cinque, Guglielmo (). ‘On the evidence for partial N movement in the Romance DP’, in Guglielmo Cinque, Jan Koster, Jean-Yves Pollock, Luigi Rizzi, and Raffaella Zanuttini (eds), Paths towards Universal Grammar: Studies in Honor of Richard S. Kayne. Washington, DC: Georgetown University Press, –. Cinque, Guglielmo (). Adverbs and Functional Heads: A Cross-linguistic Perspective. Oxford/New York: Oxford University Press. Cinque, Guglielmo (). ‘A note on mood, modality, tense and aspect suffixes in Turkish’, in Eser Erguvanlı Taylan (ed.), The Verb in Turkish. Amsterdam/Philadelphia: John Benjamins, –. Cinque, Guglielmo (). ‘Restructuring and functional structure’, in A. Belletti (ed.), Structures and Beyond. The Cartography of Syntactic Structures, Volume . Oxford/New York: Oxford University Press, –. Cinque, Guglielmo (). ‘Deriving Greenberg’s universal and its exceptions’, Linguistic Inquiry : –. Clackson, James, and Geoffrey Horrocks (). The Blackwell History of the Latin Language. Oxford/New York: Blackwell. Clark, Robin, and Ian Roberts (). ‘A computational approach to language learnability and language change’. Linguistic Inquiry : –. Coelho, António Borges (). Inquisição de Évora: Dos primórdios a . Volume . Lisboa: Caminho. Conradie, Simone (). ‘The setting of the Split-IP parameter in Afrikaans: evidence from transitive expletive constructions and object shift constructions’, Stellenbosch Papers in Linguistics PLUS , –. Cook, Philippa, and Felix Bildhauer (). ‘Annotating information structure: the case of topic’, in Stephanie Dipper and Heike Zinsmeister (eds), Beyond Semantics: Corpus based Investigations of Pragmatic and Discourse Phenomena (Proceedings of the DGfS Workshop Göttingen, February –, ). Ruhr Universität Bochum, Bochumer Linguistische Arbeitsberichte, –. Cormack, Annabel, and Neil V. Smith (). ‘Fronting: the syntax and pragmatics of “focus” and “topic” ’. UCL Working Papers in Linguistics , –. Corver, Norbert (). ‘Left branch extraction’, in NELS : Proceedings of the nd Annual Meeting of the North East Linguistic Society, University of Delaware, Newark, –. Costa, João (). Word Order Variation: A Constraint-based Approach. The Hague: Holland Academic Graphics.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Costa, João (). Subject Positions and the Interfaces: The Case of European Portuguese. Berlin: De Gruyter. Costa, João, and Ana Maria Martins (). ‘Middle scrambling with deictic locatives in European Portuguese’. in Reineke Bok-Bennema, Brigitte Kampers-Manhe, and Bart Hollebrandse (eds), Romance Languages and Linguistic Theory : Selected Papers from ‘Going Romance’, Groningen . Amsterdam/Philadelphia: John Benjamins, –. Costa, João, and Ana Maria Martins (). ‘On focus movement in European Portuguese’, Probus .: –. Crisma, Paola, and Giuseppe Longobardi (eds) (). Historical Syntax and Linguistic Theory. Oxford/New York: Oxford University Press. Cruschina, Silvio (). Discourse-Related Features and Functional Projections. Oxford/New York: Oxford University Press. Cyrino, Sonia (). ‘On richness of tense and verb movement in Brazilian Portuguese’, in Victoria Camacho-Taboada, Ángel L. Jiménez-Fernández, Javier Martín-González, and Mariano Reyes-Tejedor (eds), Information Structure and Agreement. Amsterdam/Philadelphia: John Benjamins, –. D. Mátai Mária (). ‘Az igekötők’ [Particles], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana I. A korai ómagyar kor és előzményei [The historical grammar of Hungarian. Volume : Early Old Hungarian and before]. Budapest: Akadémiai Kiadó, –. D. Mátai Mária (). ‘Az igekötők’ [Particles], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana II/. A kései ómagyar kor. Morfematika [The historical gramamr of Hungarian. Volume /: Late Old Hungarian. Morphemics]. Budapest: Akadémiai Kiadó, –. D. Mátai Mária (). ‘Szófajtörténet’ [The history of word classes], in Jenő Kiss and Ferenc Pusztai (eds) Magyar nyelvtörténet [Hungarian language history]. Budapest: Osiris Kiadó, –. Danckaert, Lieven (). On the Left Periphery of Latin Embedded Clauses. PhD dissertation, Ghent University, Belgium. Danckaert, Lieven (). Latin Embedded Clauses: The Left Periphery. Amsterdam/Philadelphia: John Benjamins. Danckaert, Lieven (). ‘The derivation of Classical Latin Aux-final clauses: implications for the internal structure of the verb phrase’, in Karen Lahousse and Stefania Marzo (eds), Romance Languages and Linguistic Theory : Selected Papers from ‘Going Romance’ Leuven . Amsterdam/Philadelphia: John Benjamins, –. Danckaert, Lieven (). The Development of Latin Clause Structure. A Study of the Extended Verb Phrase. Oxford/New York: Oxford University Press. De Haan, Germen J. (). ‘De friese imperativus-pro-infinitivo’. Taal en Tongval Themanummer Dialectsyntaxis : –. De Haan, Germen J. (). Studies in West Frisian Grammar: Selected Papers by Germen J. De Haan, ed. Jarich Hoekstra, Willem Visser, and Goffe Jensma. Amsterdam/Philadelphia: John Benjamins. De Vries, Mark (). The Syntax of Relativization. Utrecht: LOT. Demirdache, Hamida (). Resumptive Chains in Restrictive Relatives, Appositives, and Dislocation Structures. PhD dissertation, Massachusetts Institute of Technology, Cambridge, MA.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Demirdache, Hamida, and Myriam Uribe-Etxebarria (). ‘The primitives of temporal relations’, in Roger Martin, David Michaels, and Juan Uriagereka (eds), Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik. Cambridge, MA: MIT Press, –. Den Besten, Hans (). ‘On the interaction of root transformations and lexical deletive rules’, in Werner Abraham (ed.), On the Formal Syntax of the Westgermania. Papers from the ‘rd Groningen Grammar Talks’ Groningen, January . Amsterdam/Philadelphia: John Benjamins, : –. Den Dikken, Marcel (). Particles: On the Syntax of Verb-Particle, Triadic, and Causative Constructions. Oxford/New York: Oxford University Press. Den Dikken, Marcel (a). Relators and Linkers: The Syntax of Predication, Predicate Inversion, and Copulas. Cambridge, MA: MIT Press. Den Dikken, Marcel (b). ‘Either-float and the syntax of co-or-dination’, Natural Language and Linguistic Theory .: –. Denison, David. (). ‘The origins of periphrastic DO: Ellegård and Visser reconsidered’, in Roger Eaton, Olga Fischer, Willem Koopman, and Frederike van der Leek (ed.), Papers from the th International Conference on English Historical Linguistics. Amsterdam/Philadelphia: John Benjamins, –. Derksen, Rick (). Etymological Dictionary of the Slavic Inherited Lexicon. Volume of Leiden Indo-European Etymological Dictionary. Leiden: E. J. Brill. Devine, Andrew, and Laurence Stephens (). Latin Word Order: Structured Meaning and Information. Oxford/New York: Oxford University Press. Dias, Augusto Epifânio da Silva ( []). Sintaxe Histórica Portuguesa. Lisbon: Clássica Editora. Dictionary of the Scots Language (n.d.) Scottish Language Dictionaries Ltd. Dobrovie Sorin, Carmen (). The Syntax of Romanian. Berlin: De Gruyter. Doetjes, Jenny (). Quantifiers and Selection: On the Distribution of Quantifying Expressions in French, Dutch and English. The Hague: HAG. Doherty, Cathal (). ‘Residual verb second in Early Irish: on the nature of Bergin’s construction’. Diachronica : –. Dömötör, Adrienne (). ‘A főnévi névmási kijelölő jelző a középmagyar korban’ [The demonstrative modifier in the Middle Hungarian period], in László Büky, Tamás Forgács, and Balázs Sinkovics (eds), A nyelvtörténeti kutatások újabb eredményei V. Szeged, Hungary: Szegedi Tudományegyetem Magyar Nyelvészeti Tanszék, –. Dömötör, Éva, and Diána Varga (). Főmondati kötőmód diskurzuspartikulákkal [Matrix subjunctive with discourse particles], in Zsuzsanna Gécseg (ed.), Lingdok . Nyelvészdoktoranduszok dolgozatai. Szegedi Tudományegyetem Nyelvtudományi Doktori Iskola, Szeged, Hungary, –. Don, Jan, Paula Fenger, and Olaf Koeneman (). ‘Restricting language change through micro-comparative analysis’. Talk presented at GLOW, Lund. Dressler, Wolfgang U. (). ‘Degrees of grammatical productivity in inflectional morphology’. Rivista di Linguistica : –. Dryer, Matthew S. (). ‘The Greenbergian word order correlations’, Language : –. Dryer, Matthew S. (). ‘The branching direction theory of word order correlations revisited’, in Sergio Scalise, Elisabetta Magni, and Antonietta Bisetto (eds), Universals of Language Today. Berlin: Springer, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Dryer, Matthew S. (). ‘Position of tense–aspect affixes’, in Matthew S. Dryer and Martin Haspelmath (eds), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Duarte, Inês (). ‘Algumas reflexões sobre ordem de palavras em línguas românicas de sujeito nulo’, in Ana Maria Brito, Isabel Margarida Duarte, and Joana Guimarães (eds), Língua Portuguesa: Estruturas, Usos e Contrastes. Porto, Portugal: CLUP, –. Dunkel, George E. (). ‘IE conjunctions: pleonasm, ablaut, suppletion’, Zeitschrift für vergleichende Sprachforschung : –. É. Kiss, Katalin (). ‘Discourse configurationality in the languages of Europe’. Constituent Order in the Languages of Europe. Berlin: De Gruyter, –. É. Kiss, Katalin (). ‘Identificational focus versus information focus’, Language : –. É. Kiss, Katalin (). ‘The Hungarian noun phrase is like the English noun phrase’, in Alberti, Gábor, and István Kenesei (eds), Approaches to Hungarian . Szeged, Hungary: JATE Press, –. É. Kiss, Katalin (). The Syntax of Hungarian. Cambridge: Cambridge University Press. É. Kiss, Katalin (ed.) (). Event Structure and the Left Periphery: Studies on Hungarian. Dordrecht: Springer. É. Kiss, Katalin (). ‘Is free postverbal order in Hungarian a syntactic or a PF phenomenon?’, in Nomi Erteschik-Shir, and Lisa Rochman (eds) The Sound Patterns of Syntax. Oxford/New York: Oxford University Press, –. É. Kiss, Katalin (). ‘Az ősmagyar SOV-tól az ómagyar (TQF)VSO-ig’ [From ProtoHungarian SOV to Old Hungarian (TQF)VSO], in Katalin É. Kiss and Attila Hegedűs (eds), Nyelvelmélet és diakrónia [Language theory and diachrony]. Piliscsaba: PPKE BTK elméleti Nyelvészeti Tanszék—Magyar Nyelvészeti Tanszék, –. É. Kiss, Katalin (). ‘From Proto-Hungarian SOV to Old Hungarian Top Foc V X*’, Diachronica : –. É. Kiss, Katalin (a). ‘The evolution of functional left peripheries in the Hungarian sentence’, in Katalin É. Kiss (ed.), The Evolution of Functional Left Peripheries in Hungarian Syntax. Oxford/New York: Oxford University Press, –. É. Kiss, Katalin (b). ‘Ways of licensing Hungarian external possessors’, Acta Linguistica Hungarica : –. Egedi, Barbara (). ‘Grammatical encoding of referentiality in the history of Hungarian’, in Anna Giacolone Ramat, Caterina Mauri, and Piera Molinelli (eds), Synchrony and Diachrony: A Dynamic Interface. Amsterdam/Philadelphia: John Benjamins, –. Egedi, Barbara (a). ‘The DP-cycle in Hungarian and the functional extension of the noun phrase’, in É. Kiss Katalin (ed.), The Evolution of Functional Left Peripheries in Hungarian Syntax. Oxford/New York: Oxford University Press, –. Egedi, Barbara (b). ‘A latin forrásszövegek hatása az ómagyar szintaktikai jelenségekre’ [The influence of Latin sources on Old Hungarian syntax], in É. Kiss Katalin (ed.), Magyar generatív történeti mondattan. Budapest: Akadémiai Kiadó: –. Egedi, Barbara, and Eszter Simon (). ‘Gradual expansion in the use of the definite article: Checking a theory against the Old Hungarian Corpus’. Talk given at the conference ‘Exploring Ancient Languages through Corpora’, June , Oslo.