Microparameters in the Grammar of Basque [1 ed.] 9789027266422, 9789027208309

Citation preview

Microparameters in the Grammar of Basque Edited by Beatriz Fernández and Jon Ortiz de Urbina Language Faculty and Beyond

Internal and External Variation in Linguistics



John Benjamins Publishing Company

Microparameters in the Grammar of Basque

Language Faculty and Beyond Internal and External Variation in Linguistics issn 1877-6531 Language Faculty and Beyond (LFAB) focuses on research that contributes to a deeper understanding of the properties of languages as a result of the Language Faculty and its interface with other domains of the mind/brain. While the series will pay particular attention to the traditional tension between descriptive and explanatory adequacy, the series will also address issues such as the level of linguistic design, through new lines of inquiry often referred to as ‘physiological linguistics’ or ‘biolinguistics’. LFAB aims to publish studies from the point of view of internal and external factors which bear on the nature of micro- and macro-variation as, for example, understood in the minimalist approach to language.

For an overview of all books published in this series, please see https://benjamins.com/catalog/lfab

Editors Kleanthes K. Grohmann University of Cyprus

Pierre Pica CNRS, Paris

Advisory Board Paola Benincà

Anders Holmberg

Cedric Boeckx

Lyle Jenkins

Guglielmo Cinque

Richard K. Larson

Noam Chomsky

Andrew Ira Nevins

University of Padova, Italy ICREA/University of Barcelona, Spain University of Venice, Italy Massachusetts Institute of Technology, Cambridge, USA

Stephen Crain

Macquarie University, Sydney, Australia

Marcel den Dikken

Hungarian Academy of Sciences & Eötvös Loránd University

Naama Friedmann

Tel Aviv University, Israel

University of Newcastle, UK Biolinguistics Institute, Cambridge, USA Stony Brook University, USA University College London, UK

Alain Rouveret

University of Paris VII, France

Esther Torrego

University of Massachusetts, Boston USA

Anna Papafragou

University of Delaware, Newark, USA

Akira Watanabe

University of Tokyo, Japan

Volume 13 Microparameters in the Grammar of Basque Edited by Beatriz Fernández and Jon Ortiz de Urbina

Microparameters in the Grammar of Basque Edited by

Beatriz Fernández UPV/EHU

Jon Ortiz de Urbina Deusto University

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

doi 10.1075/lfab.13 Cataloging-in-Publication Data available from Library of Congress isbn 978 90 272 0830 9 (Hb) isbn 978 90 272 6642 2 (e-book)

© 2016 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com

Table of contents Part 1.  Introductory chapters Microparameters in the Grammar of Basque: Looking back and moving forward Beatriz Fernández & Jon Ortiz de Urbina On Basque dialects José Ignacio Hualde

3 15

Part 2.  Case and Agreement On the relation between ergativity, stativity and the verbal configuration of Basque Ane Berro

39

Datives in Basque bivalent unergatives Jon Ortiz de Urbina & Beatriz Fernández

67

Differential object marking in Basque varieties Beatriz Fernández & Milan Rezac

93

Gaps and stopgaps in Basque finite verb agreement Milan Rezac

139

Part 3.  Determiners Nominals in Basque and their existential interpretation Urtzi Etxeberria

195

Part 4.  Word order and left periphery Strategies of verb and verb phrase focus across Basque dialects Arantzazu Elordieta & Bill Haddican

221

Strategies for argument and adjunct focalization in Basque Aritz Irurtzun

243

 Microparameters in the Grammar of Basque

Microsyntactic variation in the Basque hearsay evidential Ricardo Etxepare & Larraitz Uria

265

Index of Basque varieties mentioned in the text

289

Name Index

291

Subject Index

297

part i

Introductory chapters

Microparameters in the Grammar of Basque Looking back and moving forward Beatriz Fernández & Jon Ortiz de Urbina UPV/EHU / Deusto University

1.  Microparameters in the Grammar of Basque This book is an endeavor to present and analyze some standard topics in the grammar of Basque from a microcomparative perspective.1 From case and agreement to word order and the left periphery, and including an incursion into determiners, the book combines fine-grained theoretical analyses with empirically detailed descriptions. The title, Microparameters in the Grammar of Basque, while self explanatory, also points back at a book published at the end of the eighties, Ortiz de Urbina’s (1989) Some Parameters in the Grammar of Basque (Foris, Dordrecht, The Netherlands), an early parametric exploration of issues such as ergativity and configurationality coached in the Principles and Parameters framework (Chomsky 1981). Since that time, our knowledge of (Basque) grammar has increased exponentially, and parameters themselves, or at least their construal as macroparameters (Baker 2008), as well as their role in the study of linguistic variation, have sparked important discussion (see for instance, Newmeyer 2005 and Roberts and Holmberg 2005). The advent of the Minimalist Program (Chomsky 1993) and the focus on language internal variation (Kayne 2000) gave a new twist to studies on linguistic variation, which thrive in the Basque Country. This is partly due to the sheer linguistic diversity found in such a small territory. Without standardizing pressures until the end of the 20th century, and largely restricted to colloquial, spoken uses, traditional dialectology, from Bonaparte’s pioneering work in the

.  This work has been partially supported by the Basque Government (IT665-13) and the Spanish Ministry of Economy and Competitiveness (FFI2014-51878-P). The research leading to these results has also received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement nº 613465. The editors would also like to thank the different anonymous reviewers whose comments and suggestions greatly improved each and all of the papers in this volume.

doi 10.1075/lfab.13.01fer © 2016 John Benjamins Publishing Company



Beatriz Fernández & Jon Ortiz de Urbina

19th ­century to our days, presents a picture of Basque as divided into numerous varieties and sub-varieties. Standardization and Basque-language schooling in the Spanish Basque Country, however, are deeply altering that situation, and dialectal variation is clearly on the wane. Nevertheless, some dialects and varieties are still robust and contact with the standard dialect is actually adding complexity to traditionally existing variation. The combination of empirically rich dialectal data domains and the presence of phenomena which are often uncommon within European languages (‘free’ word order, extensive agreement with subjects, direct and indirect objects, as well as with addressees, ergative marking, etc.) make of Basque fertile ground for studies dealing with language variation from a syntactic point of view, as this volume strives to show. In this introduction, we will first provide a minimal background to the development of research on major aspects of Basque syntax where parameter-based analyses have been proposed, turning later to an equally cursory description of the contributions of the articles in this volume. We expect this will help the reader to better understand and locate the latter in the larger picture of Basque generative studies in general and parametric variation in particular. 2.  Looking back and moving forward While the boom in syntactic research on Basque dates mostly to the late 80’s and 90’s, many (if not all) of the empirical domains analyzed in this volume already featured prominently in earlier work, and illustrate well what we consider progress in both our understanding of Basque and of syntax/grammar in general. Ergativity was the topic of de Rijk’s MIT manuscript as early as in 1966, as well as of Sarasola’s (1977) discussion arguing for the subject status of the ergative argument. Basque unergatives and unaccusative predicates were prominent in Levin’s (1983) dissertation, and Ortiz de Urbina (1989) attempted a general description of Basque within a parametric approach, addresssing ergativity (Dixon 1979), the configurationality parameter (Hale 1981, 1983), the pro-drop parameter (Chomsky 1981; Rizzi 1982) or the Focus parameter. However, the macroparametric approach declined with the inception of the Minimalist Program at the beginning of the nineties (Chomsky 1993, 1995). As Boeckx (2011) puts it, there is some ‘uneasiness’ between minimalism and the Principles and Parameters model. Complex parametrized principles contrast with the ‘minimal’ type of basic operations that figure at the core of current theoretical discussion in the generative framework. An important move to constrain parameters was the so-called Borer-Chomsky conjecture (Baker 2008), originating with Borer’s (1984) hypothesis, later adopted by Chomsky (1995), that parametric



Microparameters in the Grammar of Basque

v­ ariation was restricted to differences in the functional heads. This reoriented much generative research on linguistic variation, from very general linguistic properties to much more specific properties of functional heads, i.e., ‘microparameters’. Thus, some of the most relevant discussions on ergativity from the early nineties (Laka 1993b, Bobaljik 1993 among others) were approached from this microparametric perspective. The same perspective was also generally assumed in the study of Basque object agreement in general and dative agreement in particular. The presence or not of a third Agreement head as discussed by Collins and Thráinsson (1993) for Georgian and also proposed by Hualde (1986) and Fernández (1997) for Basque (the latter capitalizing on Larson’s (1988) VP-shells), centered the debate on the appropriate analysis of Basque agreement facts. While typologically uncommon in Europe, direct and indirect object agreement patterns in Basque bring to light analytical and theoretical problems also addressed in object complementation analyses in general, so Basque data related to datives and dative agreement were fruitfully connected with the theoretical issues surrounding phenomena such as the Me-lui (Person Case) Constraint and clitic phenomena in general or double object constructions (Bonet 1991; Laka 1993a; Albizu 1997; Ormazabal and Romero 1998). At the beginning of this century, running parallel to minimalist studies, microvariation in agreement patterns spawned a large amount of work and interest in Basque linguistics. The number of dialectal agreement phenomena and subpatterns is proportional to the ‘morphological’ complexity of standard patterns, and had been only partially studied from a traditional dialectological point of view. Once functional categories began to increase in number and inflectional morphology began to be analyzed from a syntactic perspective, novel and systematic analyses of many dialectal phenomena from a microparametric view began to be published. Some of those phenomena had earlier either been overlooked or, when noticed, branded as ‘deviations’ and even stigmatized, and only under the new microvariationist perspective did they receive the attention they deserved. Among these were phenomena such as dative displacement (Fernández 2001; Rezac 2008a; Rezac and Fernández 2013) and displacement in general (Rezac 2006, 2008b), differential object marking (Fernández and Rezac 2010, this volume; Odria 2014, in progress) and dative agreement drop (Etxepare and Oyharçabal 2013). The change from macroparametric to microparametric analyses of linguistic diversity is also clear in the realm of word order. Basque non-focal major constituents display what is often characterized as free word order, and nonconfigurationality as in Hale (1981, 1983) figured prominently in early GB discussions (Rebuschi 1984; Ortiz de Urbina 1989), even though constituent splitting, one of the hallmarks of the macroparameter, was absent from the language. In spite of the rather salient role accorded to nonconfigurationality in Chomsky (1981), its





Beatriz Fernández & Jon Ortiz de Urbina

technical implementation as a macroparameter was ill at ease with standard analytical practice at the time, and the term was often loosely used as a descriptive label for ‘free word order’. Scrambling, dislocations and different types of movement operating on standard configurational structures were more often resorted to as explanations for apparently free linearizations. This was so mostly because, in spite of the apparent freedom of word order, Basque was shown to exhibit fairly standard configurational c-command relations in binding, crossover, etc., just as it displays ‘nominative’ syntax in spite of the pervasive morphological ergativity. With regard to constituent internal order, where ‘freedom’ is largely absent, even the head parameter, which captured in a general way intuitions about ‘harmonic’ orders going back to Greenberg (1963), and which had in Basque a rather well-behaved example of a head-last language, gave way to antisymmetric analyses after Kayne (1994), where remnant roll-up movements on a universal head+complement structure, conspired, as it were, to produce OV harmonic orders. The reception of antisymmetric syntax in Basque studies was rather mixed, judging from the articles presented in Arteatx et al. (2008). The proliferation of functional categories that made it possible, on the other hand, increased the descriptive possibilities of microparameters, although it also diluted the strength of Borer’s initial intuition. While non-focal constituents display virtual freedom of permutation, Basque foci and wh-words occur preverbally in a structure which was recognized and became part of traditional Basque grammar since Altube’s (1929) description. Generative accounts of these adjacency facts have not usually followed macroparametric approaches like that in Kiss’s (1995) discourse configurationality hypothesis, which assumes the possibility of a drastically different clausal organization in some languages, based on semantics and discourse as opposed to ϑ-roles and case. Rather, they have usually identified a functional specifier in the left periphery to host the focus and/or the wh-word (CP or Force/Focus phrases as in Rizzi 1997), achieving adjacency with the verbal element by movement of the latter to the head of the functional projection (Ortiz de Urbina 1989) or by principles that ensure intervening elements may not be overt (Uriagereka 1992). Microvariation in this area is also attracting much recent research, as the papers in this volume show. The following section gives a brief overview of the latter. 3.  Contributions to the volume The chapters of this book cover a range of topics related to three main areas: case and agreement, determiners, and word order and left periphery. All the contributions to the volume analyze in detail the exuberant variation attested in Basque



Microparameters in the Grammar of Basque

language and its varieties, and in order to do so, a microparametric perspective is assumed. Although the contributions focus mainly on Basque data, crosslinguistic evidence is also presented and discussed. After all, the goal pursued in this book is to attempt to explain variation in Basque as a particular instantiation of human language variation in general. On Basque dialects, an introductory chapter to Basque dialects by José ­Ignacio Hualde, provides the reader with an overview of Basque internal variation. Hualde exhaustively presents the most salient morphosyntactic properties of Basque varieties, distinguishing both eastern (east vs. center and west) and western (western vs. center and east) properties. A phonological inventory of features is also provided. In addition, the overview includes interesting notes on the origin and historical evolution of Basque varieties, and the consequences of Basque standardization. From Louis-Lucien Bonaparte’s (1831–1891) Map of the Seven Basque Provinces to Zuazo’s (2013, 2014) current classification of Basque dialects, the interest on the study of Basque internal variation has not declined. The first of the chapters related to case and agreement is authored by Ane Berro. On the relation between ergativity, stativity and the verbal configuration of Basque presents a novel and cutting-edge analysis of (Basque) ergativity. Departing from analyses that establish correlations between ergative case and agentivity (Holmer 1999, Laka 2006), Berro shows that ergative subjects can be holders of verbal states, and even subjects of non-verbal stative predicates. In order to explain such instances, Berro further proposes that ergative case is related to a head of central coincidence. This head incorporates onto BE, yielding a HAVE auxiliary à la Kayne (1993). Thus, Berro’s analysis, also presented in her recent Ph.D. dissertation (Berro 2015), is radically different from those which consider little v to be responsible for ergative assignment (Legate 2008 and Aldridge 2004, among others). Her proposal is also generalized to eventive predicates headed by the suffix -tu. Contrary to the standard assumption in the Basque grammatical tradition, Berro convincingly shows that eventive predicates headed by -tu are nominal in nature, another piece of evidence that supports the lack of correlation between ergativity and verbal categories. Two chapters deal with dative-marked objects. Datives in Basque bivalent unergatives, coauthored by Jon Ortiz de Urbina and Beatriz Fernández, focuses on the analysis of dative-marked objects and their syntactic behavior. The pattern exhibited by these objects is unexpected in both ergative and accusative languages, as absolutive and accusative, respectively, are the cases expected to be assigned to the sole object. Instead, in both ergative and accusative languages, some predicates mark dative what looks like the first object. Thus, Basque bivalent unergative predicates of the deitu ‘call’ type, for instance, give rise to configurations with ergative subjects and dative objects (Etxepare 2003). These ergative-dative





Beatriz Fernández & Jon Ortiz de Urbina

c­ onfigurations are also accompanied by ditransitive-like verbal forms including ergative and dative agreement markers. Basque bivalent unergative predicates fall into the semantic classes identified by Blume (1998) and are, therefore, an exponent of crosslinguistic variation and not an idiosyncratic Basque pattern. In the chapter, it is argued that these dative-marked objects are not only morphologically but also syntactically similar to indirect objects, as shown in several syntactic contexts, such as: (i) secondary predication, (ii) impersonal/passive clauses, (iii) adnominals, and (iv) causativization and relativization (see also McFadden 2004 for ­German). In a nutshell, apparent ‘first’ complements actually behave like ‘second’. A second article deals with dative-marked objects, namely, Differential object marking in Basque varieties, which presents the other side of the coin discussed in the previous article. Contrary to ‘first’ dative complements, these dialectal dative objects align with absolutive objects, as also observed and argued by Odria (2014). The chapter, coauthored by Beatriz Fernández and Milan Rezac, is an attempt to analyze a highly stigmatized Basque dialectal exponent of a typologically wellattested phenomenon (Aissen 2003). In particular, Basque differential object marking (DOM) is strikingly similar to other well-known instances of the phenomenon, such as Spanish leísmo and a-marking (Torrego 1998; Ormazabal and Romero 2007). The chapter also argues for the structural nature of the dative, as it is independent of argumenthood relations and so, available for Exceptional Case Marking. It is further shown that Basque DOM is sensitive to and parametrizable by tense, finiteness and agreement (differing from Spanish and Hindi-Urdu DOM). Thus, DOM objects, although similar to canonical absolutive objects, depart from the latter in their sensitivity to properties of the entire agreement complex hosting the probe (from little v to T). The section on case and agreement closes with a chapter by Milan Rezac entitled Gaps and stopgaps in Basque finite verb agreement. In this chapter, Rezac presents an exhaustive overview of gaps in finite verb agreement, analyzing their causes and the consequences for the modular nature of the language faculty. Three are the gaps under analysis: first of all, interpretive gaps in Condition B contexts, We chose *us/(*)me, with no correlation in Basque agreement; second, gaps related to conditions on syntactic dependencies (particularly those resulting from the impossibility of dative agreement with a dative across an intervening absolutive) and to the Person Case Constraint, which bans certain combination of 1st and 2nd person absolutives and datives; and third and last, morphological gaps attested in western varieties, where some or all 1st person plural and 2nd person combinations do not exist and where stopgaps emerge instead, that is, forms in which some or all features of one of the arguments do not appear.



Microparameters in the Grammar of Basque

As a bridge between the first section on case and agreement and the section on word order and left periphery, a chapter on determiners is also included. Urtzi Etxeberria’s Nominals in Basque and their existential interpretation distinguishes the Souletin dialect, where bare nouns are allowed in argument position, from the remaining dialects, where bare nouns are disallowed. Assuming that Souletin keeps a system historically previous to the one which developed in the rest of the dialects (Michelena 1964), Etxeberria argues that this historical change is due to the semantic weakening and subsequent loss of a null D. Etxebarria further argues that the appearance of the overt D in Standard Basque in the indefinite/existential object position is the consequence of a reanalysis of the available definite article -a as a D with existential interpretation (with narrow scope). The last section in the volume brings together three contributions on word order and the left periphery. Two articles deal with focalization, one of the major organizing factors for word order in Basque. Since foci occupy a preverbal position, verb (phrase) focalization itself differs slightly from argument and adjunct focalization, and Arantzazu Elordieta and Bill Haddican’s article Strategies of Verb and Verb Phrase Focus across Basque Dialects delve on the four possible strategies that can be found across dialects, namely verb doubling, insertion of a particle ba, insertion of dummy egin ‘do’ and V1. The analysis provides a simple and unifying approach which hinges on PF requirements that finite verbs do not bear main stress and that instead the focalized element must contain the element bearing main stress. The four strategies would then follow from repair operations at PF which spell out different parts of the derivations involved so that those requirements are met. The second article on word order and the left periphery, Aritz Irurtzun’s Strategies for argument and adjunct focalization in Basque, offers a summary of the microparametric variation found across Basque dialects for both Wh-questions and focalization, paying special attention to the emergence of new strategies, whose syntactic and semantic characteristics are explored in detail and contrasted with the standard strategies with which they coexist. The latter, described as A’-movement of both wh-words and foci to a functional projection in the left periphery and T-to-C rendering the observed adjacency, display earmarks for movement, such as cyclicity, island effects and the possibility of complex clausal pied-piping of the whole clause containing the wh-word or focus. When dialectal microvariation is taken into account, however, a more complex picture emerges. Thus, young Labourdin Basque speakers employ a wh-in-situ strategy, without overt movement and, consequently, without V2 effects. This strategy displays properties similar to French wh-in-situ, such as intervention effects with negation, the inability of argumental wh-phrases to remain in situ in wh-islands and their ability to move out of them and, finally, the grammaticality of wh-words in this strategy even when they occur within strong islands. The similarities point



 Beatriz Fernández & Jon Ortiz de Urbina

to contact with French as a possible trigger for the emergence of this new strategy. Finally, the article describes two further types of dialectal ‘reinforced’ focalization strategies, which are claimed to be more contrastive and presuppositional than the standard strategies. These are the ‘rightward’ focalization of southern dialects and a Navarro-Labourdin strategy where T-to-C only moves the auxiliary, leaving the lexical verb non-adjacent with the focus or wh-word. More restrictive focalization strategies are also outlined, providing a particularly complex and rich picture of this area of Basque microvariation. A third article is included in the word order and left periphery section, Ricardo Etxepare and Larraitz Uria’s Microsyntactic variation in the Basque hearsay evidential. The article deals with the hearsay evidentiality marker omen ‘reportedly, they say’, present in central and eastern dialects but displaying more complex word order possibilities in the latter dialects, spoken on the French side of the Basque Country. Observed differences are attributed to a syntactic microparameter, according to which omen would be part of the ‘clausal spine’ in central varieties but not in eastern ones. In the former, this would be a complex morphological element corresponding to two adjacent heads preceding the temporal and modal projections of the auxiliary. In the latter varieties, on the other hand, omen would constitute a separate syntactic term associated with an independent Evidentiality head. The independent nature of eastern omen would then account for its richer distributional possibilities, accounted for in terms of local computations arising from Agree. Thus, among other features, eastern omen may occupy more than one possible position in the clause, it does not necessarily co-occur with a finite auxiliary, and it may be used as a parenthetical element. The article illustrates well the drive towards microparametric explanations of language variation, which are here shown to be derivable from the ability of functional projections such as EvP to occupy different positions in the cartography of the clause. We believe that the articles in this book show that Basque microvariation in general presents us with a fascinating laboratory to experiment with a vast and rich empirical domain, from which descriptively nuanced and theoretically relevant analyses can derive. Like all theoretical constructs, always under revision, the notion of parameter itself or its ability to account for language variation can and has been called into question, hence Boeckx’ (2011) uneasiness and even Newmeyer’s strong rejection (2004, 2005). Nobody knows for certain what the future will look like, but if the future is already written in the recent past or even in the present, this volume shows that microparametric approaches (as in Kayne 2000, 2005) to the linguistic variation in Basque continue to return a rich yield. It also hopes to vindicate, if need be, the use of microdialectal data both for descriptive coverage and for theoretical inspiration. After all, we believe that, adapting Stephen Jay Gould’s dictum, the message of language lies in generality, but its beauty lies in detail.