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Eide, Kristine (). Word Order Structures and Unaccusative Verbs in Classical and Modern Portuguese: The Reorganisation of Information Structure. PhD dissertation, University of Oslo, Norway. Eide, Kristine (). ‘Prosody, information structure and word order changes in Portuguese’, in G. Ferraresi and Rosemarie Luhr (eds), Diachronic Studies on Information Structure: Language Acquisition and Change. Berlin: De Gruyter, –. Elenbaas, Marion (). The Synchronic and Diachronic Syntax of the English Verb-Particle Combination. PhD dissertation, Radboud University, Nijmegen, Netherlands. Ellegård, Alvar (). ‘The auxiliary do: the establishment and regulation of its use in English’, in Frank Behre (ed.), Gothenburg Studies in English. Stockholm: Almqvist and Wiksell. Emonds, Joseph (). A Unified Theory of Syntactic Categories. Dordrecht: Foris. Enkvist, Nils Erik (). ‘Marked focus: function and constraints’, in Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik (eds), Studies in English Linguistics for Randolph Quirk. London: Longman, –. Ernout, Alfred, and Antoine Meillet (). Dictionnaire étymologique de la langue latine: histoire des mots (retirage de la e édition augmentée d’additions et de corrections par Jacques André). Paris: Klincksieck. Eska, Joseph (). ‘Bergin’s Rule’. Diachronica : –. Fanselow, Gisbert (). ‘Aufspaltung von NPn und das Problem der “freien” Wortstellung’, Linguistische Berichte : –. Fanselow, Gisbert, and Caroline Féry (). ‘Prosodic and morphosyntactic aspects of discontinuous noun phrases: a crosslinguistic perspective’. Manuscript. University of Postdam; Goethe-Universität Frankfurt. Fanselow, Gisbert, and Damir Ćavar (). ‘Distributed deletion’, in Artemis Alexiadou (ed.), Theoretical Approaches to Universals. Amsterdam/Philadelphia: John Benjamins, –. Farkas, Donka, and Henriette de Swart (). The Semantics of Incorporation: From Argument Structure to Discourse Transparency. Stanford, CA: CSLI. Feagin, Crawford (). Variation and Change in Alabama English: A Sociolinguistic Study of the White Community. Washington, DC: Georgetown University Press. Fiéis, Alexandra (). ‘(XP)VS em português medieval’. in Maria Helena Mira Mateus, Anabela Gonçalves, and Clara Nunes Correia (eds), Actas do XVII Encontro Nacional da Associação Portuguesa de Linguística. Lisbon: Colibri/APL, –. Fiéis, Alexandra (). ‘On the position of sempre in Medieval Portuguese and in Modern European Portuguese’. The Linguistic Review : –. Fiengo, Robert (). ‘On trace theory’, Linguistic Inquiry : –. Fitzgerald, Colleen M. (). ‘An optimality treatment of syntactic inversions in English verse.’ Language Sciences : –. Foreman John, (). ‘Syntax of negative inversion in non-standard English’, in Shahin, Kimary N., Blake, Susan, and Kim, Eun-Sook (eds), in Kimary N. Shahin, Susan Blake, and Eun-Sook Kim (eds), WCCFL : Proceedings of the th West Coast Conference on Formal Linguistics. Stanford, CA: CSLI. Forrer, Emil (–). Die Boǧazköy Texte im Umschrift. Leipzig: J.C. Hinrich. Fowlie, Meaghan (). ‘Multiple multiple spellout’, in Theresa Biberauer and Ian Roberts (eds), Challenges to Linearization. Berlin: De Gruyter, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Franks, Steven, and Ljiljana Progovac (). ‘On the displacement of Serbo-Croatian clitics’, Indiana Linguistic Studies : –. Frascarelli, Mara, and Roland Hinterhölzl (). ‘Types of topics in German and Italian’, in Suzanne Winkler and Kerstin Schwabe (eds), On Information Structure, Meaning and Form. Amsterdam/Philadelphia: John Benjamins, –. Frota, Sónia (). Prosody and Focus in European Portuguese. PhD dissertation, University of Lisbon. Frota, Sónia (). ‘Nuclear falls and rises in European Portuguese: a phonological analysis of declarative and question intonation’, Probus : –. Frota, Sonia, Charlotte Galves, Marina Vigário, Verónica González-López, and Maria Bernadete Abaurre (). ‘The phonology of rhythm from Classical to Modern Portuguese’. Journal of Historical Linguistics : –. Fruyt, Michèle (). ‘Grammaticalisation in Latin’, in Philip Baldi and Pierluigi Cuzzolin (eds), New Perspectives on Historical Latin Syntax. Volume : Complex Sentences, Grammaticalization, Typology. Berlin: De Gruyter, –. Galambos, Dezső (). ‘Tanulmányok a magyar relatívum mondattanáról’. [Studies on the syntax of Hungarian relatives]. Budapest: Athenaeum. Gallasy, Magdolna (). ‘A névelők’ [The articles], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana II/. A kései ómagyar kor. Morfematika. Budapest: Akadémiai Kiadó, –. Galves, Antonio, and Charlotte Galves (). ‘A case-study of prosody-driven language change: from Classical to Modern European Portuguese’. Manuscript. University of São Paulo, University of Campinas, Brazil. Galves, Charlotte (). ‘Clitic-placement and parametric change in Portuguese’, in Claudia Parodi, Carlos Quicoli, Mario Saltarelli, and María Luisa Zubizarreta (eds). Aspects of Romance Linguistics: Selected Papers from the th Linguistic Symposium on Romance Languages, . Georgetown University Press, –. Galves, Charlotte (). ‘Agreement, predication and pronouns in the history of Portuguese’, in João Costa (ed.), Portuguese Syntax: New Comparative Studies. Oxford/New York: Oxford University Press, –. Galves, Charlotte, Helena Britto, and Maria Clara Paixão de Sousa (). ‘The change in clitic placement from Classical to Modern European Portuguese: results from the Tycho Brahe Corpus’. Journal of Portuguese Linguistics : –. Galves, Charlotte, and Maria Clara Paixão de Sousa (). ‘Clitic placement and the position of subjects in the history of Portuguese’, in Twan Geerts, Ivo van Ginneken, and Haike Jacobs (eds), Romance Languages and Linguistic Theory: Selected Papers from ‘Going Romance’ , Nijmegen – November. Amsterdam/Philadelphia: John Benjamins, –. Galves, Charlotte, and Maria Clara Paixão de Sousa (). ‘The change in the position of the verb in the history of Portuguese: subject realization, clitic placement, and prosody’. Language : e–e (online only). Galves, Charlotte, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar (eds) (). Parameter Theory and Linguistic Change. Oxford/New York: Oxford University Press. Garzonio, Jacopo, and Cecilia Poletto (). ‘How bare are bare quantifiers? Some notes from diachronic and synchronic variation in Italian’, Linguistic Variation : –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Geldner, Karl Friedrich (). Avesta, the Sacred Books of the Parsis. Stuttgart, Germany: W. Kohlhammer. Gergel, Remus (). ‘Towards notions of comparative continuity in English and French’, in Anne Breitbarth, Christopher Lucas, Sheila Watts, and David Willis (eds), Continuity and Change in Grammar. Amsterdam/Philadelphia: John Benjamins, –. Gibrail, Alba. (). Contextos de formação de estruturas de tópico e foco no português clássico. PhD dissertation, University of Campinas, São Paulo, Brazil. Giorgi, Alexandra, and Fabio Pianesi (). Tense and Aspect: From Semantics to Morphosyntax. Oxford/New York: Oxford University Press. Giusti, Giuliana (). ‘The birth of a functional category: from Latin ILLE to the Romance article and personal pronoun’, in Guglielmo Cinque, and Giampaolo Salvi (eds), Current Studies in Italian Syntax: Essays Offered to Lorenzo Renzi. Amsterdam: Elsevier, –. Giusti, Giuliana, and Renato Oniga (). ‘La struttura del sintagma nominale latino’, in Renato Oniga and Luigi Zennaro (eds), Atti della Giornata di Linguistica Latina, Venezia, maggio . Venice: Cafoscarina, –. Giusti, Giuliana, and Renato Oniga (). ‘Core and periphery in the Latin noun phrase’, in Gérald Purnelle and Joseph Denooz (eds), Ordre et cohérence en latin: communications présentées au e colloque international de linguistique latine (Bruxelles-Liège, – avril ). Geneva: Droz, –. Giusti, Giuliana, and Iovino, Rossella (). ‘Evidence for a Split DP in Latin’, University of Venice Working Papers in Linguistics : –. Goldstein, David M. (). ‘Variation versus change: clausal clitics between Homer and Herodotus’, Indo-European Linguistics : –. Gonda, Jan (). ‘The history and original function of the Indo-European particle *kwe, especially in Greek and Latin’, Mnemosyne .: –. Gonzaga, Manuela (). Aspectos da sintaxe do advérbio em português. MA dissertation, University of Lisbon. Götze, Michael, Thomas Weskott, Cornelia Endriss, Ines Fiedler, Stefan Hinterwimmer, Svetlana Petrova, Anne Schwarz, Stavros Skopeteas, and Ruben Stoel (). ‘Information structure.’, in Stefanie Dipper, Michael Götze, and Stavros Skopeteas (eds), Interdisciplinary Studies on Information Structure, volume of Working Papers of the SFB . Potsdam, Germany: Universitätsverlag Potsdam, –. Green, Lisa (). African American English: A Linguistic Introduction. Cambridge: Cambridge University Press. Green, Lisa (). ‘Force, focus, and negation in African American English’, in Zanuttini, Raffaella and Laurence Horn (eds), Micro-Syntactic Variation in North American English. Oxford/New York: Oxford University Press, –. Greenberg, Joseph H. (). Universals of Language. Cambridge, MA: MIT Press. Grewendorf, Günther, and Cecilia Poletto (). ‘Von OV zu VO: ein Vergleich zwischen Zimbrisch und Plodarisch’, in Ermenegildo Bidese, James Dow, and Thomas Stolz (eds), Das Zimbrische zwischen Germanisch und Romanisch. Diversitas Linguarum vol. . Bochum, Germany: Brockmeyer Verlag, –. Griffith, Aaron (). ‘The animacy hierarchy and the distribution of the notae augentes in Old Irish’. Ériu : –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Griffith, Aaron (a). ‘Old Irish pronouns: agreement affixes vs. clitic arguments’, in Andrew Carnie (ed.), Formal Approaches to Celtic Linguistics. Newcastle upon Tyne, UK: Cambridge Scholars Publishing, –. Griffith, Aaron (b). Dictionary of the Old Irish Glosses in the Milan MS Ambr. C inf. . Grimshaw, Jane (). ‘Projections, heads, and optimality’. Linguistic Inquiry : –. Grimshaw, Jane (). ‘Extended projection’, in Jane Grimshaw, Words and Structure. Stanford, CA: Center for the Study of Language and Information, –. Grohmann, Kleanthes (). Prolific Peripheries: A Radical View from the Left. PhD dissertation, University of Maryland, College Park. Grohmann, Kleanthes (). Prolific Domains: On the Anti-locality of Movement Dependencies. Amsterdam/Philadelphia: John Benjamins. Haader, Lea (). ‘Az alárendelő mondatok: Az alanyi, állítmányi, tárgyi és határozói mellékmondatok’ [Subordinate clauses: subjective, predicative, objective and adverbial subclauses], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana I.: A korai ómagyar kor és előzményei. [The diachronic grammar of the Hungarian language I: The Early Old Hungarian period and its antecedents]. Budapest: Akadémiai Kiadó, –. Haader, Lea (). ‘Az alárendelő mondatok: Az alanyi, állítmányi, tárgyi és határozói mellékmondatok’ [Subordinate clauses: subjective, predicative, objective and adverbial subclauses], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana II/.: A kései ómagyar kor: Mondattan. Szöveggrammatika. [The diachronic grammar of the Hungarian language II/: The Late Old Hungarian period: syntax. Textual grammar]. Budapest: Akadémiai Kiadó, –. Haader Lea (). ‘Mikrodiakrónia és változásvizsgálat (az összetett mondatokban)’ [Microdiachrony and the analysis of changes (in complex sentences)], Magyar Nyelvőr : –. Haader, Lea (). ‘Az ómagyar kor: Mondattörténet: Az összetett mondat’ [The Old Hungarian period: the history of sentences: the complex sentence], in Jenő Kiss and Ferenc Pusztai (eds), Magyar nyelvtörténet. [The history of the Hungarian language]. Budapest: Osiris Kiadó, –. Haegeman, Liliane (). ‘Evidential mood, restructuring, and the distribution of functional sembrare’, in Paola Benincà and Nicola Munaro (eds), Mapping the Left Periphery. The Cartography of Syntactic Structures, Volume . Oxford/New York: Oxford University Press, –. Hale, Ken (). ‘Warlpiri and the grammar of non-configurational languages’, Natural Language and Linguistic Theory : –. Hale, William, and Carl Buck (). A Latin Grammar. Tuscaloosa, AL: University of Alabama. Halle, Morris, and Alec Marantz (). ‘Distributed morphology and the pieces of inflection’, in Kenneth Hale and S. Jay Keyser (eds). The View from Building , Cambridge, MA: MIT Press, –. Halle, Morris, and Alec Marantz (). ‘Some key features of distributed morphology’, MIT Working Papers in Linguistics , –. Hancock, W. Neilson, Thaddeus O’Mahony, Alexander George Richey, and Robert Atkinson (eds) (–). Ancient Laws of Ireland. Dublin: Stationery Office.