Microparameters in the Grammar of Basque

References Aissen, Judith. 2003. “Differential object marking: Iconicity vs. Economy.” Natural Language and Linguistic Theory 21: 435–483.  doi: 10.1023/A:1024109008573 Albizu, Pablo. 1997. “Generalized person case constraint: A case for a syntax-driven i­ nflectional morphology.” In Theoretical issues on the morphology-syntax interface [Supplements of ASJU XL], Myriam Uribe-Etxebarria and Amaia Mendikoetxea (eds.), 1–33. Donostia: UPV/EHU. Aldridge, Edith. 2004. Ergativity and Word Order in Austronesian Languages. Ph.D. dissertation, Cornell University. Altube, Severo. 1929. Erderismos. Bermeo: Gaubeka. Arteatx, Iñigo, Artiagoitia, Xabier and Elordieta, Arantzazu (eds.). 2008. Antisimetriaren hipotesia vs. buru parametroa: euskararen oinarrizko hitz hurrenkera ezbaian [The antisymmetry hypothesis vs. the head parameter: Basque basic word order under discussion]. Bilbo: UPV/EHU. Berro, Ane. 2015. Breaking Verbs. From Event Structure to Syntactic Categories in Basque. Ph.D. dissertation, UPV/EHU and Université Bordeaux Montaigne (UBM). Baker, Mark. 2008. “The macroparameter in a microparametric world.” In The Limits of Syntactic Variation, Theresa Biberauer (ed.), 351–374. Oxford: Oxford University Press.

doi: 10.1075/la.132.16bak

Blume, Kerstin. 1998. “A contrastive analysis of interaction verbs with dative complements.” Linguistics 36 (2): 253–280.  doi: 10.1515/ling.1998.36.2.253 Bobaljik, Jonhattan D. 1993. “On ergativity and ergative unergatives.” In Papers on Case and Agreement II [MITWPL 18], Jonathan D. Bobaljik (ed.), 45–88. Cambridge, MA: MIT Press. Boeckx, Cedric. 2011. “Approaching parameters from below.” In The Biolinguistic Enterprise: New Perspectives on the Evolution and Nature of the Human Language Faculty, Anna Maria di Sciullo and Cedric Boeckx (eds.), 205–221. Oxford: Oxford University Press. Bonet, Eulalia. 1991. Morphology after Syntax: Pronominal Clitics in Romance. Ph.D. d ­ issertation, MIT. Borer, Hagit. 1984. Parametric syntax: Case Studies in Semitic and Romance languages. Dordrecht: Foris. Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris. Chomsky, Noam. 1993. “A minimalist program for linguistic theory.” In The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger, Kenneth L. Hale, and Samuel Jay Keyser (eds.), 1–52. Cambridge, MA: MIT Press. Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: The MIT Press. Collins, Chris and Thráinsson, Höskuldur. 1993. “Object shift in double object constructions and the theory of case.” In Papers on Case and Agreement II [MITWPL 19], Colin Philips (ed.), 131–174. Cambridge, MA: MIT Press. de Rijk, Rudolph.1966. “Redefining the ergative.” Ms., MIT. Dixon, Robert M.W. 1979. “Ergativity.” Language 55: 59–138.  doi: 10.2307/412519 Etxepare, Ricardo. 2003. “Valency and argument structure in the Basque Verb.” In A grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 363–425. Berlin: Mouton de Gruyter. Etxepare, Ricardo and Oyharçabal, Beñat. 2013. “Datives and adpositions in north-eastern Basque.” In Variation in Datives: A Microcomparative Perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 50–95. New York: Oxford University Press.



 Beatriz Fernández & Jon Ortiz de Urbina Fernández, Beatriz. 1997. Egiturazko Kasuaren Erkaketa Euskaraz. Bilbo: UPV/EHU. Fernández, Beatriz. 2001. “Absolutibo komunztaduradun ergatiboak, absolutibo komunztaduradun datiboak: ergatiboaren lekualdatzetik datiboaren lekualdatzera.” In Kasu eta komunztaduraren gainean. On Case and Agreement, Beatriz Fernández and Pablo Albizu (eds.), 147–165. Bilbo: UPV/EHU. Greenberg, Joseph H. 1963. “Some Universals of Grammar with Particular Reference to the Order of Meaningful Elements.” In Universals of Human Language, Joseph H. Greenberg (ed.), 73–113. Cambridge, MA: MIT Press. Hale, Kenneth L. 1981. On the Position of Walbiri in a Typology of the Base. Indiana Univ. Linguistics Club, Bloomington Indiana. Hale, Kenneth L. 1983. “Warlbiri and the grammar of non-configurational languages.” Natural Language and Linguistic Theory 1: 5–47.  doi: 10.1007/BF00210374 Holmer, Anders. 1999. “An Active Analysis of Basque Ergativity.” Fontes Linguae Vasconum 31: 189–225. Hualde, José Ignacio. 1986. “Case asignment in Basque.” ASJU XXII: 313–330. Kayne, Richard S. 1994. The Antisymmetry of Syntax. Cambridge, MA: The MIT Press. Kayne, Richard S. 1993. “Towards a modular theory of auxiliary selection.” Studia Linguistica 47 (1): 3–31.  doi: 10.1111/j.1467-9582.1993.tb00837.x Kayne, Richard S. 2000. Parameters and Universals. New York: Oxford University Press. Kayne, Richard S. 2005. “Some notes on comparative syntax, with special reference to E ­ nglish and French,” In The Oxford Handbook of Comparative Syntax, Guglielmo Cinque and ­Richard S. Kayne (eds.), 3–69. New York: Oxford University Press. Kiss, Katalin É. 1995. Discourse Configurational Languages. Oxford: Oxford University Press. Laka, Itziar. 1993a. “The Structure of Inflection: a Case Study in X0 Syntax.” In Generative Studies in Basque Linguistics, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 21–70. ­Amsterdam: John Benjamins.  doi: 10.1075/cilt.105.02lak Laka, Itziar. 1993b. “Unergatives that assign ergative, unaccusatives that assign accusative.” In Papers on Case and Agreement I [MITWPL 18], Jonathan D. Bobaljik and Colin Phillips (eds.), 149–172. Cambridge, MA: MIT Press. Laka, Itziar. 2006. “On the nature of case in Basque: structural or inherent?” In Organizing Grammar. Linguistic Studies in Honor of Henk van Riemsdijk, Hans Broekhuis, Norbert Corver, Riny Huybregts, Ursula Kleinhenz and Jan Koster (eds.), 374–382. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110892994.374 Larson, Richard K. 1988. “On the double object construction.” Linguistic Inquiry 19 (3): 335–391. Legate, Judith. 2008. “Morphological and Abstract Case.” Linguistic Inquiry 39.1: 55–101.

doi: 10.1162/ling.2008.39.1.55

Levin, Beth. 1983. On the Nature of Ergativity, Ph.D. dissertation, MIT. McFadden, Thomas. 2004. The Position of Morphological Case in the Derivation: a Study on the Syntax-Morphology Interface. Ph.D. dissertation, Univ. of Pennsylvania. Michelena, Luis. 1964. Sobre el pasado de la lengua vasca. Donostia: Auñamendi. Newmeyer, Frederick J. 2004. “Against a parameter-setting approach to language variation.” In Language Variation Yearbook 4, Pierre Pica, Johan Rooryck and Jeroen van Craenenbroeck (eds.), 181–234. Amsterdam: John Benjamins.  doi: 10.1075/livy.4.06new Newmeyer, Frederick J. 2005. Possible and Probable Languages. A Generative Perspective on Linguistic Typology. New York: Oxford University Press.

doi: 10.1093/acprof:oso/9780199274338.001.0001



Microparameters in the Grammar of Basque 

Odria, Ane. 2014. “Differential Object Marking and the nature of dative case in Basque varieties.” Linguistic Variation, 14–2. Amsterdam/Philadelphia: John Benjamins. 289–317.

doi: 10.1075/lv.14.2.03odr

Ormazabal, Javier and Romero, Juan. 1998. “On the Syntactic Nature of the me-lui and the Person-Case Constraint.” ASJU 32 (2): 415–433. Ormazabal, Javier and Romero, Juan. 2007. “Object agreement restrictions.” Natural Language and Linguistic Theory 25: 315–347. Ortiz de Urbina, Jon. 1989. Some Parameters in the Grammar of Basque. Dordrecht: Foris. Rebuschi, Georges. 1984. Structure de l’énoncé en basque. Paris: SELAF. Rezac, Milan. 2006. “Agreement displacement in Basque.” Ms., UPV/EHU. Rezac, Milan. 2008a. “The forms of dative displacement: From Basauri to Itelmen.” In Gramatika jaietan, Xabier Artiagoitia and Joseba A. Lakarra (eds.), 709–724. Bilbo: UPV/EHU. Rezac, Milan. 2008b. “The syntax of eccentric agreement: The Person Case Constraint and Absolutive Displacement in Basque.” Natural Language and Linguistic Theory 26: 61–106. doi: 10.1007/s11049-008-9032-6 Rezac, Milan and Fernández, Beatriz. 2013. “Dative Displacement in Basque.” In Variation in Datives: a Microcomparative Perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 256–282. New York: Oxford University Press. Rizzi, Luigi. 1982. “Null objects in Italian and the theory of pro.” Linguistic Inquiry 17–3: 501–557. Rizzi, Luigi. 1997. “The fine structure of the left periphery.” In Elements of Grammar, Liliane Haegeman (ed.), 281–337. Dordrecht: Kluwer.  doi: 10.1007/978-94-011-5420-8_7 Roberts, Ian and Anders Holmberg. 2005. “On the role of parameters in Universal Grammar: A reply to Newmeyer.” In Organizing Grammar: Linguistic Studies in honor of Henk van Riemsdijk, Hans Broekhuis, Norbert Corver, Riny Huybregts, Ursula Kleinhenz and Jan Koster (eds.), 538–553. Berlin: Mouton de Gruyter. Sarasola, Ibon. 1977. “Sobre la bipartición inicial en el análisis de consituyentes.” ASJU 11–1. 49–90. Torrego, Esther. 1998. The Dependencies of Objects. Cambridge, MA: MIT Press. Uriagereka, Juan. 1992. “The syntax of movement in Basque.” In Syntactic Theory and Basque Syntax, Joseba A. Lakarra and Jon Ortiz de Urbina (eds.), 417–445. Donostia: Gipuzkoako Foru Aldundia. Zuazo, Koldo. 2013. The dialects of Basque. Reno: Univ. of Nevada. Zuazo, Koldo. 2014. Euskalkiak. Donostia: Elkar.

On Basque dialects José Ignacio Hualde University of Illinois at Urbana-Champaign

This chapter offers an overview of geographical variation in Basque. After discussing several proposals for the classification of Basque dialects, the main phenomena showing dialectal variation are introduced. Variation in morpho‑syntactic features, nominal inflection, verbal inflection, derivational morphology and phonology are discussed in separate sections. Two additional sections are devoted to the historical development of Basque dialects and to the nature of the written regional standards known as “literary” dialects. Keywords:  Basque dialects; dialectology; isoglosses; morpho-syntactic variation; phonological variation; dialectal classification

1.  Geographical variation and Basque dialects1 The first authors to write in Basque had to confront the difficulty of selecting what words, forms of words, and constructions to use, given the fact that the language, like others that lack a standard variety, differed “almost from house to house,” in the words of the first Basque translator of the Bible, Joanes Leizarraga (1571). It is indeed the case that some differences were and are found when we compare the spoken Basque of any two towns, even neighboring ones. If we undertake the job of listing and systematizing geographical differences, however, it is apparent that differences among Basque dialects are relatively superficial; certainly not large enough to have made the creation and spread of a common standard variety an unviable project or to make communication in Basque between speakers from different geographical areas impossible (after a longer or shorter accommodation period).

.  This work has been partially supported by the Spanish Ministry of Economy and Competitiveness (FFI2014-51878-P9). For comments I am grateful to Koldo Zuazo, two anonymous reviewers and the editors of this volume. All errors are mine. The following glosses have been used: a ‘absolutive’, e ‘ergative, d ‘dative’, det ‘determiner’, aux ‘auxiliary’, ipfr ‘imperfective participle’.

doi 10.1075/lfab.13.02hua © 2016 John Benjamins Publishing Company

 José Ignacio Hualde

The very real difficulties that Basque speakers from different regions have nonetheless often found in communicating with each other at first encounter are for the most part due to relatively recent phonological changes, together with some small differences in vocabulary and even smaller ones in morphology and syntax. Now that standard Basque is widely available, these difficulties in communicating across dialects have become much smaller, as many speakers are aware of ways of making their speech more intelligible to interlocutors from other regions by approaching the standard. Educated Basque speakers from distant areas no longer experience much trouble in communicating with each other in Basque (except, perhaps, those deriving from bilingualism in French vs. Spanish). In a language, like Basque, spoken in contiguous territories, dialectal areas and boundaries do not have the same reality as political and administrative entities and their borders. The question of how many Basque dialects there are and where exactly their borders lie can have no objective answer. However, even if the reality of dialects as linguistic varieties spoken in territories with well-defined borders is fictional, dialects are useful fictions. First of all, dialects have a certain reality for the speakers of the language, who are often acutely aware of different ways of speaking their language and typically associate these differences with specific regions. Secondly, in the study of a language, stating that a given phenomenon is a feature of a given dialect provides us with useful information, even if we cannot determine from this statement what the exact geographical extension of this phenomenon may be. How many dialects is it, then, useful to distinguish within the Basque-speaking territory? The first researcher to seriously engage in Basque dialectology, Louis-Lucien Bonaparte (1813–1891), distinguished eight dialects, each of which he subdivided into a number of subdialects, which, in their turn, included geographically smaller varieties. His Carte des sept provinces basques (Map of the Seven Basque Provinces) showing the geographical extension of Basque dialects, subdialects and varieties, which was first published in 1869, has been reproduced profusely up to the present day and can be readily found online. Later researchers have tended to reduce the number of main dialects, which Bonaparte sometimes established on the basis of very few features. R. M. de Azkue, first president of the Basque Academy and author of an influential dictionary (1905–6), modified Bonaparte’s classification slightly, distinguishing seven dialects, from West to East: Bizkaian, Gipuzkoan, High Navarrese, Lapurdian, Low Navarrese, Zuberoan (Souletin) and Roncalese, this last one now extinct. Azkue’s classification was adopted by the great linguist Koldo Mitxelena (Luis Michelena in the Spanish form of his name) in much of his work (see, e.g. Michelena 1985) and has become widely used.



On Basque dialects 

FRANCE

Bay of Biscay Bilbao Bizkaian (Western) BIZKAIA

lapurdianLow Navarrese Zuberoan

GIPUZKOA Gipuzkoan (Central)

High Navarrese

Vitoria-Gasteiz Iruña-Pamplona NAVARRE

SPAIN

Figure 1.  Map of the Basque Country and its administrative provinces. The grey line e­ ncloses the area where traditional dialects are still spoken. The discontinuous lines are sharply ­delineated dialectal boundaries

The more recent dialect classification of Zuazo (2008, 2014) groups Lapurdian and Low Navarrese as a single dialect and thus distinguishes only five main living dialects: Western (the Bizkaian of other classifications, which extends over parts of neighboring provinces), Central (or Gipuzkoan), (High) Navarrese, (Low) Navarrese-Lapurdian and Zuberoan. Since robust bundles of isoglosses are rare in the Basque-speaking territory, on the border between the main dialectal areas that he proposes Zuazo recognizes a number of transitional zones. For the most part, important isoglosses within the Basque Country run in a north-south direction, or so has been the case until the strengthening of the linguistic significance of the political border between Spain and France in recent time. Although it is not easy to find large sets of dialectal phenomena with exactly the same geographical distribution, sets of isoglosses separating dialects do bundle up at some locations. One such bundle of isoglosses runs across southwestern Gipuzkoa, neatly separating a western or Bizkaian dialectal area (including all of Bizkaia and also a western part of the province of Gipuzkoa, as well as the still Basque-speaking zone of Alava/Araba) from the Central or Gipuzkoan dialect. The difficulty for drawing the exact border between the western and central dialects on the map is that this dialectal boundary is indeed sharp in the southern,

 José Ignacio Hualde

mountainous region of Gipuzkoa, but the isoglosses that define it fan out and the dialectal boundary dissipates as we approach the coast. Similarly, the boundary between Zuberoan and Low Navarrese is much easier to trace in the foothills of the Pyrenees than further north in the plains. Even if older isoglosses have a different distribution, nowadays, the most important dialectal boundary within Basque is increasingly the one that runs along the Spanish/French border. The reason for this sharp discontinuity is bilingualism: it separates the Basque speech of Basque-Spanish bilingual speakers from that of Basque-French bilinguals. Until recently, before universal bilingualism on the part of Basque speakers became a reality, the linguistic importance of the political border was much slighter and older isoglosses often strand it (as manifested in Bonaparte’s map). Among the main dialectal boundaries in Zuazo’s classification, the least clear one is that between Central (Gipuzkoan) and High Navarrese. High Navarrese speech becomes less and less like Gipuzkoan as we move away form the provincial boundary, but there is little homogeneity within the High Navarrese dialect area. The dialectal picture is summarized in the map included in Figure 1. A caveat is in order. This map can be said to reflect reality only if we are solely concerned with “traditional Basque speakers,” those born in the 1950s or before, who learned Basque from their parents. But these speakers are becoming a minority. Since standard Basque (codified by the Basque Academy in the late 1960’s and 70’s) was introduced in the educational system, the media and the administration starting around 1980, the picture has changed drastically (see Zuazo 2014: 7). Nowadays, there is a very large number of people who have learned standard Basque through the school system, including both people whose home language is Basque and people whose home language is Spanish or French. One now finds Basque speakers both in areas where the transmission of Basque from parents to children had been uninterrupted throughout the centuries and outside of this area, within the administrative boundaries of the Basque County. In areas where traditional Basque had been maintained, the local variety now interacts with standard Basque in the repertoire of younger speakers. Some young speakers are able to move from something close to the local dialect of older speakers to standard Basque, depending on the context and can be said to be bidialectal. Other speakers show less flexibility in their linguistic performance. Depending on factors such as the prior social vitality of the local variety and how different it is from standard Basque, speakers may be able to keep them separate to different degrees. In some areas, where the language was dying before it was reintroduced through the school system, there has been a clear break between the traditional dialect and the speech of the younger generations, who have learned the language in the schools. On the other hand, in towns and villages where intergenerational transmission was never interrupted, young speakers often show much pride in the local



On Basque dialects 

variety, although commonly a linguist can detect important differences between the speech of older and younger generations in the same town. Much inter-speaker variation is also often now found within the same town, so that geographical dialect is perhaps no longer the main feature of variation among Basque speakers. A  20-year old Basque speaker from, say, Gernika, may or may not command a form of speech similar to the language of Basque speakers in their 60s from the same town (see Ensunza Aldamizetxebarria 2016). At present, all geographically-based dialectal classifications of Basque are, thus, dated and frozen in time, given the large number of speakers who speak standard Basque both within and outside of the area where local dialects have been preserved. In this chapter nothing will be said about variation among Basque speakers who have learned the language in childhood, but through the school system or from parents who spoke Basque as a second language, although these types of speakers constitute nowadays a large percentage of all speakers of Basque. 2.  Main dialectal phenomena Many differences between dialects, and the ones that are most salient for speakers as dialectal markers have to do with lexical choices: the use of different words for the same concept (e.g. western bari(a)ku vs. central/eastern ostiral/ortzirale ‘Friday’), or different word variants (e.g. western and central esan vs. eastern erran ‘to say’). There are also a few differences in lexical meaning (e.g. lotsa western/­ central ‘shame’ vs. eastern ‘fear’). Lexical differences in content words, salient as they are for speakers, are, however, hard to systematize. Traditionally, starting with Bonaparte’s work, in tracing dialectal boundaries much attention has been paid to relative small differences in high frequency items, such as those found in the most common forms of the auxiliary verbs. One particularly conspicuous difference is found in the present-tense forms of the transitive auxiliary. Thus, whereas in eastern dialects (and Standard Basque) we find the forms dut ‘I have it’, du ‘s-he has it’, in part of Gipuzkoa these forms are, respectively, det, du and in the western area they are dot, dau. In traditional dialectological work, this feature has been given considerable weight in determining the boundaries of the Gipuzkoan and Bizkain dialects. Here we provide a list of the main dialectal differences in morphosyntax, inflectional morphology and phonology. Starting with syntactic variation, it is useful to distinguish two groups of isoglosses: a first group of eastern features that serve to characterize most or all of the varieties spoken in the territory north of the Pyrenees (in France), including sometimes parts of (High) Navarre, and a second set of western features that separate the Western (or Bizkaian) dialect from the rest. In the next subsection, thus, we will consider, in this order, eastern

 José Ignacio Hualde

morphosyntactic features (2.1), western morphosyntactic features (2.2), variation in nominal inflection (2.3), variation in verbal inflection (2.4), differences in derivational morphology (2.5), and (morpho-)phonological variation, (2.6). 2.1  Eastern morphosyntactic features (East vs. Center and West) The following features separate eastern Basque (Lapurdian-Low Navarrese and Zuberoan, and to some extent also High Navarrese) from the language of central and western areas (Unless variation in the shape of words is relevant, we use standard Basque spelling throughout): 1. The direct object of certain nonfinite nominalized clauses may optionally appear in the genitive case, instead of the absolutive, in the eastern area (Heath 1972; Artiagoitia 2003: 678). Thus, whereas in a finite clause such as the example in (1a) the object alaba ‘daughter’ is in the absolutive case in all Basque dialects, in the example in (1b–c), with a nonfinite verb ikustera ‘(in order) to see’, the object may be marked in the genitive case in eastern varieties: (1) a. alaba ikusten dut daughter- a see-iprf I-have ‘I see the daughter’ b. alaba ikustera noa daugher- a to see I-go ‘I am going to see the daughter’ c. alabaren ikustera noa (eastern) daughter- g to see I-go ‘I am going to see the daughter’



There is textual evidence that in older times the use of the genitive to mark the object in nonfinite clauses had a greater geographical extension (Camino 2011: 104). 2. In part of the eastern area (in the dialects North of the Pyrenees), the auxiliary may be preposed to the main verb in affirmative clauses, to follow a focalized constituent immediately, as in (2a) (see Etxepare and Ortiz de Urbina 2003: 466–467; Zuazo 2008: 235, 2014: 261). In all other dialects, (2b) must be used instead. On the other hand, movement of the auxiliary to the left of the main verb in negative main clauses, as in (2c), is found in all Basque dialects: (2) a. alaba dut ikusi (eastern) ‘I have seen the daughter’ daughter aux see b. alaba ikusi dut ‘I have seen the daughter’ c. alaba ez dut ikusi ‘I have not seen the daughter’.