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Harley, Heidi (). ‘External arguments and the mirror principle: on the distinctness of Voice and v’, Lingua : –. Harley, Heidi (). ‘On the identity of roots’, Theoretical Linguistics : –. Hegedűs, Veronika (). ‘The grammaticalization of postpositions in Old Hungarian’, in Theresa Biberauer, and George Walkden (eds), Syntax over Time: Lexical, Morphological and Information–Structural Interactions. Oxford/New York: Oxford University Press, –. Heine, Bernd, and Tania Kuteva (). World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Hellwig, Oliver (–). The Digital Corpus of Sanskrit (DCS). University of Heidelberg: Department of Classical Indology. Hendrickson, George Lincoln, and Harry Mortimer Hubbell (). Brutus: Orator, Cicero. Cambridge, MA: Harvard University Press. Henry, Alison, and Siobhan Cottell (). ‘A new approach to transitive expletives: evidence from Belfast English’, English Language and Linguistics : –. Herman, József (). ‘La disparition de la déclinaison latine et l’évolution du syntagme nominal’, in Christian Touratier (ed.), Syntaxe et latin: actes du IIème congrès international de linguistique latine, Aix-en-Provence, – mars . Marseille: Université de Provence, –. Herman, József (). Vulgar Latin. University Park, PA: The Pennsylvania State University Press. Himmelmann, Nikolaus P. (). ‘Articles’, in Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds), Language Typology and Language Universals: An International Handbook. Volume. Berlin: De Gruyter, –. Hoeksema, Jack (). ‘Verb movement in Dutch present-participle clauses’, in Jan Koster and Henk van Riemsdijk (eds), Germania et alia: A Linguistic Webschrift for Hans den Besten. Electronic publication. Groningen University, Netherlands. Hoekstra, Eric, and Caroline Smits (). ‘Everything you always wanted to know about complementizer agreement’, in Elly van Gelderen and Vida Samii (eds) Proceedings of WECOL , Fresno, CA: California State University, –. Holmberg, Anders (). ‘Deriving OV order in Finnish’, in Peter Svenonius (ed.), The Derivation of OV and VO. Amsterdam/Philadelphia: John Benjamins, –. Holmberg, Anders, Aarti Nayudu, and Michelle Sheehan (). ‘Three partial null–subject languages: a comparison of Brazilian Portuguese, Finnish and Marathi’. Studia Linguistica : –. Holmberg, Anders, and Christer Platzack (). The Role of Inflection in Scandinavian Syntax. Oxford/New York: Oxford University Press. Holmberg, Anders, and Christer Platzack (). ‘The Scandinavian Languages’, in Guglielmo Cinque and Richard S. Kayne (eds), The Oxford Handbook of Comparative Syntax. Oxford/ New York: Oxford University Press, –. Holmberg, Anders, and Ian Roberts (). ‘The syntax–morphology relation’. Lingua : –. Hopper, Mike, and Elizabeth Traugott (). Grammaticalization. Cambridge: Cambridge University Press.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Horvath, Julia (). FOCUS in the Theory of Grammar and the Syntax of Hungarian. Dordrecht: Foris. Huber, Joseph (). Gramática do Português Antigo. (Portuguese translation of Altportugiesisches Elementarbuch, Carl Winters, ). Lisbon: Fundação Calouste Gulbenkian. Iatridou, Sabine, and Anthony Kroch (). ‘The licensing of CP-recursion and its relevance to the Germanic verb-second phenomenon’, Working Papers in Scandinavian Linguistics , –. Ilari, Rodolfo (). ‘Sobre os advérbios aspectuais’, in Rodolfo Ilari (ed.), Gramática do Português Falado. Volume : Níveis de análise linguística. Campinas, Brazil: Editora Unicamp, –. Ingham, Richard (). ‘Negation and OV order in Late Middle English’. Journal of Linguistics : –. Ingham, Richard (). ‘The development of Middle English expletive negative sentences’, Transactions of the Philological Society : –. Iovino, Rossella (). ‘Word order in Latin nominal expressions: the syntax of demonstratives’, in Renato Oniga, Rossella Iovino, and Giuliana Giusti (eds), Formal Linguistics and the Teaching of Latin: Theoretical and Applied Perspectives in Comparative Grammars. Newcastle upon Tyne, UK: Cambridge Scholars Publishing, –. Iovino, Rossella (a). La Sintassi delle Espressioni Nominali Latine. PhD dissertation, University of Venice. Iovino, Rossella (b). La Sintassi dei Modificatori Nominali in Latino. Munich: Lincom. Ishii, Toru (). ‘Cyclic spell-out and the that-t effects’, in Sonya Bird, Andrew Carnie, Jason Haugen, and Peter Norquest (eds), Proceedings of the West Coast Conference on Formal Linguistics . Somerville, MA: Cascadilla, –. Jackendoff, Ray (). X 0 -Syntax: A Study of Phrase Structure. Cambridge, MA: MIT Press. Jackson, John James (). The Iliad of Homer: a Parsed Interlinear Text. . Jäger, Agnes (). History of German Negation. Amsterdam/Philadelphia: John Benjamins. Jagić, Vatroslav I. (ed.) (). Quattuor evangeliorum versionis palaeoslovenicae codex Marianus glagoliticus. St Petersburg, Russia: Berolini. Jelinek, Eloise (). ‘Empty categories, case, and configurationality’, Natural Language and Linguistic Theory : –. Jespersen, Otto (² []). Negation in English and Other Languages. Copenhagen: A. F. Høst. Jøhndal, Marius L., Dag T. T. Haug, and Anders Nøklestad (). PROIEL: Pragmatic Resources in Old Indo-European Languages. University of Oslo. Jonas, Dianne (). Clause Structure and Verb Syntax in Scandinavian and English. PhD dissertation, Harvard University, Cambridge, MA. Jonas, Dianne, John Whitman, and Andrew Garrett (eds) (). Grammatical Change: Origins, Nature, Outcomes. Oxford/New York: Oxford University Press. Juhász, Dezső (). ‘A kötőszók’ [Conjunctions], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana I.: A korai ómagyar kor és előzményei. [The diachronic grammar of the Hungarian language I: The Early Old Hungarian period and its antecedents]. Budapest: Akadémiai Kiadó, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Juhász, Dezső (). ‘A kötőszók’ [Conjunctions], in Loránd Benkő (ed.), A magyar nyelv történeti nyelvtana II/.: A kései ómagyar kor: Morfematika. [The diachronic grammar of the Hungarian language II/: The Late Old Hungarian period: Morphology]. Budapest: Akadémiai Kiadó, –. Julien, Marit (). ‘The syntax of complex tenses’. The Linguistic Review : –. Julien, Marit (). Syntactic Heads and Word Formation. Oxford/New York: Oxford University Press. Kahane, Ahuvia, and Martin Mueller (). The Chicago Homer. Electronic corpus. Chicago: Northwestern University. . Kaiser, George (). ‘A ordem de palavras e a posição do verbo finito no Português Antigo’, in Actas do Congresso Internacional Organizado por Motivos dos Vinte Anos do Português no Ensino Superior, Budapest: University Eötvös Loránd, –. Kálmán, László (a). ‘Word order in neutral sentences’, in István Kenesei (ed.), Approaches to Hungarian . Szeged, Hungary: JATE, –. Kálmán, László (b). ‘Word order in non-neutral sentences’, in István Kenesei (ed.), Approaches to Hungarian . Szeged, Hungary: JATE, –. Kayne, Richard (). French Syntax: The Transformational Cycle. Cambridge, MA: MIT Press. Kayne, Richard (). ‘Romance clitics, verb movement and PRO’, Linguistic Inquiry : –. Kayne, Richard (). ‘Italian negative infinitival imperatives and clitic climbing’, in Liliane Tasmowski and Anne Zribi-Hertz (eds), Hommages à Nicolas Ruwet, –. Ghent, Belgium: Communication and Cognition. Kayne, Richard (). The Antisymmetry of Syntax. Cambridge, MA: MIT Press. Kayne, Richard (). ‘Some notes on comparative syntax, with special reference to English and French’, in Guglielmo Cinque and Richard S. Kayne (eds) The Oxford Handbook of Comparative Syntax. Oxford/New York: Oxford University Press, –. Kayne, Richard (). ‘Comparative syntax’. Lingua : –. Kenesei, István (). ‘On the logic of word order in Hungarian’, in Werner Abraham and Sjaak de Meij (eds), Topic, Focus and Configurationality. Amsterdam/Philadelphia: John Benjamins, –. Kenesei, István (). ‘Subordinate Clauses’, in Ferenc Kiefer and Katalin É. Kiss (eds), The syntactic structure of Hungarian. San Diego/New York: Academic Press, –. Kessler, Brett (). ‘Discontinuous constituents in Latin’. Manuscript. Washington University in St Louis. MO. Keverling Buysman, F. (ed.) (). Ordelen van de Etstoel van Drenthe – and –. Zutphen, Netherlands: Walburgpers. Kiparsky, Paul (). ‘Grammaticalization as optimization’, in Dianne Jonas, John Whitman, and Andrew Garrett (eds), Grammatical Change: Origins, Nature, Outcomes. Oxford/New York: Oxford University Press, –. Klein, Jared S. (a). Toward a Discourse Grammar of the Rigveda. Part , Volume , Heidelberg: Carl Winter Universitätsverlag. Klein, Jared S. (b). Toward a Discourse Grammar of the Rigveda. Part , Volume , Heidelberg: Carl Winter Universitätsverlag. Klein, Jared S. (). ‘Coordinate conjunction in Old Persian’, Journal of the American Oriental Society : –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Klein, Jared S. (). ‘Some Indo-European systems of conjunction: Rigveda, Old Persian, Homer’, Harvard Studies in Classical Philology : –. Klein, Jared S. (). ‘Indefinite pronouns, polarity and related phenomena in Classical Armenian: a study based on the Old Armenian gospels’, Transactions of the Philological Society : –. Koeneman, Olaf (). ‘Verb movement in Germanic and Celtic languages: a flexible approach’. Lingua : –. Komlósy, András (). ‘Complements and adjuncts’, in Ferenc Kiefer and Katalin É. Kiss (eds), The Syntactic Structure of Hungarian. San Diego/New York: Academic Press, –. Koster, Jan (). Domains and Dynasties: The Radical Autonomy of Syntax. Dordrecht: Foris. Koster, Jan (). ‘Predicate Incorporation and the Word Order of Dutch’, in Guglielmo Cinque, Jan Koster, Jean-Yves Pollock, Luigi Rizzi, and Raffaella Zanuttini (eds) Paths Towards Universal Grammar: Studies in Honor of Richard S. Kayne. Washington, DC: Georgetown University Press, –. Kratzer, Angelika (). ‘Severing the external argument from its verb’, in J. Rooryck and L. Zaring (eds), Phrase Structure and the Lexicon. Dordrecht: Kluwer, –. Krifka, Manfred (). ‘Basic notions of information structure’, in Caroline Féry, Gisbert Fanselow, and Manfred Krifka (eds), Working Papers of the SFB, Interdisciplinary Studies on Information Structure (ISIS) . Potsdam, Germany: Universitätsverlag Potsdam, –. Kroch, Anthony (). ‘Reflexes of grammar in patterns of language change’. Language Variation and Change : –. Kroch, Anthony, and Taylor, Ann (). ‘Verb movement in Old and Middle English: Dialect variation and language contact’, in Ans van Kemenade and Nigel Vincent (eds), Parameters of Morphosyntactic Change. Cambridge: Cambridge University Press, –. Kroch, Anthony, and Ann Taylor (). ‘Verb-object order in Early Middle English’, in Susan Pintzuk, George Tsoulas, and Anthony Warner (eds), Diachronic Syntax: Models and Mechanisms. Oxford/New York: Oxford University Press, –. Kroch, Anthony, Ann Taylor, and Don Ringe (). ‘The Middle English verb-second constraint’, in Susan C. Herring, Pieter van Reenen, and Lene Schøsler (eds), Textual Parameters in Older Languages. Amsterdam/Philadelphia: John Benjamins, –. Kühner Raphael, and Stegmann Karl (²). Ausführliche Grammatik der lateinischen Sprache. Hannover, Germany: Hahn. Labov, William, Cohen, Paul, Robins, Clarence, and Lewis, John (). A Study of the Nonstandard English of Negro and Puerto Rican Speakers in New York City. Final Report, Cooperative Project No. , United States Office of Education. Lambova, Mariana (). ‘On triggers of movement and effects at the interfaces’, in Anne Breitbarth and Henk van Riemsdijk (eds), Triggers. Berlin: De Gruyter, –. Landau, Idan (). ‘EPP extensions’, Linguistic Inquiry : –. Langslow, David (). ‘Praetor Urbanus—Urbanus Praetor: some aspects of attributive adjective placement in Latin’, in Philomen Probert and Andreas Willi (eds), Laws and Rules in Indo-European. Oxford/New York: Oxford University Press, –. Lane, George S. (). ‘The Tocharian Puṇyavantajātaka: text and translation’. Journal of the American Oriental Society, .: –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Larrivée, Pierre, and Richard Ingham (eds) (). The Evolution of Negation: Beyond the Jespersen Cycle. Berlin: De Gruyter. Lash, Elliott (). POMIC: The Parsed Old and Middle Irish Corpus, Version .. Dublin Institute for Advanced Studies, School of Celtic Studies, –present. Lausberg, Heinrich (). Elementos de Retórica Literária. (Portuguese translation of Elemente der Literarischen Rhetorik, Munich, Max Hueber, ). Lisbon: Fundação Calouste Gulbenkian. Ledgeway, Adam (). ‘Syntactic and morphosyntactic typology and change in Latin and Romance’, in Martin Maiden, John Charles Smith, and Adam Ledgeway (eds), The Cambridge History of the Romance Languages. Cambridge: Cambridge University Press, –, –. Ledgeway, Adam (). From Latin to Romance: Morphosyntactic Typology and Change. Oxford/New York: Oxford University Press. Ledgeway, Adam (). ‘Syntheticity and analyticity’, in Andreas Dufter and Elisabeth Stark (eds), Manual of Romance Morphosyntax and Syntax. Berlin/Boston: De Gruyter. –. Legate, Julie Anne (). Voice and v: Lessons from Acehnese. Cambridge, MA: MIT Press. Lightfoot, David (). The Development of Language: Acquisition, Change and Evolution. Malden, MA: Blackwell. Lightfoot, David (ed.) (). Syntactic Effects of Morphological Change. Oxford/New York: Oxford University Press. Longo, Beatriz Nunes de Oliveira, and Odette G. L. A. de Souza Campos (). ‘A auxiliaridade: perífrases de tempo e aspecto no português falado’, in Maria Bernadete M. Abaurre and Angela C. S. Rodrigues (eds), Gramática do Português Falado. Volume : Novos estudos descritivos. Campinas, Brazil: Editora da Unicamp, –. Longobardi, Giuseppe (). ‘Reference and proper names: a theory of n-movement in syntax and logical form’, Linguistic Inquiry : –. Lopes, Ana Maria M. (). ‘Contribuição para o estudo dos valores discursivos de sempre’, in Maria Antónia Mota and Rita Marquilhas (eds), Actas do XIII Encontro Nacional da Associação Portuguesa de Linguística. Volume . Lisbon: Colibri, –. Lopes-Rossi, Maria Aparecida Garcia (). ‘Estudo diacrônico sobre as interrogativas do português do Brasil’, in Ian Roberts and Mary A. Kato (eds), Português Brasileiro: Uma viagem diacrônica—Homenagem a Fernando Tarallo. Campinas, Brazil: Editora Unicamp, –. MacCana, Proinsias (). ‘On Celtic word-order and the Welsh “abnormal” sentence’. Ériu : –. MacCana, Proinsias (). ‘Narrative openers and progress markers in Irish’, in Kathryn A. Klar, Eve E. Sweetser, and Claire Thomas (eds), A Celtic Florilegium: Studies in Memory of Brendan O Hehir, Celtic Studies Publications , Lawrence, MA: Celtic Studies Publications, –. MacCoisdealbha, Pádraig ( []). The Syntax of the Sentence in Old Irish: Selected Studies from a Descriptive, Historical and Comparative Point of View. Buchreihe der Zeitschrift für celtische Philologie . Tübingen, Germany: Max Niemeyer Verlag. MacGiolla Easpaig, Dónall (). ‘Aspects of variant word order in Early Irish’. Ériu : –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Macafee, Caroline (). ‘A history of Scots to ’. DSL—Dictionary of the Scots Language/ Dictionar of the Scots Leid. . Mackenzie, Ian, and Wim van der Wurff (). ‘Relic syntax in Middle English and Medieval Spanish: parameter interaction in language change’. Language : –. Manzini, Rita, and Leonardo Savoia (). Grammatical Categories: Variation in Romance Languages. Cambridge: Cambridge University Press. Martin, Stefan E. (). Topics in the Syntax of Nonstandard English. PhD dissertation, University of Maryland, College Park. Martins, Ana Maria (). Clíticos na História do Português. PhD dissertation, University of Lisbon. . Martins, Ana Maria (). Documentos Portugueses do Noroeste e da Região de Lisboa: Da produção primitiva ao século XVI. Lisbon: Imprensa Nacional—Casa da Moeda. Martins, Ana Maria (). ‘The loss of IP scrambling in Portuguese: clause structure, word order variation and change’, in David Lightfoot (ed.), Syntactic Effects of Morphological Change. Oxford/New York: Oxford University Press, –. Martins, Ana Maria (). ‘A emergência do português escrito na segunda metade do século XII’, in Rosario Álvarez, and Antón Santamarina (eds), (Dis)cursos da Escrita: Estudos de Filoloxía Galega Ofrecidos en Memoria de Fernando R. Tato Plaza. A Coruña, Spain: Fundación Pedro Barrié de la Maza, –. Martins, Ana Maria (). ‘Scrambling and information focus in Old and Contemporary Portuguese’. Catalan Journal of Linguistics : –. Martins, Ana Maria (). ‘Deictic locatives, emphasis and metalinguistic negation’, in Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar (eds), Parameter Theory and Linguistic Change. Oxford/New York: Oxford University Press, –. Martins, Ana Maria (). ‘How much syntax is there in metalinguistic negation?’ Natural Language and Linguistic Theory : –. Martins, Ana Maria (forthcoming). ‘Against V as a general property of Old Romance languages’, in Martin Elsig, Ingo Feldhausen, Imme Kuchenbrandt, and Mareike Neuhaus (eds), Romance Languages and Linguistic Theory : Selected Papers from ‘Going Romance’ , Frankfurt. Amsterdam/Philadelphia: John Benjamins. Marvin, Tatjana (). Topics in the Stress and Syntax of Words. PhD dissertation, Massachusetts Institute of Technology, Cambridge, MA. Massam, Diane (). ‘On the left in Niuean’. Lingua : –. Matushansky, Ora (). No More No Less: Existential Comparatives Revisited. Manuscript, Utrecht University. Mioto, Carlos, and Mary A. Kato (). ‘As interrogativas Q do português europeu e do português brasileiro atuais’. Revista da ABRALIN : –. Mitrović, Moreno (). Morphosyntactic Atoms of Propositional Logic: A Philological Programme. PhD dissertation, University of Cambridge. Mitrović, Moreno, and Uli Sauerland (). ‘Decomposing coordination’, in Jyoti Iyer and Leland Kusmer (eds), NELS : Proceedings of the th Meeting of the North East Linguistic Society. Volume . University of Connecticut, Storrs, –. Mitrović, Moreno, and Uli Sauerland (). ‘Two conjunctions are better than one’. Acta Linguistica Hungarica : –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Miyagawa, Shigeru (). Why Agree? Why Move? Unifying Agreement-based and Discourse Configurational Languages. Cambridge, MA: MIT Press. Montgomery, Michael (). ‘Exploring the roots of Appalachian English’. English WorldWide : –. Montgomery, Michael (). ‘Making transatlantic connections between varieties of English: the case of plural verbal-s’. Journal of English Linguistics : –. Montgomery, Michael (). ‘Notes on the development of existential they’. American Speech : –. Montgomery, Michael, and Joseph S. Hall (eds) (). Dictionary of Smoky Mountain English. Knoxville: The University of Tennessee Press. Moraes, Maria Aparecida Torres (). Do português clássico ao português europeu moderno: um estudo diacrônico da cliticização e do movimento do verbo. PhD dissertation, University of Campinas, Brazil. Moran, Pádraic (). ‘Language interaction in the St Gall Priscian Glosses’. Peritia : –. Motapanyane, Virginia (). Theoretical Implications of Complementation in Romanian. Padua: Unipress. Motut, Alexandra (). ‘Merge over move and the empirical force of economy in minimalism’. Toronto Working Papers in Linguistics : –. Munn, Alan (). Topics in the Syntax and Semantics of Coordinate Structures, PhD dissertation, University of Maryland. Myler, Neil (). ‘A note on Den Dikken’s () arguments for a J(unction) head’. Manuscript. New York University. Nakamura, Chiye (). ‘On the cleft sentence and the nominalized sentence in Irish’. 京都大 学言語学研究 [Kyoto University Linguistic Research] : –. . Neu, Erich (). Studien zu den Boǧazköy-Texten. Althethitische Ritualtexte in Umschrift, . Wiesbaden, Germany: Harrassowitz. Newton, Glenda (). The Development and Loss of the Old Irish Double System of Verbal Inflection. PhD dissertation, Cambridge University. Nunes, Jairo (). ‘Linearization of chains and phonetic realization of chains links’, in Samuel Epstein and Norbert Hornstein (eds), Working Minimalism. Cambridge, MA: MIT Press, –. Nunes, Jairo (). Linearization of Chains and Sideward Movement. Cambridge, MA: MIT Press. O’Donovan, John (). A Grammar of the Irish Language. Dublin: Hodges and Smith. Oniga, Renato (). Il Latino: Breve introduzione linguistica. Milan: Franco Angeli. Oosterhout, Meinte (). Snitser Recesboeken –. Assen, Netherlands: Van Gorcum. Overdiep, Gerrit Siebe (). Stilistische grammatica van het moderne Nederlandsch. Zwolle, Netherlands: Tjeenk Willink. Paixão de Sousa, Maria Clara (). Língua barroca: Sintaxe e história do português nos . PhD dissertation, University of Campinas, São Paulo, Brazil. Parrott, Jeffrey K. (). ‘Negative inversion in African American Vernacular English: a case of optional movement?’ in Nancy Mae Antrim, Grant Goodall, Martha SchulteNafeh, and Vida Samiian (eds), Proceedings of the th Western Conference on Linguistics
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