On Basque dialects 

3. Agreement with a dative argument is optional in eastern dialects (including part of High Navarrese) (see Etxepare and Oyharçabal 2013). Thus, whereas in central areas one can only have (3a) (and its morphological equivalent in Bizkaian), where the auxiliary shows agreement with a third person singular indirect object, in eastern varieties (3b) is also grammatical: (3) a. alabari eman diot (Bizk. alabeari emon dotsat) daughter-d give aux b. alabari eman dut (eastern) ‘I have given it to the daughter’

4. There is a geographically restricted verbal prefix bait- with two main distinct functions, in each of which it has a different geographical extension (Zuazo 2014: 238). It is found in free or appositive relative clauses in a large eastern area, including all dialects north of the Pyrenees and most of High Navarrese. An example is given in (4a). The central and western equivalent is as in (4b), with the relative marker -en- suffixed to the inflected verb: (4) a. Jon, burusoila baita, … (< bait-da) (eastern) ‘Jon, who is bald,…’ b. Jon, burusoila dena, … (< da-en-a) (central and western)



A second use of this verbal prefix is in some causal clauses, as in the example in (5a). This usage is geographically more widespread, including parts of Gipuzkoa. To the west of this isogloss, we find -eta ‘and’ suffixed to an inflected verb or auxiliary in this function, as in (5b):

(5) a. euskaldunak baikara (< bait-gara) (eastern and central) ‘since we are Basque’ b. euskaldunak gara-eta (western)

5. North of the Pyrenees and in parts of High Navarre, there is an interrogative suffix -a added to the inflected verb in yes/no questions. The interrogative counterpart of (6a) is thus (6b). In central (Gipuzkoan) areas, the particle al is (optionally) used before the auxiliary or inflected verb for the same purpose, as in (6c), whereas Bizkaian and other varieties mark interrogativity in this case by purely intonational means: (6) a. ikusi duzu see aux ‘you have seen it’ b. ikusi duzua? (eastern) ‘have you seen it?’ c. ikusi al duzu? (central)

 José Ignacio Hualde

6. In eastern dialects, the bare perfective participle can be used in uninflected relative clauses, as in (7a) whereas in central and western regions a suffix containing relational -ko must be added to the participle, as in (7b): (7) a. nik ikusi jendea (eastern) I-e see people-det ‘the people I saw’ b. nik ikusitako/ikusiriko jendea

7. In eastern dialects, there is a contrast between the bare radical of verbs, which is used in the subjunctive, imperative and potential (e.g. ikus dezagun ‘let’s see it’, har ezazu ‘take it!’) and the perfective participle (e.g. ikus-i dugu ‘we have seen it’, har-tu duzu ‘you have taken it’). In the Gipuzkoan and Bizkaian area the radical has been lost (except in fixed expressions) and replaced with the perfective participle: ikusi dugu ‘we have seen it’, ikusi dezagun/daigun ‘let’s see it’, hartu duzu ‘you have taken it’, hartu ezazu/egizu ‘take it!’. 8. In dialects spoken in contact with Spanish, there is a semantic contrast between a present perfect/recent or hodiernal past and a non-hodiernal perfective past, e.g. ikusi dugu (gaur) ‘we have seen/we saw it (today)’ vs. ikusi genuen (atzo) ‘we saw it (yesterday)’. In dialects in contact with French (although it may have been an independent development), both forms are also employed, but with a less clear-cut difference in meaning, since the present perfect can also be used as a non-hodiernal past. 9. There are some dialectal differences in morphological argument structure. For instance, verbs like bazkaldu ‘to have dinner’ take intransitive morphology in the East and transitive morphology in the Center and West; e.g. ni bazkaldu naiz (where ni ‘I’ is in the absolutive case and naiz is the intransitive auxiliary) vs. nik bazkaldu dut (where nik ‘I’ bears ergative marking and dut is the transitive auxiliary) ‘I have had dinner’ (Oyharçabal 1992; Alberdi 2003; Berro this volume). Many of these differences in argument marking are found in recent borrowings. Speakers who are bilingual in Spanish tend to use intransitive morphology with verbs that in Spanish take reflexive se and transitive morphology with other syntactically intransitive verbs; e.g. intransitive graduatuko da ‘s-he will graduate’ (Sp. se graduará) vs. transitive emigratuko du ‘s-he will emigrate’ (Sp. emigrará). Basque-French bilinguals usually make a more systematic use of intransitive morphology with syntactically intransitive verbs. 2.2  Western morphosyntactic features (West vs. Center and East) 1. In eastern and central dialects there is a progressive construction with the auxiliary ari. This construction is not found in Bizkaia and neighboring areas of Araba and Gipuzkoa, where it is replaced by constructions with egon ‘to



2.

3.

4.

5.

On Basque dialects 

be/stay’ and ibili ‘to go about, walk’; e.g. eastern and central kantatzen ari da ‘s-he is singing’ vs. western kantatzen dago ~ dabil (cf. Sp. está/anda cantando). Western varieties have a couple of procedures to signal the focalization of the event expressed by the verb. In the western dialectal area, with synthetic verbs, the perfective participle is added immediately before the conjugated form, as in jakin dakit lit. ‘know, I know it’, ibili dabil lit. ‘walk about, s-he walks about’, eroan daroaz lit. ‘carry, s-he carries them’. A second strategy, with a wider geographical extension, is one found with analytical verbs. In this construction, the “dummy verb” egin ‘do’ is inserted between main verb and auxiliary, e.g. ekarri egingo dugu ‘what we will do is bring it’ (see Etxepare and Ortiz de Urbina 2003: 470–471; Elordieta and Haddican this volume). In all Basque dialects the numeral bat ‘one’ is postposed to the noun (and adjectives), sagar (gorri) bat ‘a/one (red) apple’, whereas numbers above three are preposed, hiru sagar (gorri) ‘three (red) apples’. The number two is postposed in Bizkaian and (optionally) in some Gipuzkoan areas, sagar bi ‘two apples’, and preposed elsewhere, bi sagar. Although postnominal bi is now, thus, a western or Bizkaian feature, examples are also found in authors from the eastern area for older periods. Demonstratives appear in noun-phrase-final position in all dialects, except in the western dialect, where they are preposed or both preposed and postposed; e.g. common gizon hori ‘that man’, with the demonstrative hori ‘that’ postposed to the noun, vs. western hori gizona ~ hori gizonori. In the West, some syntactically intransitive verbs, including urten ~ irten ‘to leave’, which elsewhere take intransitive morphology, take bivalent ergativeabsolutive morphology instead, as in (8) (See 2.1.9 above):

(8) a. (Nik) etxetik urten dot (western) I-e house.from leave aux-tr b. (Ni) etxetik irten/atera naiz (common) I (a) house. from leave aux-intr ‘I have left from my house’

2.3  Dialectal variation in nominal inflectional morphology There is some dialectal variation in nominal case suffixes. Only some of the main differences will be mentioned here. 1. Standard Basque has adopted a distinction between the ending -ak for the absolutive plural and -ek for the ergative plural. Thus, we find absolutive plural sagarrak ‘the apples’ in (9), but ergative plural sagarrek ‘the apples’ in (10). In making this morphological distinction between absolutive and ergative plural, Standard Basque has followed eastern usage. In central and western

 José Ignacio Hualde

dialects, on the other hand, this morphological distinction is not made and in our examples the form would be sagarrak in both cases. (9) sagarrak onak dira apple- a.pl good-pl are ‘The apples are good’ (10) sagarrek kolore ona dute apple-e.pl color good-sg have ‘The apples have good color’

2. In the dative plural, on the other hand, Standard Basque has favored the central/western ending -ei (e.g. sagarrei ‘to the apples’) over the eastern ending -er (e.g. sagarrer ‘to the apples’). 3. In the comitative, Standard Basque uses the suffix -kin: gizon-are-kin ‘with the man’, gizon-e-kin ‘with the men’ added to genitive ending (minus the -n). The easternmost dialect, Zuberoan, has the variants -ki and -kila. In the western area, instead, the suffix -gaz is used: gizonagaz ‘with the man’, gizonakaz (< gizonak-gaz) ‘with the men’. The isogloss for singular -gaz runs further East than that for plural -kaz, which is found only in the westernmost part of Bizkaia; that is, there are local western varieties that oppose singular gizonagaz to plural gizonakin. 4. Singular and plural forms are morphologically distinguished for all nominal cases in Standard Basque, but this is not always so in local dialects. In some Gipuzkoan, Bizkaian and High Navarrese areas, as well as in Zuberoan, singular and plural are accentually distinguished. For instance, in some Gipuzkoan varieties we find gizónai ‘to the man’ vs. gízonai ‘to the men’. In parts of Bizkaia, there is a contrast between unaccented and accented words. For most nouns and adjectives, singular and indefinite forms are unnaccented in most morphological cases and plural forms are accented instead, e.g. lagunen alabiá ‘the friend’s-sg daughter’ vs lagúnen alabiá ‘the friends’-pl daughter’. These accentual contrasts are being lost in most areas. 5. Basque in contact with Spanish makes a contrast between a benefactive form, e.g. lagunarentzat ~ lagunarendako ‘for the friend’ (Sp. para el amigo) and a prolative form, e.g. laguntzat ~ lagunetako ‘as friend’(Sp. por amigo), whereas bilingual Basque/French speakers do not make such a distinction. 6. The ablative suffix -tik ‘from; through’ (e.g. menditik ‘from the mountain’, leihotik ‘through the window’) may derive from an sequence -ti-(r)ik (where -ti may originally have had a wider locational meaning and -rik is the partitive suffix). Dialectally forms like menditi are found in a number of geographically discontinuous areas, and menditikan is found (optionally) in an area of Gipuzkoa and bordering towns in Navarre. In the ablative plural, -etatik competes with eastern -etarik.



On Basque dialects 

7. To form the locative cases of animates western and central varieties use the suffix -gan, added to genitive forms (e.g. seme-aren-gan ‘in the son’, semearengana ‘to the son’), whereas in the East the postposition baita is used instead (semearen baitan ‘in the son’, semearen baitara ‘to the son’). For reasons of space, I leave other, perhaps less significant, dialectal differences in nominal morphology unmentioned here (see Hualde 2003: 179–186; Zuazo 2008, 2014). 2.4  Dialectal variation in verbal morphology Verbal morphology is notoriously variable among Basque dialects. For this reason, scholars like L.L. Bonaparte (1813–1891) and P. Yrizar (1910–2004) devoted many years in an effort to document the diverse forms of the auxiliary verbs in local varieties. Yrizar’s life work, Morfología del verbo auxiliary vasco, which was published in a series of volumes over the years, is now conveniently available from the webpage of the Basque Academy, Euskaltzaindia. A recent theoretical treatment of some aspects of this variation can be found in Arregi and Nevins (2012). See also the chapters by Rezac and Fernández and Rezac in this volume. Here we will touch upon only the main facts of variation in this domain. 1. Perhaps the most conspicuous domain of variation in verbal morphology has to do with the shape of trivalent auxiliaries (signaling agreement with subject, direct object and indirect object). In Standard Basque, central variants, formed on the root -i- have been selected. In the western area conjugated forms of eutsi ‘to hold’ (* dabe ‘they have it’, duzue ~ duzute ‘you-pl have V’ed it’, dakite ~ dakie ‘they know it’, zarete ~ Bizk. zaree > zarie ‘you-pl are’. The strongest preference for -te across verbal paradigms is found in the central border region between Gipuzkoa and Navarre, whereas -e prevails in both peripheral areas, Bizkaia and Zuberoa (Zuazo 2008: 191–194). Most likely both forms derive from earlier *de. In Standard Basque, imperative, subjunctive and potential transitive auxiliary forms belong to a verb whose perfective participle is reconstructed as *ezan;



On Basque dialects 

e.g. dezagun ‘let’s V it’, dezaket ‘I can V it’. This standard choice reflects the most widespread usage. In the western dialectal area, however, forms of egin ‘to do’ are employed: daigun ( /x/ is a regular sound change, it is found only in borrowings and a few native words,), and none of those spoken in French territory do. That is, the political border between Spain and France coincides perfectly with the presence of the phoneme /x/. 4. The distribution of palatalization processes has a much more complicated geography. There are two distinct palatalization phenomena in Basque (see



On Basque dialects 

Oñederra 1990). One is, “affective” palatalization, i.e., the replacement of dental and alveolar segments with (pre)palatals in diminutives and child-directed speech, e.g. bero ‘hot’, bello [beʎo] ‘a little hot’. Nowadays this rule has greatest strength in parts of the High Navarrese and Lapurdian-Low Navarrese area. The other phenomenon is the palatalization of dentals and alveolars after /i/ or a palatal glide. We find substantial variation both in the set of segments that undergo the process and in the specific context of the rule (see Hualde 1991). The most commonly palatalizing segments are /n/ and /l/, e.g. mina [miɲa] ‘the pain’, mutila [mutiʎa] ‘the boy’, but /t/ also undergoes palatalization in some areas, e.g. ditut [dicut] ([c] is the IPA symbol for a voiceless palatal stop) ‘I have them’ and sibilants palatalize in most of the western dialectal area, e.g. gizon [giʃon] ‘man’. 5. Differences in vowel inventory are few, almost all dialects having a simple 5-vowel inventory /i e a o u/. In Zuberoan and some Lapurdian-Low Navarrese areas, there is a sixth vowel phoneme /y/, orthographical represented as 〈ü〉. Historically all varieties may have had nasalized vowels (as a consequence of the historical weakening of intervocalic /n/), but these have been preserved only in Zuberoan. 6. We find a great variation of solutions regarding the pronunciation of sequences of vowels arising in morphological concatenation. For instance, the sequence /i+a/, as in mendia ‘the mountain’, may be pronounced with an epenthetical (pre)palatal consonant in many western and central varieties [mendija] ~ [mendiʒa] ~ [mendiʃa] and as a diphthong in a large High Navarrese area, [mendja]. A phonological rule with a very irregular geographical distribution (covering almost all of the province of Bizkaia and a non-contiguous Gipuzkoan and High Navarrese area, see Zuazo 2014: 267 for a map) is the raising of /a/ to /e/ after a high vowel, either immediately preceding, as in mendia [mendiʒe], [mendje], etc., or across consonants, as in laguna [laɣune] ‘the friend’. When the singular determiner -a is added to nouns or adjectives also ending in -a we find a dissimilation rule, e.g. alaba+a  alabea ‘the daughter’ in a western area that includes all of Bizkaia and also extends through southern Gipuzkoa up to a small part of Navarre (see Zuazo 2014: 212). Unlike the other phenomena of vowel interaction mentioned above, this one is restricted to the singular paradigm, and does not affect plural forms, e.g. alaba+ak  alabak ‘the daughters’. At the other end of the Basque territory, in Zuberoan (as well as in the extinct Roncalese dialect), contraction is reflected in accentuation: alhába ‘daughter’ vs. alhabá ‘the daughter’. In the rest of the Basque territory, contraction results in homophony between uninflected form and absolutive singular for words ending in -a, alaba ‘(the) daughter’.

 José Ignacio Hualde

7. There are also substantial differences in accentual systems across Basque varieties (see Hualde 1999), including some typologically very interesting pitch-accent systems both in the north of Bizkaia and in the west of Navarre. Nowadays, however, we find a strong tendency towards the loss of accentual distinctions in the speech of the younger generations of most areas. A salient stress-related feature, widespread in the High Navarrese area is the deletion of word-initial and other unstressed vowels, e.g. eman > man ‘give’, zaree > zrei (Roncalese) ‘you-pl are’. 3.  Origin and development of Basque dialects From the point of view of its dialectal development, we may distinguish two main periods in the documented history of Basque: a first period from the first texts until approximately 1980, and a second period from that date until now. In the first of these two periods, we find an increasing tendency towards dialectal diversification, with local varieties becoming more differentiated with the passage of time. In the last few decades, starting from a date that we can set around the year 1980, we find a reversal of this phenomenon, with a great deal of dialectal convergence and much erasure of traditional dialectal boundaries, as was mentioned before. This reversal in tendencies has been brought about by a number of important changes in Basque society, including the use of Standard Basque in the school system and the media, an astonishing increase in the number of speakers who have learned Basque as a second language (who now constitute a large percentage of all Basque speakers), the virtual disappearance of monolingual Basque speakers, and improved communication together with the modern need to travel beyond the local area for all kinds of purposes. Given the fact that the first texts in Basque that we have already show dialectal differences, we can also speak of a prehistoric or undocumented period or dialectal diversification from an assumed common form of the language to the first historical documentation. The first phrases in Basque that have arrived to us are included in the Glossae Æmilianenses (in Sp. glosas emilianenses), a set of marginal annotations or glosses made in a document from La Rioja containing a series of sermons in Latin, including two glosses in Basque: jzioqui dugu, partially interpretable as ‘(we) have V’ed it’ and guec ajutu ez dugu ‘we have not V’ed it’. Arguably these two glosses already show signs of dialectal diversification in the form of the auxiliary dugu ‘we have it’. As already mentioned, in modern dialects, together with this form, we have the variants dogu (in western varieties) and degu (in Gipuzkoan). All three forms are reconstructable to an earlier *daugu (and an even earlier *da-du-gu, see



On Basque dialects 

Gómez and Sainz 1995). It would seem that we can conclude that dialectal diversification in Basque was already underway in the 11th century. Nevertheless, from the attestation of the form dugu in this text we cannot be certain that a clear-cut geographical isogloss already existed for this form at this time, given the fact that, as Zuazo (2013: 51) points out, in works of 15th century writers from Alava we find a plurality of forms, e.g. dot ~ dut ~ det ~ deut ‘I have it’ and polymorphism even within the speech of a single town is still found in some local dialects of Navarra bordering Alava. It thus appears that different evolutions of the diphthong /au/ may have coexisted for centuries in at least some areas, before this variation was solved by the different dialectal choices we find nowadays. That is, perhaps dugu was not the only variant of this auxiliary form that the author of the glosses employed in speech. Some authors have assumed that dialectal diversity in Basque is ancient, going so far as to seek a connection between present-day dialects and the different tribes or nations that in Roman times inhabited the territory of the modern Basque Country (Caro Baroja 1943). However, Michelena (1981) offered convincing arguments against this view. All Basque dialects are too similar in structure and vocabulary for the hypothesis of a pre-Roman origin to be sensible. Basque syntax and morphology is essentially the same across geographical varieties, including most of the details of its complex verbal morphology. The sound adaptations in old borrowings from Latin are also the same everywhere (e.g. preservation of intervocalic voiceless stops, as in Lat. lacu- > laku ‘lake’, but initial voicing, as in Lat. causa > gauza ‘thing’). Some late developments that appear to be due to contact with Romance are also common to all Basque dialects. These would include the grammaticalization of demonstratives as articles (cf. gizon hari ‘to that man’ > gizonari ‘to the man’, see Manterola 2015) and the adoption of the old second person plural pronoun and verb forms as second person singular polite forms. Given all of this, we must conclude, with Michelena and modern scholarship (Zuazo 2008, 2014, Lakarra 2011), that geographical differentiation from a common form of Basque started at some point in the Middle Ages, not in Roman times or before. An important difficulty for dating Ancient Common Basque2 and determining its features is that, as Basque has continued being spoken in geographically contiguous territories, newer changes may have spread throughout the Basquespeaking area across older dialectal boundaries, so that features that are found in all dialects are not necessarily to be traced back to Ancient Common Basque.

.  This has become the usual term in academic publications to refer to the hypothesized stage before dialectal differentiation (in Basque, euskara batu zaharra). See Lakarra (2011) for discussion.

 José Ignacio Hualde

As an example, let us consider the evolution of the second person forms, which we have already mentioned. Present-day Basque has three such forms, singular familiar, singular polite and plural, as in hi haiz ‘you-sg are (familiar)’, zu zara ‘you-sg are (polite)’ and zuek zarete ‘you-pl are’. That the second person singular polite form zu was originally plural can be deduced from the fact that when it is a direct object it takes plural agreement in the verb; e.g. na-u ‘s-he has me’ d-u ‘s-he has it’ vs. za-it-u ‘s-he has you-sg’ with the same pluralizer -it- as we find in ga-it-u ‘s/has us’, d-it-u ‘s-he has them’. A first step was the extension in the use of these originally plural forms as polite forms to address a single individual, as happened in Romance and we still find in present-day French vous. We have clear documentation from several dialectal areas that until relatively recent times forms like zara ‘you are’, zaitugu ‘we have you’ were ambiguous between second person plural and second person singular polite reference. Nevertheless, in present day Basque, such forms are unambiguously singular in all dialects. New forms like zare(t)e ‘you-pl are’, zaituztegu ‘we have you-pl’ (with two pluralizers, -it- and -zte) were created very recently, and spread throughout the entire Basque Country, so that without historical documentation we would be led to postulate them for Ancient Common Basque. It is sensible to assume that Ancient Common Basque was spoken in a geographically compact area, as Michelena does (although there is abundant epigraphical evidence that the area where an older stage of Basque and related languages were spoken was much larger in Roman times, on both sides of the Pyrenees). Zuazo (2010, 2014) argues that the original koiné may have been the speech of the area around the city of Pamplona, which became politically dominant in the early Middle Ages. A linguistic reason for postulating an eastern origin for Basque is that the Romance variety with which Basque shares the greatest similarity is Pyrenean Gascon and, to a somewhat smaller extent, Pyrenean Aragonese dialects (see Allières 1992). It thus makes sense to locate the geographical origin of common Basque in the Pyrenees. In its subsequent westward expansion, Basque may have replaced both related dialects (Bascoid) and unrelated languages (e.g. Celtic and even Latin/Romance). 4.  The written or “literary” dialects and regional norms In past centuries, the written production of authors stemming from different Basque-speaking regions whose intended readership was geographically constrained served to establish local written models. These have come to be known as



On Basque dialects 

the written or literary dialects of Basque. The written dialects with the strongest tradition are, South of the Pyrenees, literary Bizkaian and literary Gipuzkoan, and, North of the Pyrenees, literary Zuberoan and literary Low Navarrese-Lapurdian (as codified in Laffite’s 1944 grammar of this dialect). An older written model with great importance in the Basque literary tradition (as well as in the development of Standard Basque) is the Classical Lapurdian of the 17th century. Some of these written varieties still have some use, especially literary Bizkaian and Zuberoan, but they have been replaced in most of their functions by Standard Basque, which, as mentioned, was developed by the Basque Academy in the late 1960’s and 70’s, primarily based on a fusion between the Gipuzkoan and the Lapurdian literary models. Unlike what is the case with oral dialects, which may gradually morph into each other, literary dialects are more distinct. It is possible to define the differences between any two literary dialects; e.g., how literary Bizkaian differs from literary Gipuzkoan or literary Zuberoan (or, in a diachronic comparison of literary dialects, how the literary Low Navarrese-Lapurdian of the first decades of the 20th century differs from Classical Lapurdian). At the end of the 19th century and beginning of the 20th we find attempts to make literary dialects as different from each other as possible, especially regarding Bizkaian, leading to the development of “hyper-Bizkaian” (see Laka 1987). This was based in part on the incorrect notion of “pure” and “mixed” dialects, already present in the work of L. L. Bonaparte and further developed by other writers, and in part on political ideology. The guiding idea was that if, for instance, in some area within the province of Bizkaia they used a form that was not found anywhere else and in the rest of the province the form that was used was common to Gipuzkoan, the former must be considered pure Bizkaian and the latter to be due to contamination from the Gipuzkoan dialect. To give an example, the verbal form ditugu ‘we have them’ is found in all Basque dialects, including parts of Bizkaia. In some areas of Bizkaia, however, an analogical form doguz ‘we have them’ has arisen by adding the pluralizer -z to dogu ‘we have it’. The logic of making Bizkaian pure and as different as possible from other Basque dialects dictates that doguz should be chosen for written Bizkaian. 5.  Summary To sum up, this chapter has offered an overview of the main points of gegraphical variation in syntax, morphology and phonology in present-day Basque. Instead of identifying a specific set of dialects and listing the distinctive features of each of

 José Ignacio Hualde

them, we have identified specific features and their geographical distribution, as this gives a more realistic view. We have also briefly offered a diachronic perspective on dialectal diversification in Basque.