(WECOL). Volume . Fresno, CA: Department of Linguistics, California State University, –. Pesetsky, David, and Esther Torrego (). ‘Tense, case, and the nature of syntactic categories’, in Jacqueline Guéron and Jacqueline Lecarme (eds), The Syntax of Time. Cambridge, MA: MIT Press, –. Picallo, Carme (). Opaque Domains. PhD Disssertation, City University of New York. Piel, Joseph (). Livro dos oficios de Marco Tullio Ciceram, o qual tornou em linguagem o Ifante D. Pedro, duque de Coimbra. Edição crítica, segundo o MS. de Madrid, prefaciada, anotada e acompanhada de glossário. Coimbra, Portugal: Universidade de Coimbra. Pimpão, Álvaro Júlio da Costa (4 []). Os Lusíadas, de Luís de Camões. Lisbon: MNE, Instituto Camões. Pinault, Georges-Jean (). Chrestomathie tokharienne: textes et grammaire. Leuven, Belgium: Peeters. Pinkster, Harm (). ‘Changing patterns of discontinuity in Latin’. Paper presented at Latling—e Colloque International de Linguistique Latine, Facultés Universitaires SaintLouis, – April. Pintzuk, Susan (). ‘Verb seconding in Old English: verb movement to Infl’, The Linguistic Review : –. Pintzuk, Susan, George Tsoulas, and Anthony Warner (eds) (). Diachronic Syntax: Models and Mechanisms. Oxford/New York: Oxford University Press. Pintzuk, Susan, and Ann Taylor (). ‘The loss of OV order in the history of English’, in Ans van Kemenade and Bettelou Los (eds), The Handbook of the History of English. Oxford: Blackwell, –. Pirvulescu, Mihaela (). ‘Agreement paradigms across moods and tenses’, in Chiyo Nishida and Jean-Pierre Y. Montreuil (eds), Morphology, Syntax, Semantics, and Pragmatics: New Perspectives on Romance Linguistics. Volume . Amsterdam/Philadelphia: John Benjamins, –. Poletto, Cecilia (). Word Order in Old Italian. Oxford/New York: Oxford University Press. Pollock, Jean-Yves (). ‘Verb movement, universal grammar, and the structure of IP’. Linguistic Inquiry : –. Postal, Paul M. (). ‘On so-called «Pronouns» in English’, in David A. Reibel and Sanford A. Schane (eds), Modern Studies in English: Readings in Transformational Grammar. Englewood Cliffs, NJ: Prentice-Hall, –. Postma, Gertjan J. (a). ‘IV in het Middeldrents en de Subjunctief ’, Taal en Tongval .: –. Postma, Gertjan J. (b). ‘IV in het Middelnederlands, IPI in het Fries en de distributie van subjunctieven’, Taal en Tongval .: –. Postma, Gertjan J., and Hans J. Bennis. (). ‘Variatie in negatie—Het gedrag van het negatieve cliticum in het Middeldrents van rond ’, Taal en Tongval .: –. Postma, Gertjan J. (). ‘The impact of failed changes’, in Anne Breitbarth, Christopher Lucas, Sheila Watts, and David Willis (eds), Continuity and Change in Grammar. Amsterdam/ Philadelphia: John Benjamins, –. Postma, Gertjan J. (a). ‘Language contact and linguistic complexity – the rise of the reflexive pronoun zich in a th century Netherlands’ border dialect’, in Dianne Jonas and
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
John Whitman (eds), Grammatical Change—Origins, Nature, Outcomes. Oxford/New York: Oxford University Press, –. Postma, Gertjan J. (b). ‘Het verval van het pronomen du – dialect geografie en historische syntaxis’. Nederlandse Taalkunde : –. Postma, Gertjan J. (). ‘Clause-typing by []—the loss of the nd person pronoun du “you” in Dutch, Frisian and Limburgian dialects’, in Victoria Camacho-Taboada, Ángel L. JiménezFernández, Javier Martín-González, and Mariano Reyes-Tejedor (eds), Information Structure and Agreement. Amsterdam/Philadelphia: John Benjamins, –. Powell, Jonathan (). ‘Hyperbaton and register in Cicero’, in Eleanor Dickey and Anna Chahoud (eds), Colloquial and Literary Latin. Cambridge: Cambridge University Press, –. Predolac, Nikol (). ‘Information structure and Serbian bipartite NPs’, in Wayles Browne, Adam Cooper, Alison Fisher, Esra Kesici, Nikola Predorac, and Draga Zec (eds), Proceedings of Formal Approaches to Slavic Linguistics . Ann Arbor: Michigan Slavic Publications, –. Prince, Ellen (). ‘Towards a taxonomy of given-new information’, in Peter Cole (ed.), Radical Pragmatics. New York: Academic Press, –. Progovac, (). Negative and Positive Polarity: A Binding Approach. Cambridge: Cambridge University Press. Quer, Josep (). Mood at the Interface. PhD dissertation, LOT, Utrecht University. Ramchand, Gillian, and Peter Svenonius (). ‘The lexical syntax and lexical semantics of the verb-particle construction’, in Line Mikkelsen and Cristopher Potts (eds), WCCFL : Proceedings of the st West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Press, –. Reinhart, Tanya (). ‘Pragmatics and linguistics: an analysis of sentence topics’. Philosophica : –. Reintges, Chris H. (a). ‘High analyticity and Coptic particle syntax: a phase-based approach’. The Linguistic Review : –. Reintges, Chris H. (b). ‘The evolution of the Ancient Egyptian Stative: diachronic stability besides inflectional change’. Folia Orientalia : –. Reintges, Chris H. (). ‘Macroparametric change and the synthetic-analytic continuum: the case of Ancient Egyptian’, in Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar (eds), Parameter Theory and Linguistic Change. Oxford/New York: Oxford University Press, –. Reintges, Chris H. (). ‘Sapirian “drift” towards analyticity and long-term morphosyntactic change in Ancient Egyptian’, in Ritsuko Kikusawa and Laurence A. Reid (eds), Historical Linguistics : Selected Papers from the th International Conference on Historical Linguistics, Osaka (– July ). Amsterdam/Philadelphia: John Benjamins, –. Reintges, Chris H., and Sonia Cyrino (). ‘Rethinking parameter theory diachronically: A macrocomparative approach to the analyticization of the verbal tense–aspect systems of Brazilian Portuguese and Coptic Egyptian’, in Luis Eguren, Olga Fernández Soriano, and Amaya Mendikoetxea Pelayo (eds), Rethinking Parameters. Oxford/New York: Oxford University Press, –. Remberger, Eva-Maria (). ‘Left peripheral interactions in Romance’. Paper presented at the Workshop on Focus, Contrast and Givenness in Interaction with Extraction and Deletion, University of Tübingen, Germany, – March .
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Ribeiro, Ilza (). ‘Evidence for a verb-second phase in Old Portuguese’ in Adrian Battye and Ian Roberts (eds) Clause Structure and Language Change. Oxford/New York: Oxford University Press, –. Richards, Norvin (). Uttering Theory. Unpublished monograph, Massachusetts Institute of Technology, Cambridge, MA. Richards, Marc David (). Object Shift and Scrambling in North and West Germanic: A Case Study in Symmetrical Syntax. PhD dissertation, University of Cambridge. Richards, Marc David (). ‘On feature inheritance: an argument from the phase impenetrability condition’, Linguistic Inquiry : –. Rinke, Esther (). ‘Verb-placement in Old Portuguese’, in Daniel Jacob and Andreas Dufter (eds), Focus and Background in Romance Languages. Amsterdam/Philadelphia: John Benjamins, –. Rissanen, Marti, and Anneli Meurman-Solin (). The Helsinki Corpus of Older Scots. [Electronic resource: –]. Oxford: University of Oxford Text Archive. Rizzi, Luigi (). Issues in Italian Syntax. Dordrecht: Foris. Rizzi, Luigi (). Relativized Minimality. Cambridge, MA: MIT Press. Rizzi, Luigi (). ‘Residual Verb Second and the wh-criterion’, in Adriana Belletti and Luigi Rizzi (eds), Parameters and Functional Heads: Essays in Comparative Syntax. Oxford/New York: Oxford University Press, –. Rizzi, Luigi (). ‘The fine structure of the left periphery’, in Liliane Haegeman (ed.), Elements of Grammar: Handbook in Generative Syntax. Dordrecht: Kluwer, –. Rizzi, Luigi (). ‘On the position of ‘Int(errogative)’ in the left periphery of the clause’, in Guglielmo Cinque, and Giampaolo Salvi (eds) Current Studies in Italian Syntax: Essays Offered to Lorenzo Renzi. Amsterdam: Elsevier, –. Rizzi, Luigi (). ‘On the forms of chains: criterial positions and ECP effects’, in Lisa Lai-Shen Cheng and Norbert Corver (eds), Wh-movement: moving on. Cambridge, MA: MIT Press, –. Roberts, Craige (). ‘Information structure in discourse: towards an integrated formal theory of pragmatics’, in Jae-Hak Yoon and Andreas Kathol (eds), OSU Working Papers in Linguistics : Papers in Semantics. Columbus: The Ohio State University, –. Roberts, Ian (). ‘Agreement parameters and the development of English modal auxiliaries’, Natural Language and Linguistic Theory : –. Roberts, Ian (). Verbs and Diachronic Syntax: A Comparative History of English and French. Dordrecht: Kluwer. Roberts, Ian (). Principles and Parameters in a VSO Language: A Case Study in Welsh. Oxford/New York: Oxford University Press. Roberts, Ian (). Agreement and Head Movement: Clitics, Incorporation, and Defective Goals. Cambridge, MA: MIT Press. Roberts, Ian (a). ‘Phases, head movement and second-position effects’, in Ángell J. Gallego (ed.), Phases: Developing the Framework. Berlin/Boston: De Gruyter, –. Roberts, Ian (b). ‘Macroparameters and Minimalism: a program for comparative research’, in Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar (eds), Parameter Theory and Linguistic Change. Oxford/New York: Oxford University Press, –.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Roberts, Ian, and Anna Roussou (). Syntactic Change: A Minimalist Approach to Grammaticalization. Cambridge: Cambridge University Press. Rögnvaldsson, Eríkur, and Höskuldur Thráinsson (). ‘On Icelandic word order once more’, in Joan Maling, and Annie Zaenen (eds), Modern Icelandic Syntax. San Diego/New York: Academic Press, –. Rohlfs, Gerhard (). Grammatica storica della lingua italiana e dei suoi dialetti. Volume : Morfologia. Turin: Einaudi. Rooth, Mats (). Association with Focus. PhD dissertation, University of Massachusetts at Amherst. Rooth, Mats (). ‘A theory of focus interpretation’, Natural Language Semantics : –. Ross, John (). Constraints on Variables in Syntax. PhD dissertation, Massachusetts Institute of Technology, Cambridge. MA. Rouveret, Alain (). ‘On verb–subject languages’. Lingua : –. Rowlett, Paul (). Sentential Negation in French. Oxford/New York: Oxford University Press. Saito, Mamoru, and Keiko Murasugi (). ‘Subject predication within IP and DP’, in Kyle Johnson and Ian Roberts (eds), Beyond Principles and Parameters. Dordrecht: Kluwer, –. Samuels, Bridget D. (). The Structure of Phonological Theory, PhD dissertation, Harvard University, Cambridge, MA. Schachter, Paul (). ‘Focus and relativization’, Language : –. Schmitt, Cristina (). ‘Cross-linguistic variation and the present perfect: the case of Portuguese’. Natural Language and Linguistic Theory : –. Schwegler, Armin (). ‘Predicate negation and word-order change: a problem of multiple causation’, Lingua : –. Schwegler, Armin (). ‘Word-order changes in predicate negation strategies in Romance Languages’, Diachronica : –. Sekerina, Irina (). ‘The scrambling complexity hypothesis and processing of split scrambling constructions in Russian’, Journal of Slavic Linguistics : –. Sells, Peter, John Rickford, and Thomas Wasow (). ‘An optimality theoretic approach to variation in negative inversion in AAVE’, Natural Language and Linguistic Theory : –. Seuren, Pieter (). ‘The comparative’, in Ferenc Kiefer and Nicolas Ruwet (eds), Generative Grammar in Europe. Dordrecht: D. Reidel, –. Shearer, Kathy (–). Dante Oral History Project. Recorded oral history interviews collected in Dante, VA, housed in the Archives of Appalachia, East Tennessee State University, Johnson City. Sieg, Emil, and Wilhelm Siegling (). Tocharische Sprachreste, I. Band: Die Texte, A. Transcription. Berlin/Leipzig: De Gruyter. Silva, Rosa Virgínia Mattos e (). O Português Arcaico: Fonologia, Morfologia e Sintaxe. São Paulo, Brazil: Contexto. Simonyi, Zsigmond (). A jelzők mondattana. Nyelvtörténeti tanulmány [The syntax of attributes. A diachronic study]. Budapest: MTA. Slade, Benjamin M. (). Formal and Philological Inquiries into the Nature of Interrogatives, Indefinites, Disjunction, and Focus in Sinhala and Other Languages, PhD dissertation, University of Illinois at Urbana-Champaign.