References Alberdi, Xabier. 2003. “The transivitity of borrowed verbs in Basque: An outline. In Inquiries into the Lexicon-Syntax Relations in Basque [Supplements of ASJU 43], Bernard Oyharçabal (ed.), 23–46. Bilbo: UPV/EHU. Allières, Jacques. 1992. “Gascón y euskera: afinidades e interrelaciones lingüísticas”. ASJU 26 (3): 801–812. Arregi, Karlos and Nevins, Andrew. 2012. Morphotactics: Basque Auxiliaries and the Structure of Spellout. Springer. Artiagoitia, Xabier. 2003. “Complementation (noun clauses)”. In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 634–711. Berlin: Mouton de Gruyter. Azkue, Resurrección M. 1905–06. Diccionario vasco-español-francés. Bilbo. Repr., 1969, Bilbo: La Gran Enciclopedia Vasca. Berro, Ane. This volumen. “On the relation between ergativity, stativity and the verbal configuration of Basque”. Bonaparte, Louis-Lucien. 1869. Carte des sept provinces basques, montrant la délimitation actuelle de l’euscara. London. Camino, Iñaki. 2011. “Ekialdeko euskararen iraganaz”. In Euskal dialektologia: lehena eta oraina, [Supplements of ASJU 69], Irantzu Epelde (ed.), 87–153. Bilbo: UPV/EHU. Caro Baroja, Julio. 1943. Los pueblos del norte de la Península Ibérica. Madrid: CSIC. Elordieta, Arantzazu and Haddican, Bill. This volume. “Strategies of verb and verb phrase focus across Basque dialects”. Ensunza Aldamizetxebarria, Ariane. 2016. “The sociolinguistic variation of palatalization: The case of Gernika-Lumo”. Dialectologia 16: 71–91. Etxepare, Ricardo and Oyharçabal, Bernard. 2003. “Focalization”. In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 459–516. Berlin: Mouton de Gruyter. Etxepare, Ricardo and Ortiz de Urbina, Jon. 2013. “Datives and adpositions in Northeastern Basque”. In Variation in Datives: a Micro-comparative perspective Beatriz Fernández and Etxepare, Ricardo (eds.), 50–95. New York: Oxford University Press. Fernández, Beatriz and Rezac, Milan. This volume. “Differential object marking in Basque varieties”. Gómez, Ricardo and Sainz, Koldo. 1995. “On the origin of the finite forms of the Basque verb”. In Towards a History of the Basque Language, José Ignacio Hualde, Joseba A. Lakarra and Robert L. Trask (eds.), 235–274. Amsterdam: John Benjamins.  doi: 10.1075/cilt.131.10gom Heath, Jeffrey. 1972. “Genitivization in Northern Basque complement clauses”. ASJU 6: 46–66. Hualde, José Ignacio. 1991. Basque phonology. London: Routledge. Hualde, José Ignacio. 1999. “Basque accentuation”. In Word Prosodic Systems in the Languages of Europe, Harry van der Hulst, (ed.), 947–993. Berlin: Mouton de Gruyter. Hualde, José Ignacio. 2003. “Case and number inflection of noun phrases”. In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina Hualde (eds.),171–186. Berlin: ­Mouton de Gruyter.  doi: 10.1515/9783110895285



On Basque dialects 

Hualde, José Ignacio. 2010. “Neutralización de sibilantes en lengua vasca y seseo en castellano”. Oihenart 25: 89–116. Hualde, José Ignacio, Elordieta, Gorka and Elordieta, Arantzazu. 1994. The Basque Dialect of Lekeitio [Supplements of ASJU 34]. Bilbo and Donostia: UPV/EHU and Gipuzkoako Foru Aldundia. Jauregi, Oroitz and Epelde, Irantzu. 2013. “Hasperenaren galera iparraldeko euskaran”. In Koldo Mitxelenaren Katedraren III. Biltzarra/III Congreso de la Cátedra Luis Michelena/3rd Conference of the Luis Michelena Chair, Ricardo Gómez, Joaquín Gorrochategui, Joseba A. Lakarra and Céline Mounole (eds.), 245–262. Vitoria-Gasteiz: UPV/EHU. Lafitte, Pierre. 1944. Grammaire basque (navarro-labourdin littéraire). Bayonne: Librairie “Le Livre”. Laka, Itziar. 1987. “Sabino Arana Goiri eta Hiperbizkaiera (Hiperbizkaieraren historiaz. III)”. ASJU 21 (1): 13–40. Lakarra, Joseba. 2011. “Gogoetak euskal dialektologia diakronikoaz: Euskara Batu Zaharra berreraiki beharraz eta haren banaketaren ikerketaz”. In Euskal dialektologia: lehena eta oraina [Supplements of ASJU, 69], Irantzu Epelde (ed.),155–232. Bilbo: UPV/EHU. Leizarraga [Leiçarrague], Joanes. 1571. Jesus Christ gure jaunaren testamentu berria. La Rochelle: Pierre Hautin. Michelena, Luis. 1985. Fonética histórica vasca, 3rd ed (1st ed., 1961) [Supplements of ASJU 4]. Donostia: Donostiako Foru Aldundia. Manterola, Julen. 2015. Euskararen morfologia historikorako: artikuluak eta erakusleak. Ph.D. dissertation, UPV/EHU. Michelena, Luis. 1981. “Lengua común y dialectos vascos”. ASJU 15: 289–354. Oñederra, Miren Lourdes. 1990. Euskal palatalizazioa. Bilbo: UPV/EHU. Oyharçabal, Bernard. 1992. “Structural case and inherent case marking: Ergaccusativity in Basque”. In Syntactic Theory and Basque Syntax [Supplements of ASJU 27] Joseba A. Lakarra and Jon Ortiz de Urbian (eds.), 309–342. Bilbo: UPV/EHU. Rezac, Milan. This volume. “Gaps and stopgaps in Basque finite verb agreement”. Rezac, Milan and Fernández, Beatriz. 2013. “Dative Displacement in Basque”. In Variation in Datives: a micro-comparative perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 256–282. New York: Oxford University Press. Trask, R. L. 1997. The history of Basque. London: Routledge. Yrizar, Pedro de. 2014. Morfología del verbo auxiliar vasco. On line: http://www.euskaltzaindia. net/index.php?option=com_contentandview=articleandid=772andItemid=467andlang=eu Zuazo, Koldo. 2008. Euskalkiak: Euskararen dialektoak. Donostia: Elkar. Zuazo, Koldo. 2010. El euskera y sus dialectos. Irun, Gipuzkoa: Alberdania. Zuazo, Koldo. 2013. El euskera de Álava. Vitoria-Gasteiz: Arabako Foru Aldundia. Zuazo, Koldo. 2014. Euskalkiak. Donostia: Elkar.

part 2

Case and Agreement

On the relation between ergativity, stativity and the verbal configuration of Basque Ane Berro University of the Basque Country (UPV/EHU) & Structures Formelles du Langage (UMR 7023, CNRS/Paris 8)

This paper shows that, in Basque, ergative case does not always correlate with agentivity and with the verbal category. Ergative subjects can be holders of verbal states, and, in some cases, subjects of non-verbal stative predicates. I argue that in non-verbal stative predicates, termed bare analytic predicates, the ergative argument is introduced by a head of central coincidence. It is proposed that this head is responsible for the spell out of a have auxiliary (à la Kayne 1993) and is also related to the ergative marking of the subject. In the last sections, this analysis is generalized to verbal states and analytic eventive verbs headed by the -tu suffix, claiming that the ergative subject is always introduced by a head of central coincidence. Keywords:  ergativity; stative predicates; nominalization; auxiliary selction; central coincidence

1.  Introduction1 Basque is an ergative language in the sense that the subject of a transitive verb is assigned ergative case (1), whereas the subject of an unaccusative verb is assigned

.  I am grateful to Beatriz Fernández, Ricardo Etxepare, Myriam Uribe-Etxebarria and Jairo Nunes for their valuable comments and suggestions. I also want to thank the interviewed speakers and consulted colleagues and friends (Ane Odria, Mikel Ayerbe and Iker Legarra) for sharing with me their judgments and intuitions. All errors are, of course, mine. The research leading to these results has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no. 613465. Moreover, this study has also been supported by the Basque Government (the post-doctoral grant POS_2015_1_0086 and the project IT665-13) and by the Spanish Ministry of Economy and Competitiveness (FFI2014-51878-P).

doi 10.1075/lfab.13.03ber © 2016 John Benjamins Publishing Company

 Ane Berro

absolutive zero case (2). The absolutive, is also assigned to the direct object of the transitive verb (1). (1) Jon-ek sagarra-ø ja-n du2 John-erg apple-abs eat-tup have:3sgerg ‘John has eaten an apple’ (2) Jon-ø jaus-i da John-abs fall-tup be:3sgabs ‘John has fallen’

The subject of unergative verbs, on the other hand, is usually3 assigned ergative, like the subject of a transitive verb. (3) Jon-ek dantza-tu du John-erg dance-tup have:3sgerg ‘John has danced’

Some works have accounted for this divergence of unergative subjects, claiming that unergative verbs are transitive verbs in nature (Bobaljik 1993; Laka 1993; Fernández 1997), following the original proposal of Hale and Keyser (1993). Other studies, on the other hand, have suggested that ergative case is inherent (or lexical) in Basque, assigned to the external argument (Oyharçabal 1992; Holmer 1999; Laka 2006a). In this paper, I am going to present a new perspective of the argument marked with ergative case in Basque, analyzing and discussing other types of verbs, and focusing on the analytic verbal configuration of Basque. Basque verbs can appear in the synthetic (4) and in the analytic configuration (1)–(3). In contemporary Basque, the synthetic configuration is only used with about fifteen verbs (e.g. joan ‘go’, etorri ‘come’, ekarri ‘bring’, eraman ‘carry, bring or wear’ and jakin ‘know’), and its interpretation is aspectually very specific: depending on the verb, it has an on-going or unbounded aspect interpretation, as can be seen in (4). (4) Neske-k sagar-poltsa bat-ø da-kar-te girl:pl-erg apple-bag a-abs 3abs-bring-3plerg ‘The girls are carrying a bag of apples’

.  Abbreviations: erg = ergative case, abs = absolutive case, tup = the -tu morpheme representing the participle, tui= the -tu morpheme representing the infinitive, iprf = imperfect suffix, irr = irrealis suffix, gen = genitive, all = allative adposition, part = partitive, nmlz = nominalizer, instr = instrumental adposition, ine = inessive adposition, abl = ablative adposition, appr = approximative adposition, sup = superlative. .  In north-eastern varieties, the subject of some unergative verbs are not marked ergative, but absolutive (see Oyharçabal 1992; Etxepare 2003; Aldai 2006, 2009, 2012; Berro 2010, 2012). I am not going to deal with this type of variation in this paper.



On the relation between ergativity, stativity and the verbal configuration of Basque 

The analytic form, on the contrary, is available for all verbs. It is the only productive configuration and it can be used in perfect, perfective, imperfective and ­prospective aspectual contexts. The analytic configuration consists of two elements: an auxiliary (which can be be or have depending on the verb) and a lexical verb. In the perfect, the lexical verb is headed by the morpheme -tu/-i/-n (as in dantza-tu (3)). For ease of exposition, I will refer to this morpheme as the -tu morpheme. When followed by the auxiliary verb -tu headed predicates act as participles (5). In other contexts, for instance when followed by a modal verb like nahi izan ‘want’, they behave as infinitives (6). Furthermore, -tu is also used for the citation form of the verbs and, in western varieties of Basque, also as the imperative. (5) Neskak-ø etorr-i dira girl:pl-abs come-tup be:3plabs ‘The girls have come’ (6) Neske-k etorr-i nahi dute girl:pl-erg come-tui want have:3plerg ‘The girls want to come’

In the imperfective (7), instead of the morpheme -tu, -t(z)en is attached (-tzen for simplicity). In future tenses (8), the irrealis -ko (Laka 1996) is added to -tu and the auxiliary selected is in the present tense. (7) Neskak-ø astelehen-etan etor-tzen dira girl:pl-abs monday-pl:ine come-iprf be:3plabs ‘The girls come on Monday’ (8) Neskak-ø astelehen-ean etorr-i-ko dira girl:pl-abs monday-ine come-tu-irr be:3plabs ‘The girls will come on Monday’

In the following pages, I am going to show that ergative case in Basque does not go hand in hand with agentivity or the verbal category. I am going to claim that the external argument is introduced by a head of central coincidence, and that this head is the source of the ergative case. More specifically, I propose that in analytic verbal forms, this head is an adposition, which selects for nominal predicates (bare predicates or nominalized instances of the verbs). In this study, I am not going to consider the analysis of the synthetic configuration. The paper starts analyzing some diastratic variation regarding the use of verbal states which take ergative subjects. Then, bare analytic predicates are presented, a type of stative predicate which is non-verbal and takes ergative and absolutive arguments. It is claimed that the external argument in these predicates is introduced externally, by an adposition of central coincidence. In the fourth section, the analysis is generalized to the verbal states attested diastratically, and also to the

 Ane Berro

eventive analytic verbs headed by -tu. Finally, in the fifth section, the main conclusions are presented. 2.  Diastratic variation: Verbal states with ergative subjects In this section, I am going to present the new use of some verbs such as usaindu ‘smell’ where the predicate conveys a property of the subject, rather than an event that is performed by it. These examples represent new uses of the language, presumably influenced by the Romance se-less configuration of counterpart verbs.4 They are more easily found in spoken and colloquial language, and the majority are not accepted in the standard. As a consequence, most of them are not found in written corpuses or dictionaries, but some speakers, especially the young, consider them totally natural. Their acceptance varies among speakers, and it also depends on the particular verb. The verbs that can occur in this structure are pisatu ‘weigh’, neurtu ‘be long’, erre ‘burn, be warm’, labaindu/irristatu/lerratu ‘be slippery’, usaindu ‘(emit) smell’, pikatu ‘be spicy’, kubritu ‘be deep’ and itxi/zarratu ‘close’.5,6 (9) Asko pisa-tzen du  [ez dakit eraman ahal izango duzun] a.lot weigh-iprf have:3sgerg ‘It is very heavy, I do not know if you are going to be able to carry it’  (A. Cano, Belarraren Ahoa: 63) (10) 27 kilometro neur-tzen ditu alde zabal-ean […] 27kilometer measure-iprf have:3sgerg:3plabs side wide-ine ‘It is 27 kilometers long in its wide side’ (Berria, 2004-12-16) .  For example, in Spanish, the verb resbalar ‘slip’ can mean ‘be slippery’ but only in a configuration without the se clitic (cf. ia–b). (i) a. El suelo resbala the floor slips b. *El suelo se resbala    the floor SE slips    ‘The floor is slippery’ The influence of the se clitic has been reported in Alberdi (2003) when dealing with loan verbs: verbs or meanings of verbs which originally take the se clitic in Spanish or French are used in Basque with an absolutive subject and be auxiliary. In contrast, those which do not take it, are used in Basque with an ergative subject and have auxiliary. .  Among these verbs, the ones that are more easily accepted are pisatu ‘weigh’, neurtu ‘be long’, erre ‘be warm’, pikatu ‘be spicy’ and itxi/zarratu ‘close’. .  The examples with no reference represent data obtained from speakers: The ones which have a reference source have been taken from Egungo Euskararen Hiztegia (Dictionary of Contemporary Basque) (Sarasola et al. 2010).



On the relation between ergativity, stativity and the verbal configuration of Basque 

(11) Platera-k erre-tzen du dish-erg burn-iprf have:3sgerg ‘The dish burns’ (12) Lurra-k laban-tzen du floor-erg slip-iprf have:3sgerg ‘The floor is slippery’ (13) Piper hauek asko pika-tzen dute pepper these.erg a.lot burn-iprf have:3plerg ‘These peppers are very spicy’ (14) Lantegi horre-tako keak oso gaizki usain-tzen du factory that-gen smoke-erg very wrong smell-iprf have:3sgerg ‘The smoke of that factory smells very bad’ (15) Kremailera hon-ek ez du ondo ix-ten zip this-erg no have:3sgerg well close-iprf ‘This zip doesn’t close properly’

As can be seen in the examples, in these configurations the auxiliary used is have and the only argument of the verb bears ergative case, which is cross-referenced on the auxiliary with ergative agreement. I claim that in this particular use of the verbs, these predicates are non-eventive: they describe a property of the subject entity and, thus, the ergative subjects are holders, not agents. This claim is supported by several facts: (i) the semantic opposition obtained if we compare the readings in the configurations above with the readings of a parallel alignment where the argument is marked absolutive and the auxiliary selected is be; (ii) the general lose of their stative interpretation when they occur in -tu analytic configurations; (iii) the unbounded or indefinite reading with the imperfect aspect suffix -tzen; (iv) the semantic p ­ arallelism with predicative structures consisting of an adjective/noun and a copula; (v) the incompatibility with the process adverb astiro-astiro/poliki-poliki ‘slowly’; (vi) the degree interpretation of the modifier apur bat/pixka bat ‘a little’; and finally, (vii) the impossibility to occur as the infinitival complement of perception verbs. Let us begin with the semantic opposition of the readings of (9)–(15) and the readings in (16)–(18). Three of the verbs presented above (neurtu, erre and labaindu/irristatu/lerratu), when consisting of a single argument, usually occur with an absolutive subject and izan be auxiliary, as in (16), (17) and (18).7 .  Note that there is not a one-to-one mapping between the absolutive case of the subject and the -tu suffix. As a matter of fact, absolutive subjects can perfectly co-occur with the -tzen suffix, e.g. with the verbs shown in examples (16)–(18) in a frequentative meaning: (ii) Jon-ø beti laban-tzen da eskilara hauetan Jon-abs always slip-tzen be:3sgabs stair these.ine ‘John always slips in these stairs’

 Ane Berro

(16) Behin bakarrik neur-tu dira gainerako talde-ekin once only measure-tup be:3sgabs rest team-pl.with ‘they have competed with the rest of the teams just once’



(Berria, 2004-03-05)

(17) Etxea-ø erre da house-abs burn-tup be:3sgabs ‘The house has burnt’ (18) Jon-ø laban-du da Jon-abs slip-tup b:3sgabs ‘John has slipped’

The interpretation of these three examples differs notoriously from the interpretation obtained with the same verbs in (10), (11) and (12). The predicates from (16) to (18) are transitions: they convey a reflexive action –‘compete with, compare with, rival’ (16)– a change of state –‘burn’ (17)– or a change of ­position –‘slip’ (18). In (10), (11) and (12), in contrast, there is no transition conveyed, but only a property of the subject: its length, its high temperature and its slippery surface. Another characteristic proving the stativity of these predicates is that they generally lose their stative interpretation in the -tu analytic configuration.8 As more common stative predicates, they must appear in the -tzen analytic configuration in order to retain their stative interpretation (19). For instance, if some stative predicates are put within a -tu configuration, they become inchoative (Berro 2015) (20). (19) a. Jon-ek Miren-ø ezagu-tzen du  John-erg Mary-abs know-iprf have:3sgerg ‘John knows Mary’

State

b. Jon-ek istorioa-ø sinis-ten du  John-erg story-abs believe-iprf have:3sgerg ‘John believes the story’

State

(20) a. Jon-ek Miren-ø ezagu-tu du  John-erg Mary-abs know-tup have:3sgerg ‘John has met Mary’

Inchoative

b. Jon-ek istorioa-ø sinis-tu du  John-erg story-abs believe-tup have:3sgerg ‘John has believed the story’

Inchoative

.  Here, I am referring to the perfect or perfective analytic configuration, and not to the prospective, where we find -tuko headed predicates. For instance, in the prospective, these predicates can normally retain their stative meaning.



On the relation between ergativity, stativity and the verbal configuration of Basque 

Like these stative verbs, the verbal states analyzed in this section need generally to occur in the -tzen configuration in order to retain their stative interpretation. Most of them are ungrammatical or, at least, very weird in the -tu analytic configuration.9 (21) Platera-k *erre-ø/erre-tzen du dish-erg *burn-tup/burn-iprf have:3sgerg ‘The dish burns’; *‘The dish has burnt’ (22) Lantegi-ko kea-k gaizki *usain-du/usain-tzen du factory-gen smoke-erg wrong *smell-tup/ smell-iprf have:3sgerg ‘The smoke of this factory smells bad’; ‘*The smoke of the factory has smelled bad’

On the other hand, another remarkable feature of these predicates is the temporal interpretation obtained with the -tzen suffix. In Basque, the combination of the suffix -tzen plus the auxiliary in the present tense gives rise to a reading similar to that of the English present tense. As in English (Dowty 1979), the verbs belonging to different aspectual classes get different temporal interpretations when combined with the Basque -tzen + present auxiliary. Leaving apart the verbs with an available synthetic form,10 analytic verbs such as ‘eat’ (23) always have a habitual meaning with the -tzen suffix, whereas stative verbs, like ‘believe’ (24) get a non-habitual stative reading (Euskaltzaindia 1987; Albizu 2001; Alcázar 2002; Berro 2015). This temporal interpretation has been termed ‘indefinite’ (Euskaltzaindia 1987). (23) Jon-ek sagarrak-ø ja-ten ditu  Habitual Jon-erg apple.pl-abs eat-iprf have:3sgerg:3sgabs ‘Jon eats apples’ (24) Jon-ek Jainkoa-rengan sinis-ten du Non-habitual reading Jon-erg God-ine believe-iprf have:3sgerg ‘Jon believes in God’

.  As an anonymous reviewer notes, usaindu ‘smell’ can be grammatical in a -tu configuration like that of (22) if we add an appropriate temporal adverbial like arratsalde osoan, meaning ‘for the whole afternoon’. More common stative predicates (e.g. ezagutu ‘know’, sinistu ‘belive) can also be put within such a configuration and retain their stative meaning, triggering a meaning where the state holds in the interval denoted by the adverbial, but no longer in the utterance time. Therefore, still, the verbal states analyzed in this section behave like common stative predicates. .  Verbs with an available synthetic form do not show this contrast. Interestingly, the stative verbs which have an available synthetic form in the language always need to be realized in their synthetic form in order to retain their stative meaning. If they are put in the analytic -tzen configuration, they get a habitual interpretation (Euskaltzaindia 1987; Albizu 2001; Alcázar 2002; Berro 2015).

 Ane Berro

The predicates that we are analyzing in this section behave like stative predicates in this sense. They have an indefinite (non-habitual) temporal interpretation (see previous examples). As I have mentioned before, not all speakers accept this use of the verbs. When questioned, some speakers prefer to produce predicative structures consisting of an adjective or a noun plus a copula, instead of the verbal states like the ones described. (25) Hogeita bost bat milimetro luze ziren twenty five a millimeter long be:3plabs ‘They were about twenty five millimeters long’  (J. Gartzia, Sistema Periodikoa. Primo Levi: 190) (26) a. Piper hauek-ø oso pikanteak dira pepper these-abs very spicy.pl be:3plabs b. Piper hauek-ø minak dira pepper these-abs spicy.pl be:3plabs ‘These peppers are very spicy’ (27) Lurra-ø labana dago floor-abs slippery be(stage level):3sgabs ‘The floor is slippery’ (28) Jertsea-k usain txarra-ø dauka jumper-erg smell bad-abs have:3sgerg:3sgabs ‘The jumper smells bad’; lit.: ‘The jumper has bad smell’

I believe that the fact that some speakers resort to predicative structures in order to convey the meaning required is another proof which shows the stative nature of the verbal predicates in question. Other pieces of evidence in favor of their stativity come from the use of adverbs. On the one hand, they are incompatible with the process adverb astiroastiro/poliki-poliki ‘slowly’. On the other, they get a degree interpretation with the modifier apur bat/pixka bat ‘a little’. As for the adverbs meaning ‘slowly’, Fábregas and Marín (2012) have claimed that they are incompatible with non-dynamic verbs, like ‘shine’ or ‘wait’. When we combine such an adverb with these predicates of Basque, the result is ungrammatical. (29) *Platera-k poliki-poliki erre-tzen du     dish-erg slowly~slowly burn-iprf have:3sgerg ‘*The dish burns slowly’ (30) *Piperr-ek poliki-poliki pika-tzen dute     pepper:pl-erg slowly~slowly be.spicy-iprf have:3sgerg *The peppers are spicy slowly’



On the relation between ergativity, stativity and the verbal configuration of Basque 

Regarding the adverb apur bat/pixka bat ‘a little’, this modifier can have two readings depending on the type of the predicate: (i) a temporal reading, where it means ‘a short time’; and (ii) a degree reading, where it means that the amount of heat, light or whatever property described by the verb is small. In events, such as in the activity ‘dance’ (31), it can have both the temporal and the degree interpretation. In contrast, in stative verbs, it can only have the degree interpretation (32) (­Maienborn 2005; Rothmayr 2009; Fabregas and Marin 2012). Our verbs only admit the degree interpretation (33), showing once again that they are true stative verbs. (31) Ane-k apur bat dantza-tzen du  Ane-erg little a dance-iprf have:3sgerg  ‘Ane dances a little’

Temporal and degree reading

(32) Jon-ek Jainkoa-rengan apur bat John-er God-ine little a sinis-ten du believe-iprf have:3sgerg

Degree reading

‘John believes in God a little’ (33) Platera-k apur bat erre-tzen du dish-erg little a burn-iprf have:3sgerg ‘The dish burns a little’

Degree reading

Finally, the last test supporting the stativity of these verbs is that they cannot serve as the infinitive complements of perception verbs. According to M ­ aienborn (2005), since Davidsonian eventualities are spatiotemporal entities, they are perceptible and, thus, can serve as the infinitive complement of perception verbs (Higginbothan 1983). These predicates are verbal states -they are not spatiotemporal entities- and, consequently, they cannot be used as infinitives of verbs such as ikusi ‘see’. (34) #Platera-ø erre-tzen ikusi dut  dish-abs burn-iprf see-tup have:1sgerg ‘#I have seen the dish burning’ (35) *Kea-ø gaizki usain-tzen ikusi dut  smoke-abs wrong smell-iprf smell-tup have:1sgerg ‘*I have seen the smoke smelling bad’

(34) would be grammatical only with an interpretation where the dish is actually undergoing a change of state, but not with the one referring to its high temperature. Summing up, in this section I have presented and analyzed some new uses attested in Basque, where the predicate describes a property of its only argument and this argument bears ergative case. As can be concluded from the tests

 Ane Berro

­ resented, these verbs behave like stative verbs, a fact which leads us to the conclup sion that Basque, together with agentive unergative verbs, has also stative unergative verbs with holder ergative subjects. As an anonymous reviewer notes, since these predicates are only attested diastratically, their analysis cannot be generalized to all Basque speakers. Nevertheless, I consider that they represent further evidence of a fact which can be otherwise attested in any Basque variant: that holders of some stative predicates are marked ergative. Some examples are going to be shown in the following sections. 3.  Bare analytic predicates In Basque, there are some analytic predicates that cannot be aspectually modified as other predicates. They cannot be headed by the suffixes -tu or -tzen. I have termed these verbs bare analytic predicates,11 since they consist of a bare noun/DP, an adjective or an inessive postpositional phrase plus an auxiliary verb or a copula. Some of these verbs are presented in (36): (36) a. noun headed ardura izan axola izan atsegin izan12 balio izan behar izan beldur izan falta izan izena izan merezi izan uste izan zor izan

‘matter, lit. have importance’ ‘matter, lit. have importance’ ‘like, lit. be pleasure’ ‘be worth’ ‘must, have the need’ ‘fear’ ‘lack’ ‘call, lit. have name’ ‘merit, deserve, lit. be worthy’ ‘think, lit. have an opinion’ ‘owe, lit. be debt’

b. Adjective headed ageri izan ‘show, manifest, lit. be obvious’ aski izan ‘be sufficient’ gogoko izan ‘like, lit. have of desire’ gustoko izan ‘like, lit. have of pleasure’

.  De Rijk (2008: 321) calls these predicates Preterito-present verbs, since as he explains and we present in the following lines, in these predicates what looks as a periphrastic perfect tense indicates present. .  Atsegin can be both a noun or an adjective, meaning ‘pleasure’ or ‘pleasant’. Thus, it can be similarly listed in (36b).





On the relation between ergativity, stativity and the verbal configuration of Basque 

maite izan nahiago izan

‘love, lit. be dear’ ‘prefer, lit. have more wanted’

c. Headed by an inessive phrase falta-n izan ‘lack, lit. be in lack’ sobera-n egon ‘be left over, be not needed, lit. be in excess’ gogo-an izan ‘remember, have in mind’ gustu-an ükan13 ‘like, lit. have in pleasure’

These predicates are not real verbs, but just nouns/adjectives/postpositional phrases acting like primary predicates.14 I follow the analysis made in Etxepare and Uribe-Etxebarria (2012) (E&UE henceforth) for the modal behar ‘must’ and propose that the internal argument of these predicates is introduced within a small clause headed by the predicate, and that the ergative argument is introduced externally. Diverging slightly from E&UE, I argue that the arguments which take ergative case in these predicates are introduced by a P of central coincidence. Let us start analyzing the differences between bare analytic predicates (36) and the rest of analytic verbs. Bare analytic predicates cannot be selected by the -tu or the -tzen suffix, so that their infinitive and participial forms differ from the rest of the analytic verbs: they need to be accompanied by the auxiliary izan be or a copula -egon ‘(the stage level ‘be’) or ukan ‘have’. Thus, in conjugated perfective and imperfective forms, -tu and -tzen suffixes attach to the auxiliary or copula. As for example, in a habitual reading, the -tzen suffix modifies the auxiliary, as can be seen in (37) and (38) for the verbs gogoan izan ‘remember’, falta izan ‘lack’. (37) Gogo-an iza-ten nuen [luzaroan maite izan ninduela] mind-ine be-iprf have:1sgerg:3sgabs:pst ‘I used to have in mind that (s)he loved me for a long time’  (K. Navarro, Kontakizunak – Edgar Allan Poe: 25) (38) Sagardo usaina ere ez zen falta iza-ten cider smell also no be:3sgabs:pst lack be-iprf gure inguru-etan our environment-pl.ine ‘The smell of cider was not lacking in our environment’  (Txillardegi, Putzu: 48). .  The verb gustuan ukan has been taken from Etxegorri (2012). It is a verb used in Souletin Basque (a north-eastern variety). .  The verbal category of the predicate can even be questioned in analytic -tu and -tzen configurations, since -tze and -tu seem to show nominal distribution (see e.g. section  4.2, and Berro 2015 for a proposal claiming that all analytic predicates in Basque are of nominal category).

 Ane Berro

These verbs differ crucially from those which can be directly selected by -tu or -tzen in that their only meaning is stative (non-dynamic and non-eventive). Unlike the stative verbs presented in the previous section, in order to convey an indefinite reading (similar to the English present simple), these predicates do not need to be selected by -tzen: they just occur bare accompanied by the conjugated auxiliary/copula (39). (39) Hemen-go beste lagun batzuk-ø ditu gogo-an here-gen other friend some-abs have:3sgerg:3plabs mind-ine ‘He or she has other friends from here in mind’  (E. Jimenez,  Hemingway eta euskaldunak zerbitzu sekretuetan: 42)

Thus, these predicates are aspectually unmarked, and they have an imperfective aspectual reading. In this sense, they are similar to the synthetic configuration. In the synthetic form, predicates are not marked for aspect and, as presented in the introduction, they have imperfective value. The difference between them is that in the synthetic configuration, predicates surface as verbs and that in the bare analytic form, they surface as nouns, DPs, adjectives or PPs. Regarding bare analytic predicates, I follow E&UE’s analysis of the modal verb behar ‘must’ (40) and of transitive predication constructions (41). The latter have been analyzed previously in Rebuschi (1984) and de Rijk (2008) and also discussed in the present volume by Fernández and Rezac. (40) Amets-ek lo egi-n behar du Amets-erg sleep do-tui must have:3sgerg ‘Amets must sleep’ (41) Ni-k Xabier-ø aspaldi-ko adiskidea-ø dut I-erg Xabier-abs long.time-gen friend-ø have:1sgerg ‘Xabier is an old friend to me’ Lit: ‘I have Xabier an old friend’

According to E&UE, the modal behar ‘must’ is a nominal (beharN ‘need’) (see also Laka in press) heading a small clause. The subject of the small clause is the content of the need, which can be either a DP or an infinitive. This small clause is selected by an adposition (akin to an Applicative head), which introduces the experiencer of the need. In turn, this PP is selected by be. (42) a. T BE [PP Amets P [SC beharN lo egin]]

b. T P+BE [PP AmetsekERG P [SC beharN lo egin]]

The adposition incorporates onto be, yielding have auxiliary (à la Kayne 1993) and the subject checks ergative case with T (42b). E&UE suggest that a similar analysis must be adopted for transitive predicative structures such as (41). According to E&UE, constructions of transitive



On the relation between ergativity, stativity and the verbal configuration of Basque 

­ redication, involve an underlying structure similar to that of the modal behar p ‘must. In (41) the nominal adiskidea ‘friend’ and the object Xabier are part of a small clause (43), and the ergative subject is introduced as an experiencer by an applicative adposition, external to the small clause. Thus, the example in (41) could be paraphrased as ‘Xabier is an old friend to me’. (43) T P+BE [PP NikERG P [SC Xabier adiskidea]]

I argue that all bare analytic verbs fall within the same analysis (as a matter of fact, behar ‘need or must’ is a predicate listed in (36)), and that the adposition which introduces the external argument is an adposition of central coincidence. Let us consider the example in (39) of the verb gogoan izan ‘remember’, repeated and simplified in (44) for convenience. The direct object lagun batzuk ‘some friends’ is introduced within a small clause, headed by the PP predicate gogoan ‘in mind’. The small clause is then selected by an aspectual head, an adposition of central coincidence (PCENTRAL), which, in turn, introduces an external argument in its specifier position. The verb be then selects for the PP introducing the external argument and the small clause. (44) (Hark) lagun batzuk-ø ditu gogo-an    he/she.erg friend some-abs have:3sgerg:3plabs mind-ine ‘He or she has some friends in mind’ (45) T BE [PP Hura P [SC [lagun batzuk]DP gogoanPP]]

The only verbal element is the auxiliary be, which selects for a PP introducing the meaning “x with a property y”, yielding: ‘there is an x with the property y’. I follow E&UE in claiming that the adposition (an applicative in their analysis and an adposition of central coincidence in this one) incorporates onto be, which gives rise to have, à la Kayne (1993). As a consequence, another way of putting the structure into words would be ‘x has a property y’.15 This analysis is compatible with an approach of the ergative case where it is assigned structurally from T (here a complex T, actually, T[P+BE+T]) as in E&UE and Rezac et al. (2014) among others, but it could also be related to an approach where ergative case is inherently assigned (as in Laka 2006a) to the argument sitting the specifier of the central coincidence P. In any case, the ergative marking of the subject must be, directly or

.  Another possibility is to consider that be is not present as such in the structure and that it is just a support root which is lexically inserted in T for morphological reasons. The derivation proposed would proceed in similar terms: T[P+T] would be lexicalized as have and the subject would surface with ergative marking. This is actually the approach that I take for -tzen and -tu analytic configurations (see Examples 49 and 62) (see also Berro 2015).

 Ane Berro

indirectly, related to the central coincidence head, since only the arguments originating in the specifier of this head get ergative case. TP

(46) DP

T T[P+BE+T]

VP BE

PP DP

P P

SC DP gogoan

The idea is that the projection of P and its raising to be and to T motivates the lexicalization of have and the realization of the ergative marking. If P is not present, like in an unaccusative predicate where no external argument is introduced, be is spelled out and the subject surfaces with absolutive zero marking.16,17 In a transitive configuration, like that of (46), the absolutive zero case would be assigned to .  As an anonymous reviewer notes, it would be interesting to analyze, within a structural approach where ergative case is assigned by T, how that approach would generalize to the configurations where we find agents of passives and agents of non-finite configurations marked with ergative case: (iii) [Ni-k eros-i-ta-ko liburua-ø] garestia zen    I-erg buy-tup-ta-ko book-abs expensive was ‘[The book bought by me] is expensive’ (iv) [Ni-k liburua-ø eros-te-a] ez zen ideia ona izan    I-erg book-abs buy-nmlz-det no was idea good be ‘[Me buying the book] was not a good idea’ In these sentences in brackets, there is not an overt exponent of T that could be argued to be assigning ergative case to the subject. Nevertheless, Duguine (2013) has argued that in (ii) T is actually present and that it assigns structural case. I leave the analysis of these structures for further research. .  The ari progressive would represent another example where P is not present. Following the analysis made in Laka (2004, 2006b), I suggest that in ari constructions, the external argument is not projected as such. For reasons of space, I am not going to go deeper on this issue. The interested reader can consult Laka (2004, 2006b).



On the relation between ergativity, stativity and the verbal configuration of Basque 

the DP within the small clause, either by the P introducing the external argument or the head of the small clause. As can be concluded from this analysis, ergativity does not correlate with agentivity or the verbal category. In these particular cases, the external ergative argument is just the holder of a property, which in bare analytic predicates, can be denoted by a bare noun, a DP, an adjective or an inessive postpositional phrase. 4.  Generalizing the analysis In the previous sections, I have shown that, in Basque, ergative subjects may appear in contexts where the predicate is non-agentive, and in the particular case of bare analytic predicates, non-verbal. I believe that the fact that ergative subjects occur in such configurations must lead us to reconsider the rest of the verbal contexts and the rest of configurations where ergative subjects are present in the language. In this section, I am going to consider two specific configurations: (i) the verbal states presented in the second section; and (ii) eventive analytic verbs headed by the -tu suffix. I am going to claim that in both cases, the verb is nominalized, and that the external argument is introduced externally to this predicate, by means of an adposition of central coincidence. 4.1  Verbal states As I have presented in the second section, those new verbal states are analytic verbs which must be headed by the aspectual suffix -tzen. Regarding the -tzen suffix, In the progressive context, this suffix has been claimed to consist of a nominalizer morpheme -t(z)e plus the locative adposition -n (‘in’) (Demirdache and UribeEtxebarria 1997; Mateu and Amadas 1999; Laka 2004, 2006b). Leaving aside the role of the inessive adposition -n for the moment,18 I claim that verbs headed by -tzen in analytic predicates are nominalized and that they do not introduce external arguments (although this nominalization may not happen in syntax, see Berro 2015). Instead, the external argument is introduced in by an adposition of central coincidence. In the particular case of the verbal states, the nominalized verb denotes a property of the subject introduced by the adposition of central coincidence: ‘a property

.  The role of the inessive adposition -n may be more closely related to outer aspect (imperfectivity and progressivity), and it is still not clear to me how it interplays with the adposition introducing the external argument. I leave this issue for further research. The interested reader may also refer to Berro (2015), where a similar but different approach is adopted.

 Ane Berro

y on x’ or as ‘x with a property y’. Thus, a sentence like (47), which means ‘the smoke smells bad’, can be paraphrased as ‘bad smell is on the smoke’ or ‘the smoke with bad smell’. (47) Kea-k gaizki usain-tzen du smoke-erg wrong smell-iprf have:3sgerg ‘The smoke smells bad’

As a matter of fact, some speakers produced for the verb irristatu/lerratu ‘slip’ an alternative configuration to convey the same meaning, instead of the one presented in the second section. In this alternative configuration, the argument representing the source of the property or the holder of the property appears headed by the inessive adposition. (48) Zorua-n irrista-tzen da floor-ine slip-iprf be:3sgabs ‘The floor is slippery’

The example in (48) was recorded in Hernani (central), but it was also produced by speakers in Leitza (central) and Baigorri (north-eastern) (with certain varying aspects, due to dialectal variation).19 Interestingly, in the example where the holder argument is selected by the adposition, the auxiliary selected is be, instead of have. The production of such examples supports my analysis in two ways. Firstly, it shows that an argument can be marked with ergative case or be headed by the inessive adposition, and that in both cases it has the same theta role of holder. Secondly, it proves that when the holder argument is headed by an overt inessive adposition, the auxiliary selected is be (48), instead of have. We can consider that this fact indirectly points out that in the ergative clause, where the presumed adposition is non-overt, the adposition is lexicalized in the auxiliary itself – together with be, and thus, supporting the analysis of have as be+P (Kayne 1993; E&UE) (49a). ­Bellow, I show the simplified structure of these two different configurations:20

.  In every town, two speakers belonging to different generations were interviewed. It is ­interesting to note that in all the cases, only the older speaker produced an example such as (48). .  As an anonymous reviewer notes, the sentence in (48) may have a more complex structure, similar to the following Spanish sentence. (v) En este suelo (uno) se resbala On this floor (one) SE slips ‘In this floor, one can slip’, lit. ‘one slips’



On the relation between ergativity, stativity and the verbal configuration of Basque  TP

(49) a. DP zorua-k

T T P+be

PP DP

P P

VP (= nP) irrista-tzen

b.

TP

T be

PP DP zorua

P P -n

VP (= nP) irrista-tzen

4.2  Eventive -tu verbs In this section, I further generalize the claim that the external ergative argument is introduced by PCENTRAL to eventive analytic verbs headed by the -tu suffix. In previous lines, I have related the ergativity of Basque with the projection of an adposition of central coincidence which is also the source of have auxiliary. This seems to be more easily obtained in the case of stative verbs, such as verbal states and bare analytic predicates.21 The predicates headed by -tu are usually eventive, but still, I believe that they fall within the same analysis. This is because verbs headed by -tu are superficially nominal (like the -tzen headed predicates analyzed in the previous section), and not verbs. As I am going to show, -tu predicates

If this is the case, the predicate would be unaccusative, and ‘the floor’ would be introduced as a locative argument. In any case, ‘the floor’ in (49a) has also a locative kind of role, and interestingly, it is marked ergative, a remarkable aspect that has to be somehow explained. .  The idea that stativity is related to adpositions of central coincidence has been advocated in many works indeed (Hale and Keyser 1993, 2002, 2005; Fábregas and Marín 2012)

 Ane Berro

behave like nouns22 or adjectives in many aspects: (i) they can be selected by the determiner -a and also superlative suffixes to give rise to predicative participles; (ii) the prospective is obtained attaching the genitive -ko to -tu headed predicates; (iii) the adjective berri ‘new’ can modify -tu predicates and obtain an adverbial meaning of ‘recently’ or ‘just’; (iv) they can act as nominals and form copulative compounds of nominal category; (v) they can be selected by Axial Parts (Svenonious 2006; ­Etxepare 2013) to form temporal clauses (Uribe-Etxebarria 2014); and (vi) they can be the complement of the equivalent of the English preposition ‘without’, gabe. I claim that, since -tu headed predicates are nominal, they are related to the subject in the same way as bare analytic verbs and verbal states are, by means of PCENTRAL. The first piece of evidence which shows the nominal nature of -tu headed predicates comes from the fact that, in order to form intransitive (50) or transitive resultative participial predicates (51), the determiner -a ‘the’ selects for -tu headed verbs. This operation can be considered equivalent to the selection of a noun or an adjective by the determiner -a, which results in a DP (52a,b). Furthermore, out of the analytic context, -tu headed verbs can also be taken by comparative or superlative suffixes (53) (Hualde 2003: 204) and behave like adjectives (52c). (50) a. bezeroak-ø etorr-i dira guest.pl-abs come-tup be:3sgabs ‘The guests have come’ b. bezeroak-ø etorr-i-a-k dira guest.pl-abs come-tup-det-pl be:3sgabs ‘The guests are come’ (51) a. (Nik) aulkiak-ø erreserba-tu ditut   I-erg chair.pl-abs reserve-tup have:1sgerg:3plabs ‘I have reserved the chairs’ b. (Nik) aulkiak-ø erreserba-tu-a-k ditut    I-erg chair.pl-abs reserve-tup-det-pl have:1sgerg:3plabs ‘I have the chairs reserved’ (52) a. mendi-a-k mountain-det-pl ‘the mountains’

.  See also Haddican and Tsoulas (2012), who have argued that the -tu morpheme represents a nominalizing head, which in the case of participial verbal forms, is selected by a silent Aspectual head.



On the relation between ergativity, stativity and the verbal configuration of Basque 

b. mendi altu-a-k mountain tall-det-pl ‘the tall mountains’ c. mendi altu-en-a-k mountain tall-sup-det-pl ‘the tallest mountains’ (53) a. irakasgai ikas-i-en-a subject study-tup-sup-det ‘the most studied subject’ b. pelikula ikus-i-en-a-k film see-tup-sup-det-pl ‘The most seen film’

The verbs headed by -tu do not agree in number with the unaccusative subject (50a) or direct object (51a). However, when selected by -a, the participle can show plural number agreement (-k) with the corresponding unaccusative subject (50b) or the direct object (51b), in the same way as common DPs (52). Secondly, I believe that the way in which future tense is constructed in Basque is also relevant for my analysis. As I have commented in the introduction, future tense is constructed adding the irrealis suffix -ko (Laka 1996) to the lexical verb and selecting an auxiliary in the present tense. The relevant data is which category the suffix -ko selects. In the third section, I have stated that bare analytic predicates cannot be aspectually modified by -tu or -tzen. Interestingly, there is one suffix which most bare analytic predicates can take: the irrealis -ko. Bare analytic verbs headed by nouns and adjectives can be modified by -ko: among nouns, axola-ko ‘matter’, atsegin-go ‘like’, balio-ko ‘worth’, behar-ko ‘must’, falta-ko ‘lack’, uste-ko ‘think, have an opinion’, ardura-ko ‘matter’, merezi-ko ‘merit, deserve’, zor-ko ‘owe’ (but *berdur-ko ‘fear’, *izena-ko ‘call’); in adjectives ageri-ko ‘show, manifest’, askiko ‘be sufficient’, maite-ko ‘love’, nahiago-ko ‘preffer’ (but *gogoko-ko or *gustokoko ‘like, lit. of pleasure’).23 Finally, those headed by the inessive adposition cannot be either selected by -ko (*falta-n-ko ‘lack, lit. in lack’, *sobera-n-ko ‘be left over, be in excess’, *gogo-an-ko ‘remember, lit. in mind’, *gustuan-ko ‘like, lit. in pleasure’). On the other hand, it is remarkable that, in analytic verbs, the irrealis -ko attaches to the verb after -tu, and not before: apur-tu-ko and not *apur-ko ‘break (pros)’. This fact leads us to a conclusion: that the surface layer of bare analytic verbs and the layer projected by the -tu suffix have the same category. Both can be selected by -ko. Note at this point that -ko is a place genitive which can select .  That gogoN-ko izan and gustoN-ko izan cannot be selected by -ko could be argued to be due to the fact that they are already headed by the genitive -ko.