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Sommerer, Lotte (). Old English ‘se’: from demonstrative to article: A usage-based study of nominal determination and category emergence. PhD dissertation, Universität Wien, Vienna, Austria. Spevak, Olga (). ‘L’hyperbate en latin tardif ’. Paper presented at Latin Vulgaire, Latin Tardif, IXe Colloque International, – September. Stenson, Nancy. (). Studies in Irish Syntax. Tübingen, Germany: Gunter Narr Verlag. Stifter, David (). Sengoídelc: Old Irish for Beginners. Syracuse, NY: Syracuse University Press. Stjepanović, Sandra (). ‘Free word order and copy theory of movement’, in Norbert Corver, and Jairo Nunes (eds), The Copy Theory of Movement. Amsterdam/Philadelphia: John Benjamins, –. Stokes, Whitley (ed.) (). The Old-Irish Glosses at Würzburg and Carlsruhe. Hertford: Stephen Austin and Sons. Stowell, Tim (). The Syntax of Tense. Manuscript, University of California at Los Angeles. Sulger, Sebastian (). Irish Clefting: The LFG Perspective. PhD dissertation, Universität Konstanz, Germany. Surányi, Balázs (). ‘Hungarian as a Japanese-type scrambling language’, in Christopher Davies, Amy Rose Deal, and Youri Zabbal (eds) NELS : Proceedings of the th Annual Meeting of the North East Linguistic Society. Amherst, MA: GLSA, –. Surányi, Balázs (). ‘Verbal particles inside and outside vP’. Acta Linguistica Hungarica : –. Sybesma, Rint P. E. (). ‘Whether we tense-agree overtly or not’. Linguistic Inquiry : –. Szabolcsi, Anna (). ‘The noun phrase’, in Ferenc Kiefer and Katalin É. Kiss (eds), The Syntactic Structure of Hungarian. San Diego/New York: Academic Press, –. Szabolcsi, Anna (). ‘What do quantifier particles do?’ Linguistics and Philosophy : –. Takita, Kensuke (). Cyclic Linearization and Constraints on Movement and Ellipsis, PhD dissertation, Nanzan University, Japan. Tanaka, Tomoyuki (). ‘On the development of transitive expletive constructions in the history of English’. Lingua , –. Taylor, Ann, and Susan Pintzuk (). ‘The interaction of syntactic change and information status effects in the change from OV to VO in English’. Catalan Journal of Linguistics : –. Teyssier, Paul (). Manuel de Langue Portugaise. Paris: Klincksieck. Teyssier, Paul (). Histoire de la Langue Portugaise. Paris: Presses Universitaires de France. Thráinsson, Höskuldur (). The Syntax of Icelandic. Cambridge: Cambridge University Press. Thurneysen, Rudolf (). A Grammar of Old Irish. Dublin: Dublin Institute for Advanced Studies. Tortora, Christina (). ‘The case of Appalachian expletive they’. American Speech : –. Travis, Lisa (). Parameters and Effects of Word Order Variation. PhD Dissertation, Massachusetts Institute of Technology, Cambridge, MA. Uriagereka, Juan (). On Government. PhD dissertation, University of Connecticut, Storrs. Väänänen, Veikko (). Introduzione al Latino Volgare. Bologna: Pàtron.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Van Gelderen, Elly (). Grammaticalization as Economy. Amsterdam/Philadelphia: John Benjamins. Van Gelderen, Elly (). ‘The definiteness cycle in Germanic’. Journal of Germanic Linguistics : –. Van Gelderen, Elly (a). ‘Negative cycles’. Linguistic Typology : –. Van Gelderen, Elly (b). ‘Linguistic cycles and economy principles: the role of universal grammar in language change’, in Eythórsson, Thórhallur (ed.), Grammatical Change and Linguistic Theory. The Rosendal Papers. Amsterdam/Philadelphia: John Benjamins, –. Van Gelderen, Elly (). ‘Renewal in the left periphery: economy and the complementiser layer’. Transactions of the Philological Society .: –. Van Gelderen, Elly (). The Linguistic Cycle: Language Change and the Language Faculty. Oxford/New York: Oxford University Press. Van Hamel, Anton Gerard (). Compert Con Culainn and other stories. MMIS . Dublin: The Stationery Office. Van Kemenade, Ans, and Los Bettelou (). ‘Particles and Prefixes in Dutch and English’, in Geert Booij and Jaap van Marle (eds), Yearbook of Morphology . Dordrecht: Kluwer, –. Van Kemenade, Ans, and Nigel Vincent (eds) (). Parameters of Morphosyntactic Change. Cambridge: Cambridge University Press. Van Nooten, Barend A., and Gary B. Holland (eds) (). Rig Veda: A Metrically Restored Text. Cambridge, MA: Harvard University Press. Van Riemsdijk, Henk (). ‘Movement and regeneration’, in Paola Benincá (ed.), Dialect Variation and the Theory of Grammar. Dordrecht: Foris, –. Van Riemsdijk, Henk (). ‘Functional prepositions’, in Harm Pinkster and Inge Genee (eds) Unity in Diversity: Papers Presented to Simon C. Dik on his th Birthday. Dordrecht: Foris, –. Vangsnes, Øystein Alexander (). ‘Icelandic expletive constructions and the distribution of subject types’, in Peter Svenonius (ed.) Subjects, Expletives, and the EPP. Oxford/New York: Oxford University Press, –. Varga, Diána (). ‘A mód, a modalitás és az imperatív operátor viszonya magyar felszólító mondatokban’ [The relation of mood, modality, and the imperative operator in Hungarian imperative sentences], in Andrea Parapatics (ed.) Félúton . Konferencia Kötet. [Proceeedings of the student conference ‘Félúton’, volume ]. Budapest, ELTE—BTK. Verdenius, Andries Anton (). ‘De ontwikkeling der Hollandse voornaamwoorden je en jij’. Tijdschrift voor Nederlandse taal- en letterkunde : –. Verdenius, Andries Anton (). ‘Over de aanspreekvorm ie (i–j) in onze oostelike provincieën’. Tijdschrift voor Nederlandse taal- en letterkunde : –. Verdenius, Andries Anton (). ‘De laatste sporen van du in Noord-Holland’, De Nieuwe Taalgids : –. Vergnaud, Jean-Roger (). French Relative Clauses. PhD dissertation, Massachusetts Institute of Technology, Cambridge, MA. Vergnaud, Jean-Roger (). Dépendences et niveaux de représentation en syntaxe. Amsterdam/Philadelphia: John Benjamins.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
References
Vermeulen, Reiko (). ‘Non-topical wa-phrases in Japanese’, in Rafaella Folli and Christiane Ulbrich (eds), Interfaces in Linguistics: New Research Perspectives. Oxford/New York: Oxford/New York University Press, –. Vikner, Sten (). Verb Movement and Expletive Subjects in the Germanic Languages. Oxford/New York: Oxford University Press. Visser, Fredericus Theodorus (). An Historical Syntax of the English Language. The Hague: Brill. Wackernagel, Jacob (). ‘Über ein Gesetz der indo-germanischen Wortstellung’. Indogermanische Forschungen : –. Ward, Gregory, and Betty J. Birner (). ‘Discourse and Information Structure’, in Deborah Schiffrin, Deborah Tannen, and Heidi Hamilton (eds), The Handbook of Discourse Analysis. Oxford: Basil Blackwell, –. Waters, Cathleen (). ‘The preposition cycle in English’, in Elly van Gelderen (ed.), Cyclical Change. Amsterdam/Philadelphia: John Benjamins, –. White-Sustaíta, Jessica (). ‘Reconsidering the syntax of non-canonical negative inversion’, English Language and Linguistics : –. Willis, David (). Syntactic Change in Welsh: A Study of the Loss of Verb-Second. Oxford/ New York: Oxford University Press. Wilson, Horace Hayman (). Ṛgveda Saṃ hitā. Volumes –. Varanasi, India: Indica. Winter, Yoad (). Flexible Boolean Semantics: Coordination, Plurality and Scope in Natural Language. PhD dissertation, Utrecht University. Wolfe, Sam (). Microvariation in Medieval Romance Syntax: A comparative study. PhD dissertation, University of Cambridge. Wolfram, Walt, and Donna Christian (). Appalachian Speech. Arlington, VA: Center for Applied Linguistics. Yang, Charles (). Knowledge and Learning in Natural Languages. Oxford/New York: Oxford University Press. Zanuttini, Raffaella (). Negation and Clausal Structure: A Comparative Study of Romance Languages. Oxford/New York: Oxford University Press. Zanuttini, Raffaella (). ‘La negazione’, in Giampaolo Salvi and Lorenzo Renzi (eds), Grammatica dell’Italiano Antico. Volume . Bologna: Il Mulino, –. Zeijlstra, Hedde (). Sentential Negation and Negative Concord. PhD Dissertation, University of Amsterdam. Zhang, Niina Ning (). Coordination in Syntax, Cambridge: Cambridge University Press. Zimmermann, Malte (). ‘Contrastive focus’, in Caroline Féry, Gisbert Fanselow, and Manfred Krifka (eds), Working Papers of the SFB, Interdisciplinary Studies on Information Structure . Potsdam, Germany: Universitätsverlag Potsdam, –. Zubizarreta, María Luisa (). Prosody, Focus and Word Order. Cambridge, MA: MIT Press. Zubizarreta, María Luisa (). ‘Las funciones informativas: tema y foco’, in Ignacio Bosque and Violeta Demonte (eds), Gramática Descriptiva de la Lengua Española. Madrid: Espasa Calpe, –. Zumpt, Karl (). A Grammar of the Latin Language. London: B. Fellowes. Zwart, Jan-Wouter (). Dutch Syntax: A Minimalist Approach. PhD dissertation, University of Groningen, Netherlands.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of names Abels, Klaus Aboh, Enoch Ackema, Peter Ackerman, Farrell Adams, James Noel , , , Adams, Marianne Agbayani, Chris Golston , , Alexiadou, Artemis , , , Algeo, John Amaral, Patrícia – Ambar, Manuela , , –, Anderson, Gregory D. S. Androutsopoulou, Antonia Antonelli, André , –, –, – Avelar, Juanito Bacskai-Atkari, Julia , , , , Baker, Mark , , , Bakró-Nagy, Marianne Bartos, Huba Bašić, Monika Basilico, David Batllori, Montse Battye, Adrian Bauer, Brigitte , , Bauer, Laurie Baunaz, Lena , Bayer, Joseph , Beekes, Robert S. P. Belletti, Adriana , –, , , –, –, Benincà, Paola , , Bennis, Hans J. – Bergin, Osborn , – Bernstein, Judy , , , , Bertinetto, Pier Marco Berwick, Robert C. Bianchi, Valentina , , , Biberauer, Theresa , , , –, , , , , , , – Birner, Betty J. , – Bisagni, Jacopo Bildhauer, Felix Blümel, Rudolf Bobaljik, Jonathan David –, , –
Boisacq, Emile Bošković, Željko , , –, , –, –, – Breitbarth, Anne , Brito, Ana Maria Broekhuis, Hans , Brunet, Étienne Buck, Carl Bures, Anton Butler, Alastair Campos, Odette G. L. A. de Souza Cardinaletti, Anna , , , , , Cardoso, Adriana , , Castilho, Ataliba Teixeira de Castillo, Juan Carlos Ćavar, Damir , , –, Cecchetto, Carlo Chomsky, Noam , , , , –, , , , , , Christian, Donna – Cinque, Guglielmo , , –, , , –, , , , Clackson, James , Clark, Robin Conradie, Simone , Cook, Philippa Corver, Norbert Costa, João , , , , , , –, , – Cottell, Siobhan Crisma, Paola Cruschina, Silvio Cyrino, Sonia , , , , , , D. Mátai, Mária Danckaert, Lieven , , , , , –, , De Haan, Germen J. –, De Swart, Henriette De Vries, Mark , Del Prete, Fabio – Den Besten, Hans Demirdache, Hamida , Den Dikken, Marcel –, , , , ,
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of names
Denison, David Derksen, Rick Devine, Andrew –, , Dobrovie-Sorin, Carmen Doherty, Cathal Doetjes, Jenny Dömötör, Adrienne Dömötör, Éva , Don, Jan Dressler, Wolfgang U. Drury, John E. Dryer, Matthew S. , – Duarte, Inês , Dunkel, George E. , – É. Kiss, Katalin , , , –, , , –, –, Egedi, Barbara , , –, , Eide, Kristine –, –, Elenbaas, Marion Ellegård, Alvar Emonds, Joseph Enkvist, Nils Erik –, Ernout, Alfred Eska, Joseph Fanselow, Gisbert –, –, Farkas, Donka Feagin, Crawford , , – Fenger, Paula Féry, Caroline , Fiéis, Alexandra , –, Fiengo, Robert Fitzgerald, Colleen M. Foreman, John Fowlie, Meaghan Franks, Steven Frascarelli, Mara –, , , Frota, Sónia –, Fruyt, Michèle – Gajewski, Jon Galambos, Dezső Gallasy, Magdolna Galves, Antonio Galves, Charlotte , , , –, , Garrett, Andrew Garzonio, Jacopo , , Gergel, Remus Gibrail, Alba , , – Giorgi, Alexandra , , – Giusti, Giuliana , Goldstein, David M.