 Ane Berro

for bare nouns (mendiNP-ko ‘from mountain’ in mendi-ko etxea ‘the house at the mountain’) and other postpositional phrases, such as the allative (etxe-raALL-ko ‘to house’ in etxe-ra-ko bidea ‘the way home’ ‘the way which goes to home’). In this sense, it is important to highlight that the similarity between the irrealis -ko and the genitive -ko is far from accidental. As a matter of fact, eastern varieties of Basque make use of another irrealis marker, -(r)en -as in egin-en dut ‘I will do it’, which is also homophone with a genitive form.24 Thirdly, another context that serves to show the nominal nature of -tu headed verbs is that it can be further modified by the adjective berri ‘new’. In order to convey that the event has been recently carried out, the adjective berri ‘new’ can be inserted just between the -tu headed predicate and the auxiliary (54,55). As de Rijk (2008: 681) points out, when the adjective berri ‘new’ combines with the participle, it acts as an adverb meaning ‘recently’ or ‘just’, a meaning which does not get in any other context.25 (54) Lantegi-tik etorr-i berri da  (TOE III, 26) factory-abl come-tup new be:3sgabs ‘He has just come from the factory’, lit. ‘He is new come from the factory’ (55) Inazio-ø ikus-i berri dut  Inazio-abs see-tup new have:1sgerg ‘I have just seen Ignatius’, lit. ‘I have Inazio new seen’

(Lapitze, 143)

The fact that the element headed by -tu can be modified by an adjective, forming the complex attribute etorri berri ‘new come’ and ikusi berri ‘new seen’ supports further our claim. Let us analyze now the contexts involving the use of -tu headed predicates out of the analytic configuration. To begin with, the verbs headed by -tu can function like nouns (56) (Hualde 2003: 204). Furthermore, some of them can form copulative compounds of nominal category (57). These copulative nouns combine with determiners (58) and can trigger plural agreement, as common direct objects do (58a). (56) a. ja-n-a eat-tup-det ‘the food’ b. eda-n-a drink-tup-det ‘the drink’

.  I want to thank the anonymous reviewer for bringing into my attention this dialectal difference. .  The examples and the English translation are taken from De Rijk (2008:  681). I have added the literal translations.



On the relation between ergativity, stativity and the verbal configuration of Basque 

c. begira-tu bat look.at-tup-det ‘a look’ (57)

a. b. c. d.

hartu-eman joan-etorri sartu-irten jan-edan

‘interchange, relation, lit. to take-to give’ ‘round trip, lit. to go-to come’’ ‘small visit, lit. to enter-to go out’ ‘diet, food and drink, lit. to eat-to drink’

(58) a. Geure hartu-eman-a-k indar-tu behar ditugu our relation-det-pl reinforce-tup must have:3plabs.1plerg ‘We have to reinforce our relations’ b. Sartu-irten bat egi-n zuen atezaina-ren-ean small.visit det do-tup have:3erg caretaker-gen-ine ‘(S)he made a small visit in the house of the caretaker’ ‘I paid a small visit to the caretaker’ (J. Urteaga, Ordaina zor nizun: 21)

On the other hand, another interesting aspect of the verbs headed by -tu is that they can be the complements of Axial Part heads (Svenonius 2006; Etxepare 2013) in temporal clauses, as discussed by Uribe-Etxebarria (2014) (59). According to Svenonius (2006), in a decomposition analysis of adpositions, PlaceP is further decomposed into Axial Part and KP. As suggested by Etxepare (2013), in a Basque adpositional Phrase such as (59), the AxialPart head aurre ‘front’ selects for the noun etxe ‘house’. (59) Etxe-aurre-an house-front-ine ‘In front of the house’ (60) a. Miren-ø etorr-i oste-an joa-n naiz etxe-ra Miren-abs come-tup back-ine go-tup be:1sgabs home-all ‘I have gone home after Miren came’ b. Miren-ø etorr-i aurre-tik joa-n naiz etxe-ra Miren-abs come-tup front-abl go-tup be:1sgabs home-all ‘I have gone home before Miren came’

In temporal clauses (60), the Axial Part heads like aurretik ‘before, lit. by the front’ and ostean ‘after, lit. in the back’ select for -tu headed predicates. This fact gives us additional support for our claim. Finally, our last piece of evidence in favor of the nominal nature of -tu predicates is that they can be the complements of gabe, the equivalent of the English preposition ‘without’. In Basque, de Rijk (2008) points out that in old words gabe seems to be a noun, meaning ‘lack’ or an adjective, meaning ‘devoid’, and that nowadays, it has evolved to a stative adverb or a postpositon. It can combine with bare nouns (61a), with partitives (61b), with the indefinite determiner bat (61c), definite DPs (61d), and also, crucially, with the -tu morpheme (61f):

 Ane Berro

(61) a. Diru gabe, beldur gabe money without, fear without ‘without money’ ‘without fear’ b. diru-rik gabe, beldur-rik gabe money-part without, fear-part without ‘without money’, without fear’ c. hitz bat gabe, euro bat gabe word det without, euro det without ‘without a word’, ‘without an euro’ d. zu gabe, hori gabe you without, that without ‘without you’, without that’ f. ezer esa-n gabe, ikus-i gabe nothing say-tui without, see-tui without ‘without saying anything’, ‘without seeing anything’

Summing up, in this sub-section I have shown that the predicates headed by -tu share many distributional features with nouns (and adjectives).26 Therefore, I propose that they have a similar configuration to that found in -tzen verbal states. TP

(62) DP

T T[P+BE]

PP DP

P P

VP (= NP) DP

V (= N) ....-tu

With respect to the projection of arguments, I consider that the internal argument is projected within -tu predicates, but, that the external argument, in contrast, is projected by an adposition of central coincidence, like in verbal states and bare

.  The fact that -tu predicates behave in many contexts like adjectives could be derived from its relation with the aspectual head responsible for viewpoint aspect (see Berro 2015).



On the relation between ergativity, stativity and the verbal configuration of Basque 

analytic predicates. This adposition raises and incorporates onto T[BE], yielding have auxiliary. 5.  Conclusion In this paper I have shown that ergativity in Basque does not always correlate with agentivity and verbal forms. As a matter of fact, I have presented several new verbal states attested in the language, where the ergative argument is a holder, rather than an agent. Furthermore, I have also discussed bare analytic predicates, which are very common in Basque and where the ergative subject is a holder of a non-verbal stative predicate. This way, ergative case cannot be assigned by the verbalizer little v -as has been suggested for other ergative languages (Legate 2002, 2008; Aldridge 2004). Instead, I have argued that in these predicates the external argument is introduced by an adposition of central coincidence, which later incorporates onto be and T, yielding have auxiliary. The ergative marking on the subject would also be consequence of the projection of this adposition, either directly or indirectly. In the last sections of the paper, I have generalized this analysis to analytic predicates headed by the -tu suffix. In order to do so, I have presented several contexts where the predicates headed by -tu behave like adjectives and nouns, supporting their nominal category. As a conclusion and as a consideration for further research, I would like to discuss briefly how this analysis of the introduction of the external argument in Basque can be related to an analysis of argument structure. In the First Phase Syntax proposed by Ramchand (2008 et seq.), predicates are decomposed into three subevents: initiation (init), process (proc) and result (res). Init introduces the initiator argument in its specifier. Interestingly, Ramchand claims that there are two basic predicates over events – states and processes – and that both init and res belong to the first class. Thus, init is basically a state, which according to Ramchand, is interpreted as init only when it selects for (implicates) a process. The argument in the specifier position of init is, then, basically a holder, which in a specific c­ onfiguration – when init selects for proc – is interpreted as an initiator. I suggest that the distribution of ergative marked external arguments in Basque can be explained in these terms (see Berro 2015 for a proposal in these lines). The opposition between a state and a process can be viewed as a central vs. non-central opposition (Hale 1986). Thus, a state can be argued to be syntactically represented by a head of central coincidence, for example, an adposition (P­CENTRAL). This is actually what I propose for Basque. The ergative argument is in origin a holder argument introduced by PCENTRAL. In a certain

 Ane Berro

c­ onfiguration -when PCENTRAL selects for procP- that holder argument is interpreted as an initiator. Finally, I believe that this analysis of the analytic configuration of Basque can be framed within a long-standing claim which states the relation between ergative marking and adpositions (see, among others, Johns 1992; Mahajan 1994, 1997; Alexiadou 2001; Markman & Grashchenkov 2012; Torrego 2012). In particular, I think it is closely related to the proposals made in Alexiadou (2001) and Salanova (2007). Alexiadou (2001) suggests that the ergative clause structure is similar to a nominalization and that ergative case is parallel to a genitive adposition. Salanova (2007), on the other hand, posits a relation between the nominal nature of the verb in certain configurations of Mẽbengokre and the ergative marking found in that context. I leave the analysis of these correspondences and its implications for the theory of ergativity for future research.

References Alcázar, Asier. 2002. “Aspectual interpretation in Basque.” Ms., University of Southern C ­ alifornia, Los Angeles. Alberdi, Xabier. 2003. “Euskal aditz mailegatuen erregimena: hurbilketa”, In P. Lafitteren sortzearen mendemugako biltzarra. Gramatika gaiak, Iker 14 (I), Jesus Mari Makazaga and Beñat Oyharçabal (eds.), 37–60. Bilbao: Euskaltzaindia. Albizu, Pablo. 2001. “Sobre la distribución sintáctica de las formas finitas del verbo vasco: condicionamiento léxico y sintáctico.” ASJU 35: 65–106. Aldai, Gontzal. 2006. “Is Basque morphologically ergative? (i) The semantic split–intransitive case marking system of Western Basque.” In Andolin Gogoan. Essays in Honour of Professor Eguzkitza, Beatriz Fernandez and Itziar Laka (eds.), 117–138. Bilbao: UPV/EHU. Aldai, Gontzal. 2009. “Euskararen kasu markaketaren aldakortasun dialektala.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai [Supplements of ASJU 52], Beatriz Fernández, Pablo Albizu and Ricardo Etxepare (eds.), 5–19. Donostia: UPV/EHU and Gipuzkoako Foru Aldundia. Aldai, Gontzal. 2012. “Euskara morfologikoki ergatiboa al da?: Ekialdeko euskalkiak vs. mendebaldeko euskalkiak.” In 2nd Conference of the Luis Michelena Chair, Joseba A. Lakarra, Joakin Gorrochategui and Blanca Urgell (eds.), 337–351. Bilbo: UPV/EHU. Aldridge, Edith. 2004. Ergativity and Word Order in Austronesian Languages. Ph.D. dissertation, Cornell University. Alexiadou, Artemis. 2001. Functional Structure in Nominals. Nominalization and Ergativity. Amsterdam: John Benjamins.  doi: 10.1075/la.42 Berro, Ane. 2010. “Unergative predicates in Basque varieties: consequences for the ergative case assignment.” Certificate of Advanced Studies (IAD/DEA) research work. UPV/EHU. Berro, Ane. 2012. “Three levels of root insertion in Basque intransitive verbs.” In Journal of Portuguese Linguistics vol. 11 (1). Special Issue, Ernestina Carrilho and Beatriz Fernández, (eds.), 7–22. Lisboa: Ediçoes Colibrí. Universidade de Lisboa.  doi: 10.5334/jpl.89



On the relation between ergativity, stativity and the verbal configuration of Basque 

Berro, Ane. 2015. Breaking Verbs. From event Structure to Syntactic Categories In Basque. Ph.D. dissertation, UPV/EHU and Université Bordeaux Montaigne. Bobaljik, Jonathan. D. 1993. “On ergativity and ergative unergatives.” In Papers on Case and Agreement II [MITWPL 19], Colin Phillips (ed.), 45–88. Cambridge, Mass: MIT Press. De Rijk, Rudolf P.G. 2008. Standard Basque. A Progressive Grammar. Cambridge/London: The MIT Press. Demirdache, Hamida and Uribe-Etxebarria, Myriam. 1997. “The Primitives of Temporal Relations.” Ms., University of British Columbia/University of California at Irvine. Dowty, D.R. 1979. Word Meaning and Montague Grammar. The Semantics of Verbs and Times in Generative Semantics and in Montague’s PTQ. Dordrecht: D. Reidel Publishing Company.  doi: 10.1007/978-94-009-9473-7 Duguine, Maia. 2013. Null Arguments and Linguistic Variation: a Minimalist Analysis of Pro-Drop. Ph.D. dissertation, UPV/EHU and Université de Nantes. Etxegorri, Philippe. 2012. Biarnoko Euskaldunak. Elkar. Etxepare, Ricardo. 2003. “Valency and argument structure in the verb.” In A Grammar of Basque, Jose Ignacio Hualde and Jon Ortiz de Urbina (eds.), 369–465. Berlin: Mouton de Gruyter. Etxepare, Ricardo. 2013. “Basque primary adpositions from a clausal perspective.” Catalan Journal of Linguistics 12: 1–42. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2012. “Denominal necessity modals in Basque.” In Noun Phrases and Nominalizations in Basque: Syntax and Semantics, Urtzi Etxebarria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 283–330. Amsterdam/ Philadelphia: John Benjamins.  doi: 10.1075/la.187.13etx Euskaltzaindia. 1987. Euskal Gramatika: Lehen Urratsak II, Bilbao: Euskaltzaindia. Fábregas, Antonio and Marín, Rafael. 2012. “Differentiating eventivity from dynamicity: the Aktionsart of Davidsonian state verbs.” LSRL 42, Southern Utah University. Cedar City, UT. Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. Ph.D. dissertation, UPV/EHU. Haddican, Bill and Tsoulas, George. 2012. “A continuum of deficiency for Basque infinitives.” In Noun Phrases and Nominalizations in Basque: Syntax and Semantics, Urtzi Etxebarria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 437–460. Amsterdam/Philadelphia: John Benjamins.  doi: 10.1075/la.187.18had Hale, Kenneth. 1984. “Notes on world view and semantic categories: some Warlpiri examples. In Features and projections, Pieter Muysken and Henk van Riemsdijk (eds.), 233–254. ­Dordrecht: Foris. Hale, Kenneth and Keyser, Samuel Jay 1993. “On argument structure and the lexical expression of syntactic relations.” In The View from Building 20, Kenneth Hale and Samuel Jay Keyser (eds.), 53–109. Cambridge, Massachusetts: The MIT press. Hale, Kenneth and Keyser, Samuel Jay. 2002. Prolegomenon to a Theory of Argument Structure. Cambridge, Massachusetts: the MIT Press. Hale, Kenneth and Keyser, Samuel Jay. 2005. “Aspect and the syntax of argument structure.” In The Syntax of Aspect: Deriving Thematic and Aspectual Interpretation, Nomi Erteschik-Shir and Tova Rapoport (eds.), 11–41. Oxford: Oxford University Press

doi: 10.1093/acprof:oso/9780199280445.003.0002.

Higginbothan, James. 1983. “The logic of perceptual reports: an extensional alternative to situation semantics.” Journal of Philosophy 80: 100–127. Holmer, Arthur. 1999. “An active analysis of Basque ergativity.” Fontes Linguae Vasconum 31: 189–225.

 Ane Berro Hualde, Jose Ignacio. 2003. “Non-finite forms.” In A Grammar of Basque, Jose Ignacio Hualde and Jon Ortiz de Urbina (eds.), 202–210. Berlin: Mouton de Gruyter. doi: 10.1515/9783110895285 Johns, Alana. 1992. “Deriving ergativity.” Linguistic Inquiry 27: 57–87. Kayne, Richard. S. 1993. “Towards a modular theory of auxiliary selection.” Studia Linguistica 47 (1): 3–31.  doi: 10.1111/j.1467-9582.1993.tb00837.x Laka, Itziar. 1993. “Unergatives that assign ergative, unaccusatives that assign accusative.” In Papers on Case and Agreement I. [MITWPL 18], Jonathan D. Bobaljik and Colin Phillips (eds.), 149–172. Cambridge, MA: MIT Press. Laka, Itziar. 1996. A Brief Grammar of Euskara, the Basque Language. Euskararako Errektoreordetza. UPV/EHU. [http://www.ei.ehu.es/p289-content/eu/contenidos/ informacion/grammar_euskara/en_doc/adjuntos/Brief_grammar_euskara.pdf] Laka, Itziar. 2004. “Ari progresiboaz: euskararen kasu markak.” In Euskal Gramatika XXI. mendearen atarian: arazo zaharrak, azterbide berriak, Pablo Albizu and Beatriz Fernández (eds.). Bilbao: UPV/EHU. Laka, Itziar. 2006a. “On the nature of case in Basque: structural or inherent?” In Organizing Grammar, Hans Broekhuis, Norbert Corver, Riny Huybregts and Ursula Kleinhenz and Jan Koster, (eds.), 374–382. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110892994.374 Laka, Itziar. 2006b. “Deriving split ergativity in the progressive: the case of Basque.” In Ergativity: Emerging Issues, Alana Johns, Diane Massam and Juvenal Ndayiragije (eds.), 173–195. Dordrecht: Springer.  doi: 10.1007/1-4020-4188-8_7 Laka, Itziar. (in press). “Ergative need not split: an exploration into the TotalErg hypothesis.” To appear in The Oxford Handbook of Ergativity, Jessica Coon, Diane Massam and Lisa Travis (eds.). Oxford University Press. Legate, Julie. 2002. Walrpiri: Theoretical Implications. Ph.D. dissertation, MIT. Legate, Julie. 2008. “Morphological and abstract Case.” Linguistic Inquiry 39 (1): 55–101.

doi: 10.1162/ling.2008.39.1.55

Maienborn, Claudia. 2005. “On the limits of the Davidsonian approach: The case of copula sentences.” Theoretical Linguistics 31 (3): 275–316. Mahajan, Anoop. 1994. “The ergativity parameter: have-be alternation.” Proceedings of NELS 24: 317–331. Mahajan, Anoop. 1997. “Universal Grammar and the typology of ergative languages.” In Studies in Universal Grammar and Typological Variation, Artemis Alexiadou and T. Alan Hall (eds.), 35–57. Amsterdam: John Benjamins.  doi: 10.1075/la.13.03mah Markman, Vita & Pavel Grashchenkov 2012. “On the adpositional nature of ergative subjects.” Lingua 122 (3): 257–266.  doi: 10.1016/j.lingua.2011.10.010 Mateu, Jaume and Amadas, Laia 1999. “Extended argument structure: progressive as unaccusative.” Catalan Working Papers in Linguistics 7: 159–174. Oyharçabal, Beñat. 1992. “Structural and inherent case-marking; ergaccusativity in Basque.” In Syntactic Theory and Basque Syntax. [Supplements of ASJU 27], Joseba A. Lakarra and Jon Ortiz de Urbina (eds.), 309–342. Donostia: UPV/EHU and Gipuzkoako Foru Aldundia. Rezac, Milan, Albizu, Pablo & Etxepare, Ricardo. 2014. “The structural ergative of Basque and the theory of Case”. Natural Language and Linguistic Theory 32: 1273–1330. Rothmayr, Antonia. 2009. The Structure of Stative Verbs. Amsterdam: John Benjamins.

doi: 10.1075/la.143

Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First Phase Syntax. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511486319



On the relation between ergativity, stativity and the verbal configuration of Basque 

Rebuschi, Georgés. 1984. Structure de l’énoncé en basque. Paris: SELAF. Sarasola, Ibon. 2010. Dictionary of Contemporary Basque. Basque Institute. UPV/EHU. On line: http://www.ei.ehu.es/p289-content/en/contenidos/informacion/ euskara_inst_erdaretan/ en_erdaret/eeh.html Salanova, Andrés Pablo. 2007. Nominalization and Aspect. Ph.D. dissertation, MIT. Svenonius, Peter. 2006. “The emergence of Axial Parts.” In Trømso Working Papers on Language and Linguistics: Nordlyd 33 (1), Special Issue on Adpositions, Peter Svenonius (ed.). Trømso: University of Trømso. Torrego, Esther. 2012. “The unaccusative case pattern of Hindi and auxiliary be.” Lingua 122 (3): 215–224.  doi: 10.1016/j.lingua.2011.10.009 Uribe-Etxebarria, Myriam. 2014. “Asymmetries in the long distance reading of temporal clauses.” Hizkuntzalaritza Mintegia/Seminario Lingüística. 31st January. Vitoria-Gasteiz. UPV/EHU.