Golston, Chris , , Gonda, Jan Gonzaga, Manuela – Götze, Michael Green, Lisa –, , –, Greenberg, Joseph H. Grewendorf, Günther Griffith, Aaron , , Grimshaw, Jane , Grohmann, Kleanthes K. , Haader, Lea –, , Haegeman, Liliane , Hale, Ken , Hale, William Hall, Joseph S. , –, – Halle, Morris , Hegedűs, Veronika , , , , Heine, Bernd Henry, Alison Harley, Heidi , Herman, József , Hernanz, Maria-Lluïsa Himmelmann, Nikolaus P. Hinterhölzl, Roland –, , , Hoeksema, Jack Holmberg, Anders , , , , , , , – Hopper, Mike Horrocks, Geoffrey Horvath, Julia – Huber, Joseph Iatridou, Sabine Ingham, Richard , , Iovino, Rossella Ishii, Toru Jackendoff, Ray , Jäger, Agnes Jelinek, Eloise Jespersen, Otto – Jonas, Dianne , –, , Juhász, Dezső – Julien, Marit , Kaiser, George Kálmán, László Kato, Mary A. Kayne, Richard , , , –, , , , , , , , , , Kenesei, István Kessler, Brett
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of names Keverling, Buysman, F. Kiparsky, Paul Klein, Jared S. , , Komlósy, András Koster, Jan , , Kratzer, Angelika Krifka, Manfred , , Kroch, Anthony , , , –, , , , – Kühner, Raphael Kuteva, Tania Labov, William – Larrivée, Pierre Lash, Elliott , , Lambova, Mariana Landau, Idan Langslow, David Ledgeway, Adam , –, , , , , –, –, , –, – Legate, Julie Anne Lightfoot, David Longo, Beatriz N. de Oliveira Longobardi, Giuseppe , Lopes, Ana Maria M. – Lopes, Ruth Lopes-Rossi, Maria Aparecida G. Los, Bettelou MacCana, Proinsias , – MacCoisdealbha, Pádraig Macafee, Caroline , Mackenzie, Ian , Manzini, Rita Marantz, Alec , Martin, Stefan E. Martins, Ana Maria , –, , –, –, , , – Marvin, Tatjana Massam, Diane – Mathieu, Eric , Matushansky, Ora Meillet, Antoine Mellet, Sylvie Mioto, Carlos Mitrović, Moreno , , –, , , , , – Montgomery, Michael –, , –, –, Moraes, Maria Aparecida Torres Moran, Pádraic Motapanyane, Virginia Motut, Alexandra
Munn, Alan Murasugi, Keiko Myler, Neil Nakamura, Chiye , , , Nayudu, Aarti Neeleman, Ad Newton, Glenda , Nunes, Jairo , , –, – O’Donovan, John Oniga, Renato , – Overdiep, Gerrit Siebe – Paixão de Sousa, Maria Clara –, , Parrott, Jeffrey K. Pesetsky, David Pianesi, Fabio , , – Picallo, Carme , Pinkster, Harm –, Pintzuk, Susan , , , , Pirvulescu, Mihaela Platzack, Christer , Poletto, Cecilia , , –, , , – Pollock, Jean-Yves , Postal, Paul M. Postma, Gertjan J. , , –, –, –, Powell, Jonathan – Predolac, Nikol Prince, Ellen Progovac, Ljljana , Quer, Josep Ramchand, Gillian Reinhart, Tanya , Reintges, Chris H. , –, , , – Remberger, Eva-Maria Ribeiro, Ilza Richards, Norvin Richards, Marc David Rinke, Esther , Rizzi, Luigi , , , , , , , –, –, , , –, , , Roberts, Craige Roberts, Ian , –, , –, , , , , –, , –, –, , – Roca, Francesc Rodrigues, Ângela
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of names
Rögnvaldsson, Eríkur Rohlfs, Gerhard Rooth, Mats Ross, John , , , Roussou, Anna , Rouveret, Alain Rowlett, Paul
Tortora, Christina –, Traugott, Elizabeth Travis, Lisa – Tsoulas, George
Saito, Mamoru Samuels, Bridget D. – Sandalo, Filomena Savoia, Leonardo Schachter, Paul –, Schmitt, Cristina Schwegler, Armin , Sekerina, Irina Sells, Peter Seuren, Pieter Shearer, Kathy , , Sheehan, Michelle Simon, Eszter Simonyi, Zsigmond – Slade, Benjamin M. , Sommerer, Lotte Spevak, Olga Stentson, Nancy – Starke, Michael , , , Stegmann, Karl Stephens, Laurence –, , Stifter, David , Stjepanović, Sandra –, Stowell, Tim Sulger, Sebastian , Surányi, Balázs –, , – Svenonius, Peter Sybesma, Rint P. E. Szabolcsi, Anna , , ,
Väänänen, Veikko Van der Wurff, Wim , Van Gelderen, Elly , –, , , –, – Van Kemenade, Ans , Van Riemsdijk, Henk , , Vangsnes, Øystein Varga, Diána , Vergnaud, Jean-Roger , Vermeulen, Reiko Vikner, Sten Vincent, Nigel Visser, Fredericus Theodorus
Takita, Kensuke Tanaka, Tomoyuki , , Taylor, Ann , , , –, , , Teyssier, Paul , Thráinsson, Höskuldur Thurneysen, Rudolf Torrego, Esther
Zanuttini, Raffaella , , , , , , Zeijlstra, Hedde Zhang, Niina Ning Zimmermann, Malte Zubizarreta, María Luisa –, , Zumpt, Karl Zwart, Jan-Wouter , , ,
Uriagereka, Juan , Uribe-Etxebarria, Myriam
Wackernagel, Jacob , , –, Walkden, George Ward, Gregory – Warner, Anthony Waters, Cathleen Webelhuth, Gert White-Sustaíta, Jessica , – Whitman, John , Willis, David Winter, Yoad Wolfe, Sam – Wolfram, Walt – Wurmbrand, Susi Yang, Charles
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of languages Avar – Avestan Afrikaans , Armenian (Classical Armenian) Bulgarian n., Calabrian Catalan Celtiberian Coptic Egyptian –, –, – Cuneiform Luwian Dutch , , , , , – dialectal Dutch –, – Middle Dutch – Modern Dutch , – English –, , n., , , , n., , , –, , , , –, , , n., –, , , , n. African American English –, – Appalachian English –, – Belfast English n. Early Modern English , , Middle English , –, –, , – Old English –, (present-day) varieties of English n., –, , – French , , , , , , , , , , n. Old French Frisian –, , –, Modern Frisian n., , , Old Frisian , Gaulish Germanic , –, –, , , –, n., , , , , , n. Greek , , –, –, , Homeric, Ancient Greek , –, Modern Greek German n., , , , , , , Gothic , , , ,
Hindi Hittite , , – Hungarian n., –, –, – Middle Hungarian –, –, –, – Modern Hungarian –, n., , n., –, –, –, –, , Old Hungarian –, –, –, –, –, – Proto-Hungarian , , , Icelandic –, , , , Indic –, (Proto-)Indo-European –, –, Indo-Iranian , Irish , , , , – Modern Irish , , , – Old Irish , , , – Italian , n., , –, , , –, , , – Florentine Modern Italian , n., –, Old Italian –, – Southern Italian Dialects Japanese , Latin –, , , –, –, , , –, , –, –, , –, n., – Archaic Latin Classical Latin , , –, –, –, , Late Latin –, n., , –, –, Limburgian Low Saxon – fifteenth-century Low Saxon , , Middle Low Saxon Modern Low Saxon – Macedonian n., Mycenaean Niuean – Occitan (Old Occitan) Persian (Old Persian)
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of languages
Portuguese –, –, –, –, –, –, Brazilian Portuguese –, , –, – Classical Portuguese – European Portuguese –, –, –, –, , –, –, earlier stages of Portuguese –, –, –, – Old Portuguese – Romance , , –, –, , , , , , –, , , , , –, n., , – Russian , Sanskrit –, , –, , – Ṛgvedic, Vedic Sanskrit –, , , – Classical Sanskrit ,
Sardinian , Scandinavian , n., , , Scotch Scots –, – Modern Scots Older Scots –, – Slavonic , , , –, n. Common Slavonic Old Church Slavonic , , –, n. Proto-Slavonic Ser-Bo-Croatian n., , , Serbo-Croatian , –, Spanish , , , , Tocharian (A, B) , , Tuscan (early Tuscan) Venetic
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of subjects A-movement see movement A0 -movement see movement adjunct(ion) , , , –, , , , , –, , , , n. Adpositional Phrase – adverb(ial) , , , , , , , , , –, , , , –, – analytic(ization) –, , , –, –, , , – antilocal(ity) , , – antisymmetry , , argument(al) –, , –, , n., , , –, –, , –, , , n., , –, , –, –, , , –, , , article , –, – see also definiteness articleless aspect(ual) , , , , , –, , –, , – auxiliary/auxiliaries , , , , –, –, –, n., –, , , , , , , –, , , n., , , , –, –, –, , , – see also negative auxiliary inversion auxiliary class , auxiliary support auxiliary verb , –, , , BE-auxiliary –, finite auxiliary , –, –, –, – modal auxiliary –, – negative auxiliary Bergin’s Construction , – binding , boundary , , , , categorial restriction, clausal conjunction see conjunction/coordinator chain –, , , –, check(ing) , , –, , , , , , –, n., n., , , , , , n., clause structure –, , , , –, , –, , , –, , –, , , ,
, , , –, –, , , , , , –, –, –, , , , , –, –, , , –, , –, –, –, – biclausal / monoclausal , , cartography , n., , , , , –, , see also left periphery contraction/expansion , –, , , , – structural/clausal position , , , , , , , , , cleft , –, –, – clitic , , , n., , , , , , , –, –, n., –, , n. see also conjunction/coordinator, enclitic; enclisis; Neg-clitic(ization)/ clitic negator clitic combinatorics cliticization , n., n., n., –, complement clause / complementation , , see also embedded clauses complementizer(s) , , n., –, , n., , , , , , , – complex predicate , , –, , –, compound tenses , , , , see also periphrastic tenses/constructions configurational(ity) , , , , –, confirmative reading of sempre – conjunction/coordinator –, , see also coordination; junction categorial restriction, clausal conjunction dual semantics of conjunction , emphatic conjunction enclitic , , , , , –, , n., – first-position conjunction , , – orthotone coordinator , , – postpositive/peninitial conjunction –, –, –, –, ,
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of subjects
conjunction/coordinator (cont.) prepositive/initial (head-initial) conjunction , , , –, , –, , , , reduplicative pattern – second-position conjunction , , n. Wackernagel’s Law , , n., –, Constant Rate Hypothesis , –, coordination –, , –, n. coordinate construction – double system of coordination –, embedded coordination – monosyndetic coordination , –, polysyndetic coordination , – copy theory of movement , , , –, corrective contexts – CP-domain/area , , , , , –, Criterial Freezing , defective , , , deictic –, , –, deletion / delete , , , , , , – scattered deletion , , demonstrative –, , n., , n. definiteness , , , , , n., , , –, definite article , , –, definite objects n. indefinite article , determiner , , , –, n., –, –, – determiner doubling , –, – discontinuous discontinuous noun phrases , –, , –, discontinuous structures , –, – discontinuous subjects , –, – phrasal discontinuity , , directionality –, –, –, discourse , n., , n., , –, , –, , n., n. do-support , , double system of coordination see coordination doubly filled , n., , drift –, , , , , , , economy n., , , n., , –, , , , , , n., –, embedded clauses , , –, –, , n., , , , n., n., –, –, – see also complement clause/ complementation
emphasis/emphatic , , , –, , –, , –, –, –, , Emphatic/Contrastive Stress Rule (E/CSR) , n. emphatic conjunction see conjunction/ coordinator empty –, , , see also null enclisis , , see also conjunction/ coordinator EPP , , , , –, , –, , , –, – evidential(lity) , , – see also particle, tense-aspect-mood-evidentiality particles extraction see movement expletive , –, – transitive expletive (constructions) , –, –, – failed change / successful change –, , , feature(s) , , , , , , , , , , , –, –, , –, , , , , , –, , , , –, –, , , , –, –, –, , –, –, , –, n., Final Over Final Constraint (FOFC) , –, , , , –, n. finite(ness) / non-finite , , –, , , n., –, , –, –, , , –, –, , , –, , , , –, , , , , first-position conjunction see conjunction focus , , , , , , , , –, –, , –, –, , , –, –, , , –, –, –, , n., –, , –, –, , , , –, , – contrastive focus , , –, , , , emphatic/contrastive focus –, , , information focus , –, , n., , – focalization / focus-movement , , identificational focus –, , free-choice construction/item , fronting argument fronting , , –, , –, – edge-fronting , –
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of subjects focus fronting –, –, , , –, –, , –, , – stylistic fronting – verb fronting / infinitival fronting , , , , functional (category/structure/item) –, –, , , –, , , , , n., –, , , –, –, , , –, , –, , , , , –, , , n., , –, , , –, , – general subordination marker – grammaticalization , –, –, –, , , –, , n., , –, , , , head adjunction , see also incorporation; movement, head movement head movement see movement Head Movement Constraint see movement Head Parameter , , –, Head Preference Principle , , head-final(ity) , –, , n., , , –, –, – head-initial(ity) , –, , , , , –, n., , –, –, – heavy NP-shift , incorporation , , , –, , –, n., , , , , n., , , – Specifier Incorporation Principle infinitive / infinitival , , , , –, , , , –, n., –, –, –, –, –, , inflected infinitive , –, information structure (IS) , –, , , , , , –, , , –, , , n. IP domain/area , , , , Jespersen’s cycle – Junction/J0 –, –, , – see also conjunction/coordinator last resort , –, laten-support , , –, –, , Left Branch Condition (LBC) , left branch extraction , –
left dislocation see topic left periphery , –, , –, , , –, –, , –, , –, –, –, , , , , , , , , –, , , , , , , –, –, , –, , , –, –, – lexical defectivity , license / licensing , , , , , , –, , n., , , – linearization / linear placement , , , , , , , – local(ly) , , , –, , , –, –, , , – logistic function / logistic peak – merge late merge merge over move n., , metalinguistic negation – modal (verb/auxiliary) , –, , –, –, –, , , , , –, , –, –, , –, , –, – modality , , , – modifier , , –, –, , –, –, –, , n., , –, –, , , –, , – see also verbal modifier monosyndetic coordination see coordination mood , , –, , morphemicity, mono-/bi- , , –, , –, –, movement see also fronting; particle, particle movement; preposing; scrambling; verbmovement; VP displacement/ movement; topic A-movement , , , , n., , , , , A0 -movement , –, n., antilocal movement , complement-to-specifier movement , , extraction , , , , , –, –, – head movement , , , , , Head Movement Constraint (HMC) –, – L-movement leftward movement , , local movement , – long movement loss/emergence of movement –, , , –, , , n., , , , , –, , , , –, –,
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of subjects
movement (cont.) movement operation , –, , , , , , , , phrasal movement / XP movement , , p-movement post-syntactic movement – predicate movement , , –, – remnant movement –, – rightward movement , –, , roll-up movement –, –, , –, –, – short movement , – subextraction , , Wackernagel movement , wh-movement , , , multiple subjects see subjects negation –, –, , , , –, , , –, , , –, , , –, n., , , , , –, , –, , , –, , –, –, negative auxiliary inversion , , –, , –, Neg-clitic(ization) / clitic negator –, –, Negative Concord , – node , , , n., , , , , , bounding node – nominalized verb nominalized sentence , –, – nominativus pendens –, noun phrase –, – NP-shift , Nuclear Stress Rule (NSR) null , –, , , n., , , , – see also empty; subject, null subject object object scrambling –, , –, , , , , – object shift , , , OV/VO , , , –, , , –, –, , – operator , , , – orthotone connective see conjunction/ coordinator particle auxiliary particle –, clause-chaining particle
discourse particle , enclitic particle initial particles , μ-particle, superparticle –, –, – particles , –, , – particle movement , – Q particle / question particle , , relativizing particle / relative particle , tense-aspect-mood-evidentiality (TAME) particles , , –, – verbal particle , , , , , – past participle , , , , –, , –, , , –, –, –, , n. periphrastic tenses/construction , , , – see also compound tenses phase(s) , , , , , –, –, , – Phase Impenetrability Condition (PIC) , – phonology / phonological , , , –, , , , , –, n., , , , –, , n. polarity , , , , –, polysyndetic coordination see coordination possessive / possessive constructions , –, , –, –, , possessor , , –, –, – dative-marked –, , –, – external , –, – extracted , –, , lexical n., – pronominal unmarked –, –, postpositive/peninitial conjunction see conjunction/coordinator pragmatic(s) , , n., –, n., , –, , , , , –, , –, –, , – predicate movement see movement preposing , prepositive/initial (head-initial) conjunction/ coordinator see conjunction probe-goal , –, , , procliticization , –, prosody , , , –, n. quantifier bare quantifier , , –, , –, distribuitive quantifier floating quantifier , n.
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of subjects indefinite quantifier negative quantifier , , – quantification(al) –, –, , quantifier raising , quantifiers , , , –, universal (bare) quantifier , –, –, –, , – reanalysis –, , , –, –, –, n., , , , , , –, , , , –, – tense-to-aspect reanalysis , – reduplication –, relative clause , –, , –, , –, –, –, , n., restrictive relative clause (RRC) –, –, relative cycle –, – relativization , , , –, relic syntax remnant movement see movement roll-up movement see movement scrambling –, , , –, , , , , – second-position , , , n., , , , n. second-position conjunction see conjunction/ coordinator, enclitic; Wackernagel’s Law semantics , , n., , , , , , n., , n., , , , , , , , , , , , , , , , , , –, , –, , –, –, , , shift , , , , , , , , , , , , , , , – specifier Spec,AspP , , n., –, Spec,CP , , , , –, , –, , n., –, –, –, , , –, –, , , Spec,DP , , , –, –, , , –, n. Spec,Fin(itness)P , , – Spec,Foc(us)P , , , , , , , Spec,ForceP , –, –, Spec,FP , , – Spec,Int(errogative)P , , –, Spec,IP n. Spec,JP Spec,ModP , , Spec,NP , n.
Spec,Operator Spec,pP SpecPredP , , , Spec,ShiftP – Spec,TP , , –, , , , , , , –, – Spec,Top(ic)P , –, , , , –, , , Spec,VoiceP , –, , –, , – Spec,VP , n., –, , , , , , , – Spec,vP , , , , , , , –, –, –, –, – Specifier Incorporation Principle see incorporation specifier positions / specifiers –, , , , , , , , , , , , , , split CP split DP n. split IP split noun phrase , – stative verbs , –, stress assignment –, subject discontinuous subjects , –, – expletive subjects , , multiple subjects , null subject n., n., , –, pre-/post-verbal subjects see subject, subject position subject position , –, , , –, –, n., , , , , –, , , –, , , –, subject–verb inversion see subject, subject position subjunctive , , , –, , – synthetic(ally) –, , –, , , –, , –, n., – synthetically infected / simple tenses , , –, TAME particles see particle, tenseaspect-mood-evidentiality (TAME) particles target –, –, , –, , , , , , n., , –, , , tense , , , –, , –, , n., , , terminal –, tmesis – transitive expletive see expletive
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Index of subjects
topic , –, , , , –, , , aboutness topic , , , , –, contrastive topic , , , –, –, –, familiar topic , –, , –, , – left dislocation , , , , topic shift , , – topicalization , , , , , , , , , , , , trigger , –, –, , –, , n., n., , n., –, –, , , , , , , , –, –, , –, typology/typological(ly) , , n., , , , , , , –, n. verb movement , –, –, , , , , , –, –, , –, –, , , , , AUX-to-Comp embedded/infinitival verb second (IV) –, , – Tense-to-Comp (T-to-C) , –, – verb fronting , , , , Verb-to-Comp (V-to-C) –, n., –, , , , , –, , Verb-to-Neg (V-to-Neg)
Verb-to-Tense (V-to-T) , –, –, –, –, , , , –, n. verb to Topic verb second (V) –, –, –, –, –, –, , , , –, , , , –, , –, –, –, , – verb-initial (V) , , , , , , n. verb second see verb movement, verb second (V) verbal modifier (VM) , , n., , n., – verbal particle see particle VO/OV see object, OV/VO VP displacement/movement –, – vP area n., , , – Wackernagel’s Law see conjunction/ coordinator wh-(items/features/structures) , , , , –, , , n., , , , , , , , , , –, wh-movement see movement word order word order change see chapter as roadmap neutral word order , , , , –, n., –, n. word order variation –, –, , –, –, , , –, –, , , , , –, , , , –,
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
OXFORD STUDIES IN DIACHRONIC AND HISTORICAL LINGUISTICS GENERAL EDITORS
Adam Ledgeway and Ian Roberts, University of Cambridge ADVISORY EDITORS
Cynthia Allen, Australian National University; Ricardo Bermúdez-Otero, University of Manchester; Theresa Biberauer, University of Cambridge; Charlotte Galves, University of Campinas; Geoff Horrocks, University of Cambridge; Paul Kiparsky, Stanford University; Anthony Kroch, University of Pennsylvania; David Lightfoot, Georgetown University; Giuseppe Longobardi, University of York; George Walkden, University of Konstanz; David Willis, University of Cambridge PUBLISHED
From Latin to Romance Morphosyntactic Typology and Change Adam Ledgeway Parameter Theory and Linguistic Change Edited by Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar Case in Semitic Roles, Relations, and Reconstruction Rebecca Hasselbach The Boundaries of Pure Morphology Diachronic and Synchronic Perspectives Edited by Silvio Cruschina, Martin Maiden, and John Charles Smith The History of Negation in the Languages of Europe and the Mediterranean Volume I: Case Studies Edited by David Willis, Christopher Lucas, and Anne Breitbarth Constructionalization and Constructional Changes Elizabeth Traugott and Graeme Trousdale Word Order in Old Italian Cecilia Poletto Diachrony and Dialects Grammatical Change in the Dialects of Italy Edited by Paola Benincà, Adam Ledgeway, and Nigel Vincent Discourse and Pragmatic Markers from Latin to the Romance Languages Edited by Chiara Ghezzi and Piera Molinelli
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Vowel Length from Latin to Romance Michele Loporcaro The Evolution of Functional Left Peripheries in Hungarian Syntax Edited by Katalin É. Kiss Syntactic Reconstruction and Proto-Germanic George Walkden The History of Low German Negation Anne Breitbarth Arabic Indefinites, Interrogatives, and Negators A Linguistic History of Western Dialects David Wilmsen Syntax over Time Lexical, Morphological, and Information-Structural Interactions Edited by Theresa Biberauer and George Walkden Syllable and Segment in Latin Ranjan Sen Participles in Rigvedic Sanskrit The Syntax and Semantics of Adjectival Verb Forms John J. Lowe Verb Movement and Clause Structure in Old Romanian Virginia Hill and Gabriela Alboiu The Syntax of Old Romanian Edited by Gabriela Pană Dindelegan Grammaticalization and the Rise of Configurationality in Indo-Aryan Uta Reinöhl The Rise and Fall of Ergativity in Aramaic Cycles of Alignment Change Eleanor Coghill Portuguese Relative Clauses in Synchrony and Diachrony Adriana Cardoso Micro-change and Macro-change in Diachronic Syntax Edited by Éric Mathieu and Robert Truswell
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
The Development of Latin Clause Structure A Study of the Extended Verb Phrase Lieven Danckaert Transitive Nouns and Adjectives Evidence from Early Indo-Aryan John J. Lowe Quantitative Historical Linguistics A Corpus Framework Gard B. Jenset and Barbara McGillivray Gender from Latin to Romance History, Geography, Typology Michele Loporcaro Clause Structure and Word Order in the History of German Edited by Agnes Jäger, Gisella Ferraresi, and Helmut Weiß Word Order Change Edited by Ana Maria Martins and Adriana Cardoso IN P R E P A R A T I O N Negation and Nonveridicality in the History of Greek Katerina Chatzopoulou Morphological Borrowing Francesco Gardani Indefinites between Latin and Romance Chiara Gianollo Nominal Expressions and Language Change From Early Latin to Modern Romance Giuliana Giusti The Historical Dialectology of Arabic Linguistic and Sociolinguistic Approaches Edited by Clive Holes A Study in Grammatical Change The Modern Greek Weak Subject Pronoun τος and its Implications for Language Change and Structure Brian D. Joseph Reconstructing Pre-Islamic Arabic Dialects Alexander Magidow Grammaticalization from a Typological Perspective Heiko Narrog and Bernd Heine
OUP CORRECTED PROOF – FINAL, 4/5/2018, SPi
Word Order and Parameter Change in Romanian Alexandru Nicolae Referential Null Subjects in Early English Kristian A. Rusten The History of Negation in the Languages of Europe and the Mediterranean Volume II: Patterns and Processes Edited by David Willis, Christopher Lucas, and Anne Breitbarth Verb Second in Medieval Romance Sam Wolfe Palatal Sound Change in the Romance Languages Diachronic and Synchronic Perspectives André Zampaulo