Datives in Basque bivalent unergatives Jon Ortiz de Urbina & Beatriz Fernández Deusto University / UPV/EHU

Bivalent predicates which mark their sole object dative rather than absolutive/ accusative are unexpected under the assumption that dative is associated with a ‘second complement’. Apparently first complements of morphologically transitive verbs are also found in Basque, in the main semantic classes already identified by Blume (1998). Morphologically, this sole object is indistinguishable from the indirect object of trivalent predicates of the give-type: both of them share dative case and trigger identical agreement marking on a ditransitive-like auxiliary form.  In this chapter, we will focus on the syntactic behavior of these datives and, following McFadden (2004), we will show that similarities with indirect objects extends also to syntax, as attested in: (i) secondary predication, (ii) impersonal/passive clauses, (iii) adnominals, and (iv) causativization and relativization. These dative objects, then, differ from DOM dative complements in Basque (Fernández and Rezac, this volume) both in syntactic behavior and in distribution. Keywords:  dative; unergatives; indirect objects; secondary predicates; impersonal clauses; adnominals, DOM

1.  Introduction1 Predicates taking dative marked complements without any other overt accusative complement, are crosslinguistically not uncommon (Blake 2001; Blume 1998). Moreover, as pointed out by Blume (1998), membership into this class is remarkably similar in languages exhibiting this phenomenon. Just as the number

.  We are very thankful to two anonymous reviewers for extremely helpful comments on earlier drafts of this paper. We are responsible for all remaining errors and inadequacies. This work has been partially supported by the Basque Government (IT665-13) and the Spanish Ministry of Economy and Competitiveness (FFI2014-51878-P9). The research leading to these results has also received funding from the European Union’s Seventh Framework ­Programme for research, technological development and demonstration under grant ­agreement nº 613465.

doi 10.1075/lfab.13.04ort © 2016 John Benjamins Publishing Company

 Jon Ortiz de Urbina & Beatriz Fernández

of ­languages exhibiting this type of marking is comparatively small, so is the number of predicates displaying it within these languages, but then, the predicates are remarkably similar across languages. Thus, the class of verbs taking dative ‘first’ complements in Basque features predicates well-known for this phenomenon such as listen, follow, help, obey or look at. For an ergative language like Basque, and assuming that absolutive in the transitive construction of an ergative language is grammatically equivalent to the accusative in a nominative-accusative language, the phenomenon in question can be described by saying that, given that dative is usually dependent on the presence of absolutive, it is surprising to find it associated with a complement without any co-occurring overt absolutive complement, as in (1): (1) a. Nik Joni begiratu diot2  I.erg Jon.dat look.at aux.3dat.1erg ‘I looked at Jon’

(bivalent unergative)

b. Nik ezin diot Joni entzun I.erg can’t aux.3dat.1erg Jon.dat hear ‘I can’t hear Jon’

We use the term ‘bivalent unergative’ as a purely descriptive label to indicate that we are dealing with a structure which is unergative in the sense that the external argument (nik ‘I’) is marked ergative, like transitive subjects, while there is no overt ‘direct object’ absolutive argument. It is bivalent because there is yet a s­ econd argument alongside the ergative subject, the dative complement in question (Joni ‘to Jon’).3 This is a marked situation in a language where most single complements of transitive verbs appear in the absolutive case and where dative is typically found only provided absolutive case is already associated with a nominal. This is so

.  Most examples come from corpora as gathered in Michelena and Sarasola’s General Basque Dictionary (Orotariko Euskal Hiztegia), or have been collected using the Corsintax corpora searching tool developed by Josu Landa for Ametzagaiña. For the sake of readability, we only give partial glosses for the auxiliaries, concentrating on the person and case morphemes which are relevant to the discussion. Abbreviations: abs = absolutive, aux = auxiliary, comp = complementizer, dat = dative, erg = ergative, inst = instrumentala, pl= plural, pst = past, gen = genitive, .  Monovalent unergatives predicates, i.e., monoargumental predicates with subjects marked ergative even though no object argument is present, are also found in Basque (Levin 1983; Laka 1993; Bobaljik 1993; Preminger 2012); examples include eskiatu ‘sky’ and disdiratu ‘shine’. They form a relatively small group since many crosslinguistically typical unergative predicates are expressed with light verb + object transitive constructions (hitz egin ‘talk’, lit. ‘do word’).



Datives in Basque bivalent unergatives 

regardless of whether the predicate is transitive or intransitive: in both (2) and (3) the theme argument is absolutive and the goal argument dative: (2) Mirenek eskutitza Joni bidali dio Miren.erg letter(abs) Jon.dat send aux.3dat.(3erg) ‘Miren sent Jon a letter’ (3) Miren Joni hurbildu zaio Miren(abs) Jon.dat approach aux.(3abs).3dat ‘Miren approached Jon’

Notice that (3) bears typical intransitive morphology, both in the absolutive marking of the subject and in the choice of intransitive auxiliary, while the constructions we will be concerned with here are morphologically transitive: the examples in (1) we will be examining are similar to the transitive in (2), but crucially lacking any overt absolutive complement.4 This paper will compare the grammatical behavior of the dative complement in bivalent clauses like (1) and in transitive clauses like (2), along the lines of McFadden’s (2004) work with German datives. The goal of the paper is to check whether single overt dative-marked complement of bivalent unergative displays grammatical properties of typical absolutive object complements or of dative indirect object complements. We will show the latter to be largely the case, so that grammatical similarities add up to observed morphological similarities. Given the descriptive thrust of this paper, we have been non-committal in calling these dative complements ‘apparent’ first objects, avoiding the analytical question of whether they are first complements with an exceptional case or second complements with an exceptional first complement (say, covert or somehow incorporated into the verb). These are in effect some of the major analytical approaches for these complements. In the first type of approach, a first complement (perhaps, as in Blake (2001) or Blume (1998) among many others, as a result of bearing a particular thematic relation to the predicate),5 receives a special (inherent) case.

.  This excludes the types of intransitive ‘dative object’ constructions which Blake (2001) describes as antipassive. .  Thus, Blume (1998) claims that “Interaction verbs with /nom/dat case frames always express complex events that involve two agents; and the activity of the nominative participant never manipulates or affects the activity of the dative participant. In other words, the participants of /nom/dat interaction verbs do not show agent-patient (dependence) relations in any of the subevents that are implied by the verb meaning.” These would be ‘weakly transitive interaction verbs’. Notice that Blume also claims the accusative nom/dat pattern corresponds more often to an intransitive ergative abs/dat pattern, rather than to the erg/dat pattern displayed by Basque.

 Jon Ortiz de Urbina & Beatriz Fernández

In an alternative approach,6 these are regular indirect objects akin to goal datives, made to look exceptional by the failure of the ‘first’ complement to appear as a regular, case-marked argument. We will be referring to these analytical issues in Section  2. In general terms, to the extent that we show that datives in bivalent unergatives share the same behavior as canonical datives in ditransitive constructions (goal datives, for instance), the facts described below seem to lend indirect support to analyses of the second type: datives with similar syntactic behavior would then have similar structural and derivational characteristics. We will briefly touch on this type of approach in an informal way as we review some of the predicates exhibiting this pattern in Section 2. Datives in bivalent unergative constructions differ from yet another type of single complement marked dative found in Basque Differential Object Marking constructions (Fernández 2008; Fernández & Rezac 2010, this volume), since the latter actually shares grammatical properties with canonical absolutive first object complements. Odria (2014, in progress), for instance, shows that DOM datives behave like canonical absolutive objects in their ability to license depictive secondary predicates, concluding that they must be generated in the same structural position. We will also be using depictive secondary predicates in Section  4.1, now to contrast datives in bivalent unergative (rather than DOM datives), with more canonical ditransitive dative complements (rather than canonical direct objects). There is great dialectal variation in the extent of first complement dative marking, and we will strive to incorporate existing microvariation into our description. In general terms, bivalent unergatives with dative complements are more common in western varieties of the language, where many such predicates have displayed similar behavior since the earliest records (16th, 17th centuries; see Mounole 2012). As we will see, dialectal patterns are sometimes affected by the interaction with the contemporary standard dialect, resulting in additional analytical complexity. The paper is organized as follows. Section 2 presents the major predicates that exhibit this marking pattern. In Section 3, we briefly turn to a contrast between DOM datives and bivalent unergative datives. Finally, S­ ection 4 concentrates on the syntactic behavior of the latter.

.  In fact, this is a fairly standard way to interpret these datives as unexceptional second complements in many traditional grammars. See Torrego (2010), or Sáez (2009) for technical implementation of some Spanish cases.



Datives in Basque bivalent unergatives 

2.  The predicates As indicated earlier, class membership into the group of predicates displaying this type of dative marking is cross-linguistically quite similar (Blume 1998). Thus, McFadden (2004) lists the following German verbs as examples of Blume’s predicate classes:

(4) Dative taking predicate classes in German:

a. Verbs of communication: zuhören ‘listen to’, antworten ‘answer’, schreiben ‘write to’, danken ‘thank’ b. Verbs of relative motion: folgen ‘follow’, begegnen ‘meet’ c. ‘Obey verbs’: gehorchen ‘obey’, dienen ‘serve’, helfen ‘help’.

Very similar predicates for these classes can be found in Basque and will be examined in this paper. However, although this pattern is still relatively uncommon in the language, more predicates than those belonging to the classes in (4) select for dative complements in Basque. Consequently, we add three more classes to host some predicates mentioned later in the article, with the caveat that class labels are impressionistically given for ease of reference (particularly so in the catch-all class ‘other’):

(5) Dative taking predicates classes in Basque

a. Verbs of communication: abisatu ‘notify’, aditu ‘listen to’, deitu ‘call’, entzun ‘listen to’, erregutu ‘pray’, eskertu ‘thank’. b. Verbs of relative motion: jarraiki/jarraitu ‘follow’, segitu ‘follow’, jazarri ‘chase’, esetsi ‘attack, flow, persecute, chase’ c. ‘Obey verbs’: manatu ‘order’, obeditu ‘obey’, lagundu ‘help’ d.  Physical contact verbs: jo ‘hit’, eutsi ‘hold’, heldu ‘hold’ oratu ‘hold’, ukitu ‘touch’ e. Aspectual verbs: ekin ‘engage in’, eman ‘start’, eutsi ‘keep on’, heldu ‘to hold on to something’ f. ‘Others’: begiratu ‘look at’, erreparatu ‘be aware of ’, iguriki ‘wait’, itxaron ‘wait’

As the number of predicates increases, so does the possibility that they may correspond to more than one single syntactic pattern; but we will proceed on the assumption that they do not.7 Since out tests in Section 4 agree with McFadden’s

.  The list is relatively extensive partly because it includes verbs which, at least in some ­dialects, can take dative marked complements.

 Jon Ortiz de Urbina & Beatriz Fernández

in showing that unergative datives pattern with (‘second’ complement) datives in ditransitive structures, they can be taken to indirectly support ditransitivelike analyses for these apparent first complements. We therefore briefly (and informally) dwell on this analytical approach throughout this section, primarily devoted to a presentation of unergative bivalent predicates in Basque. At least in types (a) and (c), we seem to be in front of verbs which involve transfer of a message or order, i.e. fairly common trivalent predicates which might be used here with a silent object of sorts, perhaps as in conflation structures along the lines of Hale & Keyser (1993, 2002). This is also the intuition behind traditional analyses which posit the existence of a ‘hidden object’ in all these cases (and which, therefore, do not recognize a bivalent ergative complementation type in Basque distinct from the ditransitive type; see Fernández & Ortiz de Urbina 2012). Physical contact verbs of type (d), on the other hand, are not mentioned in Blume (1998) and seem to differ semantically from the complex event characterization proposed there. In fact, on a first approach, they correspond to fairly prototypical transitive configurations where the theme would be marked absolutive/ accusative. However, the class is not restricted to Basque and can also be found with dative complements in Spanish. Thus, Fernández-Ordóñez (1999) mentions ­Spanish tocar ‘touch’ as a verb co-occurring with dative clitic complements providing examples like the following: (6) María se ha caído. No le toques en la María has fallen not her.dat touch in the herida/la herida/ø wound/the wound/ø ‘María has fallen. Do not touch her/her wound’

In fact, Fernández-Ordóñez explicitely argues for a structure with three arguments, There would be a complement which ‘belongs or is part of the entity denoted by the indirect object’ and which may also be expressed by a canonical object (la herida), a locative phrase (en la herida) or be left unexpressed (but implicit). A similar approach may be extended to Basque examples with similar predicates like the following: (7) a. telefono-listinarekin jo zidaten, buruan batez ere phone-book.with hit aux.1dat.3plerg.pst head.on specially ‘they hit me with the phone-book, especially on the head’ b. tukurrukutukutuku/ ez neri ikutu not I.dat touch ‘tukurrukutukutuku, do not touch me’ (folk song)

No overt absolutive object is found in the clause containing the dative in (a), while the extension introduces a locative inalienably possessed body part which would



Datives in Basque bivalent unergatives 

then be understood as the covert object for the verb.8 By localizing the contact to a part of the indirect object, this analysis fits well with the traditional view of dative objects as less directly affected by the event (Blake 2001) than regular accusative/ absolutive direct objects. Among the three basic classes described by Blume (1998), relative motion verbs like follow (class (b) above) seem perhaps least amenable to an ‘exceptional’/missing argument approach. However, even for these Fernández-Ordóñez (1999:1329) claims that a hidden object like ‘route, path’ may be understood in the Spanish verb seguir ‘follow’, so that this predicate is interpreted as ‘walk behind, follow one behind the other’ when it takes a dative complement, but as ‘chase’ when it takes an accusative one. This does not fully coincide with the Basque pattern, and a look at the other predicates in the (b) class above shows that in fact chase verbs are among the typical dative-taking members of this class. The meanings ‘chase’, ‘linearly/temporally follow’ or ‘be a follower of x’ are all compatible with both the western bivalent unergative construction and the eastern/northern bivalent unaccusative one. In trying to find possible candidates for a ‘deep’ object, one may point out that there are examples of this predicate with overt absolutive objects like bidea ‘way’, ohitura ‘custom’ or moda ‘fashion’. However, these are found in regular monotransitive constructions, with no other argument, rather than in ditransitive constructions overtly displaying the purported three participants that would provide the basis for a regularizing account of the unergative bivalent construction: (8) Arraldeko bizilagunek moda jarraitu nahi izan dute9 Arralde.of dwellers.erg fashion follow want have aux.(3abs.)3plerg ‘Arralde inhabitants have wanted to follow the fashion’

It is unclear that this absolutive argument could provide a third participant for the analysis of unergative bivalent constructions with jarraiki/jarraitu as canonic ditransitive structures, since, in fact, it may appear as the dative argument in the former: (9) idatziak hutsunea duenean, ohiturari jarraitu writing.erg gap.abs have.when, custom.dat follow behar zaio need aux ‘when the writing has a gap, one should follow the custom’

.  Examples like (b) may be also close to DOM phenomena in Basque, to which we will return below in Section 3. .  The source is a Facebook message which includes a twit by a Basque TV program. The twitted message has absolutive moda jarraitu, while the Facebook message referring to it, a couple of lines above, has dative modari jarraitu. https://es-es.facebook.com/goenkale/ posts/486117588121765

 Jon Ortiz de Urbina & Beatriz Fernández

(10) sentimentuari nahiz  modari jarraitu barik feeling.dat or fashion.dat follow without ‘without following neither feelings nor fashion’

It would be contrived to posit yet another hidden absolutive of this kind in bivalent unergatives like (9,10). Notice also that this predicate occurs with a bivalent unaccusative complementation pattern (intransitive verb with absolutive subject and dative complement) in the earliest texts, a pattern which remains at present in northern dialects, which use jarraiki. No question of any extra argument arises in that unaccusative configuration, and there does not seem to exist any semantic difference which would warrant positing any extra argument. Moreover, in those dialects, the path (or similar) argument, appears also in the dative, as in the early northern examples (11) and in (12), by now close to an idiomatic expression: (11) Jesus yarraiki zitzaion bere bideari Jesus.abs follow aux.3abs.3dat his way.dat ‘Jesus continued on’

(Leiçarraga, II 208)

(12) Aspaldiko ohiturari jarraiki… long.ago.of tradition.dat follow ‘Following an old tradition…’

This seems to indicate that the dative in the bivalent unergative use corresponds to this same argument and could hardly be posited as an underlying absolutive hidden argument in such constructions. It is, however, tempting to extend to this predicate an approach which seems generally promising for other predicates and verb classes. It seems to us that an approach similar to that taken for verbs of the ‘touch’ type may also be attempted. Remember that Fernández-Ordóñez points out that a body part or possessed object argument is often found overtly with these predicates, providing support for the covert presence of an extra argument in the apparent bivalent unergative uses of verbs like ukitu ‘touch’. In the case of a predicate like jarraitu/jarraiki, in turn, we may capitalize on its inherent ‘relational’ semantics: it is not only that there are two movement subevents (in the basic meaning), as indicated by Blume (1998), but the participants are also involved in a spatial or temporal relationship of ‘being behind’ or ‘after’. Thus, locative phrases containing relational locative nouns like atzetik, gibeletik ‘behind’ are also often found with this verb, and this inalienable ‘possession’ relation may provide some basis for an extra ‘argument’ required by the ‘hidden’ object approach to bivalent unergatives. The dative argument is also found, generally, with the aspectual usage of this verb (‘continue’), where it falls into the broader predicate class (5e) above. Other aspectual predicates which behave in this way are ekin ‘engage in’, eman ‘start’, eutsi ‘keep on’, or heldu ‘to hold on, to keep on’:



Datives in Basque bivalent unergatives 

(13) Eta irakurketari jarraitu zion and reading.dat follow aux.3dat.3erg.pst ‘And s/he continued (with the) reading’ (14) nasatik askatzeko maniobrari ekin dio dock.from free.to maneuver.dat engage aux.3dat.3erg ‘it turned to the maneuver of departing from the dock’

The doubling of jarraitu ‘follow’ as both a fully thematic and an aspectual predicate is not an isolated feature of this predicate, since the last three examples above share this feature: eman has a basic meaning and ditransitive use as ‘give’, and both eutsi and heldu also have a basic meaning ‘hold’ (taking dative complements, see (5d) above). Similar dative-like markings are common for aspectual verbs (Spanish empezar a ‘begin’ or even English begin to). Aspectual verbs differ from the other predicates in (5) in that their typical complements, unlike those exemplified in (13) and (14), are clausal rather than nominal (or PPs, if datives are so analysed). They present problems which fall outside of the main topic discussed in this paper and we will therefore not be dealing with aspectual datives in the remainder of this paper. To finish up with this summary review of objectless dative-taking unergative Basque predicates, we must add another group of verbs, perhaps related to the ‘­contact’ group and containing predicates like begiratu ‘look at’, erreparatu ‘realize, see’, iguriki ‘wait’ and itxaron ‘wait’. Again, these also have similar counterparts in other languages, not only among those displaying dative first objects but also, from a different perspective, even in the occurrence with prepositional complements with markings similar to datives, as in Spanish mirar a ‘look at’, esperar a ‘wait’. It is not difficult to justify in an impressionistic way that datives are actually second objects for these verbs. In fact, this is diachronically the case for iguriki ‘wait’ and itxaron ‘wait’, if they derive from compounds like egun eduki ‘hold the day’ and hitz edun ‘have the word’, respectively (Michelena 1977; Mounole 2012). A different question is whether this can be justified on synchronic evidence, but we will not attempt to do this here. Instead, we will concentrate on the issue whether datives in bivalent unergatives like those in (5) pattern with canonical direct objects or goal-like ‘second’ complements. We will also touch on the question of marking variability across dialects to see whether the grammatical behavior in these constructions covaries with marking. But first, it is important to differentiate these dative complements from a related group of first complement datives which we believe are best treated independently, namely, Differential Object Marking (DOM) datives. 3.  Bivalent unergative datives and DOM datives Many Basque dialects exhibit Differential Object Marking (DOM) phenomena (Bossong (1991, 1997), Lazard (2001), Aissen (2003), among many others, and

 Jon Ortiz de Urbina & Beatriz Fernández

Fernández (2008); Fernández & Rezac (2010, this volume); Mounole (2012); Odria (2014, in progress) for Basque. The following examples illustrate the two types of datives, a bivalent unergative in (15) and a DOM dative in (16b): (15) Nik zuri begiratu dizut I.erg you.dat look.at aux.2dat.1erg ‘I looked at you’

(bivalent unergatives)

(16) a. Nik zu ikusi zaitut I.erg you(abs) see aux.2abs.1erg ‘I saw you’

(standard marking)

b. Nik zuri ikusi dizut I.erg you.dat see aux.2dat.1erg ‘I saw you’

(DOM)

Alongside standard (15), some dialects assign dative to human, definite first complements of verbs like ikusi ‘see’, as in (16b), contrasting with standard (16a). Notice, however, that marking in bivalent dative configurations like (15) is precisely not ‘differential’ at all, that is, we do not find complement subsets marked differently on the basis of factors such as animacy, definiteness or grammatical person. Moreover, while dative marking in a bivalent unergative like begiratu has been ‘standard’ in the dialects where the verb is used, DOM datives in examples like (16b) represent a relatively recent phenomenon. Basque DOM patterns can probably be related to increased contact with Spanish, where animate direct objects (but not inanimate ones) are marked with the preposition a, identical to that found with datives. Moreover, it is significant that in the Spanish of the Basque Country (Fernández-Ordónez 1999; Ormazabal & Romero 1998, 2001, 2007, 2013a,b), animate direct objects, whether masculine or feminine, tend to be pronominalized with the dative clitic le, especially in colloquial registers. This points at a DOM phenomenon perhaps associated to Spanish and not independently developed (Comrie 2012). While the distinction between the two types of phenomena is quite robust, it is nonetheless true that it seems to be occasionally blurred. A case in point is that of itxaron ‘wait for’. As a bivalent unergative, dative marking can be found with both animate and inanimate complements, as shown in (17): (17) a. Geltokian itxaron zion Mikeli stop.at wait aux.3dat.1sgerg.pst Mikel.dat ‘He waited for Mikel at the station’ b. Hurrengo emaitzari itxaron zion next result.dat wait aux.3dat.1erg.pst ‘He waited for the next result’



Datives in Basque bivalent unergatives 

However, some varieties within western dialects also admit absolutive marking of the complement, but in those cases it can exclusively be assigned to inanimates. This is not exactly a DOM restriction in that dative is compatible with both animates and inanimates, but, in those dialects admitting absolutive marking with this predicate, there is a differential effect based on animacy to the extent that this case can only mark inanimates. DOM-like effects attributable to language contact seem to be on the rise from the 20th century. Thus, the predicate begiratu ‘to look at’ itself presents a traditional pattern with dative and allative complements, as exemplified in. (18) Begiratzen nien, look.at aux.1erg.3pldat.pst begiratzen nuen gero egutegira look.at aux.1erg.pst then calendar.to ‘I looked at them, and then I looked at the calendar’

Dative marking has been traditionally associated with both animate and inanimate objects, while allative marking typically indicates general direction as in look back, look around etc. Absolutive complements have been only very occasionally attested until the 20th century, when, while still marginal, they increase in numbers and proportion (see the entry for begiratu in the OEH [General Basque ­Dictionary]). However, a cursory look at contemporary corpora reveals that the vast majority of absolutive complements are again inanimate, often with meanings like ‘examine’, ‘look up’, etc.: (19) a. Eta alde batetik eta bestetik and side one.from and other.from begiratu behar dira gauzak look.at need aux.3plabs things.abs

‘And things should be looked at from one side and the other’

b. Ozono begiratu nuen hiztegian ozone.abs look.at aux.3erg dictionary.in ‘I looked at [up] ‘ozone’ in the dictionary’

There are very occasional examples of animate absolutive complements. Interestingly, they may stem from hypercorrection: with DOM effects spreading in colloquial registers in western dialects, dative complements of bivalent unergatives like begiratu resemble, especially when animate, DOM patterns, perceived as influenced by Spanish. Although not always evident, then, DOM datives represent a related but distinct phenomenon from the bivalent unergative patterns we are discussing here.

 Jon Ortiz de Urbina & Beatriz Fernández

Odria (2014) shows that in fact, DOM datives pattern with canonical absolutive direct objects as opposed to goal indirect objects. In what remains of this chapter, we will show that bivalent unergative datives, on the other hand, pattern with the latter, justifying an analysis which treats them as being generated in the same type of configuration. 4.  Syntactic behavior of dative objects in bivalent unergatives We turn now to a comparison of the grammatical properties of datives in bivalent unergative and in ditransitive constructions. We will examine in some detail the behavior of these datives in secondary predication (4.1), impersonal/­­ medio-passive clauses (4.2) and adnominal contexts (4.3), also discussed in McFadden (2004), where the same comparison is made for single dative complements of predicates in German. In (4.4), we check, albeit in a cursory way, dative behavior with respect to accessibility to relativization and doubling in causatives. The fact that in some of the predicates to be examined dative marking may alternate with absolutive marking and the fact that we are not restricting out attention to a single dialect introduces a fair amount of complexity into the description. As we will show, however, once we tease out different dialectal factors, unergative datives pattern mostly with canonical indirect object datives, supporting a similar treatment. 4.1  Dative first objects and secondary predication Depictive adjectives and some adverbial modifiers10 are restrictive as to the type of ‘controller’ they are predicated about. Typically, they can refer back to the subject or direct object of the primary predication, and Haider (1985) and McFadden (2004) show that German indirect objects and (inherent) datives cannot license secondary predicates. Similarly, neither Basque indirect objects (Zabala 1993; Arregi & Molina-Azaola 2004; Oyharçabal 2010; and Odria 2014) nor datives of bivalent unergatives (Fernández and Ortiz de Urbina 2012; Odria in progress) can license them: (20) a. Nik Jon biluzi ikusi dut I-ergi Jon.absj nakedi/j see aux ‘I saw Jon naked’ .  See E. Schultze-Berndt and N. Himmelmann (2004) for similarities and differences from a cross-linguistic perspective.



Datives in Basque bivalent unergatives 

b. Nik Joni biluzi lagundu diot I.ergi Jon.datj nakedi/*j help aux ‘I helped Jon naked’ c. Nik Joni arropa biluzi erosi diot I.ergi Jon.datj clothes.abs nakedi/*j buy aux ‘I bought Jon clothes naked’

Plank (1985) and Simpson (2005), quoted in Müller (2008), showed that German datives can act as secondary predicate controllers when a competing possible controller is eliminated, as in impersonal constructions, for instance: (21) a. Mani half ihmj erst halbtoti,j one helped him.dat only half-dead b. You can’t give them injections unconscious

We have not been able to replicate these data in Basque, but the relative pragmatic prominence of some arguments does interfere with secondary predicate licensing in other contexts. Thus, consider the predicate haserre ‘angry’ in (23), in the intended construal where it is predicated of the dative ugazabari ‘(to) the boss’. The primary predicate entzun ‘hear’ exhibits variation in its complement, which can be absolutive and also dative. Such variation is not related to meaning, so that in either complement configuration the verb can mean both hear and listen to.11 One would expect the same clear-cut distinction in Basque as one finds for English; however, the dative controller is only slightly worse than the absolutive controller: (22) I heard/*listened to the bossi angryi (23) ugazabai/ (?)ugazabarii haserrei entzun boss.abs    boss.dat angry hear ‘to hear the boss angry’

The same is not true of ikusi ‘see’ versus begiratu ‘look at’: the dative controller is as implausible as the prepositional phrase in English: (24) I saw/*looked at the bossi angryi (25) a. Ugazabai haserrei ikusi nuen boss.abs angry see aux ‘I saw the boss angry’

.  The dative pattern is often explained (away) in traditional grammar as corresponding to ‘hear someone (say) something’. This may be true at some level, but prima facie many of the examples are cases which would be simply translated as ‘listen to’.

 Jon Ortiz de Urbina & Beatriz Fernández

b. *Ugazabarii haserrei begiratu nion     boss.dat angry look.at aux ‘I looked at the boss angry’

One possible difference between hearing (25b) on one hand and looking at (25b) on the other is that, from a complementation point of view, the former but not the latter may take the clausal type of complement found with perception verbs, similar to causatives; from a pragmatic point of view, the agentive subject is prominent in the latter, while in the former there is an experiencer subject and prominence lies on what is perceived. If this is at all related to the question at hand, we would expect the entzun ‘hear, listen on’ examples in (23) with a depictive to mean hear rather than listen to, since the predicate listen to does not introduce a perception causative-like complement and seems to be agent prominent in the pragmatic perspective. This is the case indeed: even with dative complement, (23) means that I heard the boss and that the boss was angry. However, while this line of analysis may help us account for the difference between entzun ‘hear’ and begiratu ‘look at’, it does not explain why ikusi ‘see’ does not have a dative marked complement in non-DOM varieties. In any event, while the contrast see/hear is not clear, in the case of see the availability of dative marking may be connected with a special perception causative-like complementation pattern, since in the normal ditransitive pattern with a dative source and an absolutive theme argument the former cannot control a secondary predicate: (26) *Jonek ugazabarii erantzuna haserrei entzun zion    Jon.erg boss.dat answer.abs angry hear aux ‘John heard the boss’s answer [the answer to the boss] angry’

There is a further dialectal quirk to these data. The verb ikusi ‘see’ takes an absolutive direct object, as shown in (25a). Colloquial varieties sometimes present DOM effects and mark animate objects dative (see Ortiz de Urbina 1987, as well as Fernández-Ordóñez 1999 and Mendikoetxea 1999 for similar data in Spanish). However, in western varieties, much more generally and even in the case of speakers that do not exhibit such DOM effects, animate objects are marked dative in impersonal constructions: (27) Hemendik ez zara/zaizu ikusten here.from not aux.2abs/2dat see.impf ‘You cannot be seen from here’

Impersonal constructions are formally intransitive, so the implicit agent is not expressed and unaccusative patterns emerge. The only overt argument corresponds to the ‘deep object’, and one would expect that in a verb like ikusi an unaccusative configuration with a single absolutive argument should emerge, with the corresponding simple unaccusative auxiliary zara. However, this is unacceptable



Datives in Basque bivalent unergatives 

for most speakers and a bivalent unaccusative configuration emerges where the deep absolutive object shows up as dative instead, with the corresponding auxiliary zaizu. This dative argument can in turn control a secondary predicate, as (28) shows: (28) Mikelii oso gutxitan ikusten zaio haserrei Mikel.dat very seldom see.impf aux.3dat angry ‘Very seldom is Mikel seen angry’

This is unexpected since datives in bivalent unaccusative constructions are as unable to control secondary predicates as in ditransitives and bivalent unergatives: (29) Joni Mikelij urdurii,*j hurbildu zitzaion Jon.abs Mikel.dat worried approach aux.3abs.3dat ‘Jon approached Mikel worried’

The dative of impersonal objects which are assigned absolutive in non impersonal contructions seems therefore to behave like normal direct objects for secondary predication. The data resemble at first sight the pragmatic effects noted for (21) above. However, although we can’t explore the issue here, they should perhaps be connected with DOM phenomena, i.e. datives which correspond to absolutives in non-colloquial dialects. As indicated in Section  1, there is a north(east)/south(west) divide with respect to many of the predicates listed there, so that dative case is typically found in the latter, while the former often display objects marked with absolutive case in a standard monotransitive configuration. An obvious question which arises is then whether the behavior of the northern absolutive complement is similar or differs from that of the southern dative complement. With respect to secondary predicate licensing, our preliminary data seem to indicate that in fact, the same arguments admit depictives in the two dialectal areas regardless of case marking. Thus, a verb like segitu ‘follow’, with the same meaning and distribution as jarraitu ‘follow’ in Section  1, takes absolutive or dative along these dialectal lines. Following the normal distribution, a depictive adjective cannot be controlled by the dative in the southern dialects (30b) and, interestingly, neither can the absolutive marked object in the north (30a),12 even though absolutive objects of normal monotrasitive verbs often admit secondary predicates: (30) a. *(Nik

zui) biluzii segitu zaitut

   I.erg you.abs naked follow aux.2abs.1erg ‘I followed you naked.’

.  Beñat Oyharçabal (p.c.).

(north)

 Jon Ortiz de Urbina & Beatriz Fernández

b. *(Nik zurii) biluzii segitu dizut you.dat aux.2dat.1erg

(south)

The situation in northern dialects is reminiscent of similar facts discussed in Sáez (2009) for Spanish verbs similar to the ones we are discussing. One example is ayudar ‘help’, whose complement, although marked accusative and, therefore, doubled by accusative clitics (a), is shown by Sáez to exhibit behavior typical of dative complements, among others, the incompatibility with secondary predicates (b): (31) a. Ana la ayudó Ana her.acc helped b. *Juan ayudó a Maríai enfadadai.    Juan helped Mary angry

Sáez argues in this way for a covert object argument, positing a conflation analysis of a cognate object help with a low applicative (Pylkkänen 2008) head and the V selecting the latter, much in line with the traditional approaches mentioned in ­Section 1: an initial double object/complement construction would surface as a single complement one. This approach may be appropriate for northern Basque dialects, especially for narrowly similar verbs like lagundu ‘help’, which behave similarly in Basque and where a nominal root (lagun ‘companion’, lagun-tza ‘help’) undergoing conflation can be easily posited.13 In spite of the many subpatterns dialectal variation introduces, we can conclude that unergative datives show the same syntactic behavior with respect to depictive secondary predicate licensing as goal indirect object datives, providing ground for a similar syntactic treatment. 4.2  Dative first objects in impersonal constructions Dative complements in bivalent unergatives pattern with ditransitive datives in their behaviour in impersonal constructions (see McFadden 2004 for similar facts in German and Sigurðsson 2004 for Icelandic). Basque impersonals are formally intransitive structures, where what would typically occur as a subject argument in a personal construction is neither present as an overt nominal nor is marked in the auxiliary (Albizu 1998; Ortiz de Urbina 2003a: 579–584). Objects, whether first or second, retain the same case marking as in personal constructions, triggering the expected agreement in the intransitive impersonal auxiliary. The examples below .  However, it would not be directly applicable to Basque verbs like segitu ‘follow’, at least if the ‘missing’ argument for these predicates in northern dialects is also an inherently possessed noun as mentioned in the previous section for western dialects, since it would probably make conflation a morphologically less plausible process.



Datives in Basque bivalent unergatives 

show a personal (a) and impersonal (b) ditransitive. The dative indirect object cannot occur in the absolutive in the impersonal construction (c): (32) a. Joni mezua posta elektronikoz bidali diot Jon.dat message mail electronic.inst send aux.3abs.3dat.1erg ‘I sent the message to Jon by e-mail’ b. Joni mezua posta elektronikoz bidali zaio Jon.dat message mail electronic.inst send aux.3abs.3dat ‘The message has been sent to Jon by e-mail’ c. *Jon bidali da    Jon.abs send aux.3abs ‘Jon was sent (something)’

Absolutive case retention in impersonals is far from unexpected in an ergative language, even under radically different analyses. If internal arguments become subjects once the external argument has been ‘neutralized’ in an impersonal construction, they would be intransitive subjects, hence absolutive. If, on the other hand, the external argument is somehow present in the impersonal, as Albizu (1998) claims, the first complement would still be associated with the absolutive case of direct objects. For our argument based on impersonals to work, however, we must assume that the first analysis is correct, since if objects remain VP internal as the second alternative claims, no prediction is made about the case marking of the relevant arguments in impersonal constructions. Under the assumptions of the first hypothesis, however, if first objects become impersonal subjects, then complements of bivalent unergatives would also be expected to do so, consequently bearing absolutive case. As the following examples show, dative complements in bivalent unergatives (a) remain dative in impersonals (b) and cannot occur in the absolutive (c): (33) a. (Guk) Mikeli ordu erdi itxaron diogu  we.erg Mikel.dat hour half wait.for aux.3dat.1plerg ‘We waited for Mikel for half an hour’ b. Mikeli ordu erdi itxaron zaio Mikel.dat hour half wait.for aux.3abs.3dat ‘One waited for Mikel for half an hour’ c. *Mikel ordu erdi itxaron da    Mikel.abs hour half wait.for aux.3abs ‘One waited for Mikel for half an hour’

The pattern in (a,b) extends to all dative objects, regardless of animacy. The DOM-like effect mentioned in the previous section, however, introduces some variability in impersonals. Thus, while all the animate datives in (34), with

 Jon Ortiz de Urbina & Beatriz Fernández

an enlarged set of exemplified predicates, behave as expected, inanimate complements display variation (35): (34) a. *Mikel ordu erdi itxaron da    Mikel.abs hour half wait.for aux.3abs ‘One waited for Mikel for half an hour’ b. *Mikel begiratu da14    Mikel.abs look.at aux.3abs ‘Mikel was looked at’ c. *Mikel jarraitu da    Mikel.abs follow aux.3abs ‘Mikel was followed’ (35) a. *Trenak ordu erdi itxaron dira    trains.abs hour half wait.for aux.3plabs ‘The trains were waited for half an hour’ b. Egunkariak begiratu dira newspaper.abs look.at aux.3plabs Intended: ‘Newspapers were looked at’ c. Ildo berriak jarraitu dira way new.abs follow aux.3plabs ‘New ways were followed’

While (a), with itxaron ‘wait for’ is still unacceptable if the complement occurs in the absolutive, (b) and (c) are perfectly grammatical. This is so due to the fact that, as also mentioned above, these predicates can occur in monotransitive configurations with absolutive complements in meanings different from the one we are examining here. Thus, begiratu ‘look at’ has an accomplishment reading different from the activity reading under consideration and similar to examine. In the same vein, jarraitu ‘follow’, as mentioned in Section 2, occurs with absolutive complements like bidea ‘way, path’, ohitura ‘custom’, and this is the type of absolutive complement that also appears in the absolutive in the impersonal example (c). As Blume (1998) claims, follow someone does involve a ‘relative motion’ type of predicate with a complex subevent structure and two independent but related movements, while follow a trend only contains one moving agent.15 The presence .  Acceptable in a reflexive meaning. The sentence is also acceptable where begiratu is ­interpreted as ‘look after’, since this meaning occurs in a regular monotransitive configuration. .  It is not easy to extend this analysis to the difference between look at someone, which (also in the impersonal) would take dative complements in Basque, and look at a newspaper in (b), which is why the difference between the two has been given in aspectual terms in the text. An observed person, whether static or engaged in yet another activity, would take dative



Datives in Basque bivalent unergatives 

of absolutives in the impersonal clauses can therefore be the consequence of the alternating nature of bivalent unergatives: in those cases where underlying transitive configurations prevail, we can expect to find absolutive marking also in the impersonal, as in (35). Bivalent datives would then remain in the dative case, just like canonical indirect object datives.16 4.3  Nominalization genitives Following the lead of Kayne’s (1984) discussion on the resistance of double object datives to genitivize in nominalizations and Maling’s (2001) analysis of adnominal genitives in Germanic languages, McFadden (2004) uses genitivization to show that complements of deverbal nouns take genitive case when they correspond to the verbal accusative complement but not to a dative complement, whether the latter is the single complement or a canonical indirect object. Thus, just as in the German equivalent to the boy’s gift discussed by Kayne, the boy may not correspond to the goal of giving, in the nominalized form of a dative taking verb like helfen ‘help’ the priest’s help the priest may not correspond to the person helped. Similarly, Basque nominal arguments are marked genitive, but verbal arguments marked dative cannot appear as genitive in the corresponding nominal: (36) Jonen laguntza/eskaintza Jon.gen help/offer ‘Jon’s help/offer’

Jon in (36) can only receive an agentive interpretation, with an objective reading also marginally possible in the case of eskaintza ‘offer’; no goal interpretation is acceptable where Jon is the person receiving the help or offer. While eskaintza ‘offer’ is a deverbal noun related to a straightforward ditransitive verb, laguntza

case in Basque. On the other hand, a verb like itxaron ‘wait for’ can also occur with absolutive complements (Fernández & Ortiz de Urbina 2010: 164–166), as in ateratzeko ordua itxaron ‘wait for the time to go out’, opening the possibility for absolutive arguments in impersonals. This follows better from Blume’s analysis in that while the train in wait for the train is engaged in an agentive-like movement subevent, and requires dative case as in (a), the time in wait for the time to arrive is only metaphorically moving, and most likely not in an agentive way. .  Mendikoetxea (1999: 1688) presents similar data for Spanish verbs with obligatory a in Spanish like avisar ‘warn’, convencer ‘convince’ or amenazar ‘threaten’ and if they are equivalent to Basque datives, the parallelism is quite strong. In any event, canonical indirect objects cannot become subjects of reflexive passive constructions in Spanish, and must remain indirect objects in impersonal constructions, just like both unergative datives and regular indirect objects in Basque.

 Jon Ortiz de Urbina & Beatriz Fernández

‘help’ is a nominal derived from, and probably inheriting the argumental structure of, one of the bivalent unergatives described in Section  2. As in German, these data support the similarity between dative first objects and regular second object datives. When more data and more varieties are taken into account, however, complicating factors arise. In particular, it is possible to find deverbal nouns related to some of our dative-taking predicates occurring with genitive complements, as in the following examples: (37) suminduen jarraipena/jazarpena/entzuketa/barkamena activists.gen persecution/persecution/phone tapping/absolution ‘the persecution/phone tapping/absolution of the activists’

The question is why these nominals differ from laguntza ‘help’ above or others derived from bivalent unergatives like begiratua ‘look’, deia ‘call’, whose genitive modifiers can only receive an agentive interpretation, and cannot correspond to the dative complement of the verb. These data can be understood if we bear in mind dialectal variation, including here consideration of the contemporary standard dialect as a variety interacting with traditional dialects in complex contact situations. The presence of alternations with absolutive objects for some of the predicates contributes to the availability of the corresponding genitive element. Thus, barkatu ‘forgive’ also occurs with absolutive complements corresponding not only to the fault forgiven but also to the person who is forgiven, so the genitive in (37) can easily derive from this type of marking. However, other factors may intervene. For instance, it is interesting to notice that most of the nominals in (37) correspond to neologisms typical of formal registers, not exactly the ‘dialectal’ varieties we are looking at. They have a decidedly bookish flavor which might indicate they correspond to a formal register which, in the case of Basque, is typically the standard literary language, and it is in this dialect that such examples can be found, and mostly with these neologisms. Traditional, neutral register nouns like laguntza ‘help’, begiratua ‘look’ or deia ‘call’, as mentioned above, follow the expected pattern. This may indicate that, like many of the absolutive complements of bivalent unergatives in that variety, these data derive from a hypercorrection tendency mentioned in Section  2 that regards dative first complements as resulting from Spanish-induced DOM phenomena. As usual, the picture is considerably complicated when microvariation is taken into account, but, bearing these provisos in mind, then, the generalization that unergative datives cannot appear as genitives in nominalizations seems to us to be quite robust, lending further support to similar syntactic treatments for unergative and ditransitive datives.



Datives in Basque bivalent unergatives 

4.4  Other contexts The previous sections describe Basque data in contexts comparable to those presented for German in McFadden (2004). We give a quick overview here of other, more language-particular data configurations such as relativization and causative case doubling. The conclusion is the same: once dialectal variation is taken into consideration, datives in bivalent unergatives pattern with canonical ditransitive datives. Beginning with relativization, it has been observed (­O yharçabal 1987, 2003) that, while subjects and direct objects are easily relativized (a), indirect objects are significantly degraded (b). This is of course in line with Keenan and Comrie’s (1977) relativization accessibility data, but it is nonetheless significant in a language where indirect objects agree just like subjects and direct objects: (38) a. ikusi dud-an gizona see aux.3abs.1erg-comp man ‘the man I saw’ b. ??liburua eman diod-an gizona    book give aux.3abs.3dat.1erg-comp man ‘the man I gave the book to’

Datives of bivalent unergatives fully pattern with those of ditransitive predicates and are judged to be equally degraded: (39) ??jarraitu/begiratu/lagundu/entzun diod-an gizona    follow/watch/help/hear aux.3abs.3dat.1erg-comp man ‘the man I followed/watched/helped/heard’

As expected, when we consider microvariation, relativization is perfectly acceptable in dialects where an absolutive complement is also found for these verbs. If the head noun, marked absolutive here, appears in the dative, case-matching will improve the acceptability of the relative construction. This is a general phenomenon in relative clauses (see de Rijk 1972 for an early description) and independent of the basic judgement we are considering here; it is a phenomenon which also affects regular second complement datives and provides therefore further motivation for the fundamental similarity of the two types of datives. Similar conclusions can be drawn from dative doubling in Basque morphological causatives (Deustuko Hizkuntzalaritza Mintegia 1989; Ortiz de Urbina 2003b; Odria 2014, in progress). Animate subjects of unaccusative verbs (in western dialects, probably a DOM effect), most subjects of unergatives and transitive subjects will appear as dative causees when their predicates combine with the causative verb arazi as in (40). If the basic predicate already contains a dative

 Jon Ortiz de Urbina & Beatriz Fernández

argument, the dative causee will cooccur with it in a dative doubling configuration which is often acceptable even if sometimes difficult to process. Agreement is reserved for the causee, helping disambiguate the meaning. The dative argument of the causativized verb behaves in the same way, whether ‘first’ or ‘second’ argument: (39) (niri) Joni hulbildu/begiratu/eman arazi zidaten.    I.dat Jon.dat approach/look at/give cause aux.1dat.3plerg.pst ‘they made me approach/look at/give (it) to Jon.’

Even though absolutive case might in principle have been available for the single complement of a bivalent unergative verb like begiratu ‘look at’, producing a canonical ditransitive configuration and avoiding a marked dative doubling pattern, this pattern is not used. Instead, the dative ‘first’ complement remains as a non-­agreeing dative cooccuring with the dative causee, just like the dative goal complements of unaccusative hurbildu ‘approach’ or ditransitive eman ‘give’. Notice that these data only argue for a syntactic similarity of the internal argument Jon for unaccusative, unergative and ditransitive complements, but do not argue for a similarity of these three with the cause argument encoded in the agreement. In fact, Odria (2014) shows that, unlike the internal argument of the embedded verbs in (40), this causee dative can license a depictive secondary predicate. The point here is that, as in previous tests, unergative datives pattern with indirect object datives, providing some extra support for analyses that posit structural similarities in the syntactic configurations in which the two appear.

5.  Conclusion In this article we have compared the grammatical behavior of dative complements of bivalent unergatives with that of canonical ‘second’ complement datives, bearing in mind part of the existing dialectal variation in Basque. Once dialectal variation is taken into account, the two types of arguments marked dative are shown to have largely similar properties. This does not mean that the two types of datives we have discussed should be treated as occurring in identical syntactic configurations. Thus, even though they behave similarly in the syntactic tests described above, they also behave differently in other respects. For instance, Ormazabal & Romero (2010) use the inability to be modified by depictive secondary predicates as evidence in favor of the PP status of many datives. Further evidence for this status would come from the dialectal failure of some datives, crucially including unergative datives, to trigger dative agreement (Etxepare & Oyharçabal 2008, Fernández, Ortiz de Urbina & Landa 2009). These two features differentiate them from causee datives (see Odria in



Datives in Basque bivalent unergatives 

progress and Zabala 1993). However, goal datives do show fairly strong agreement behavior, so a single explanation like the PP vs. DP nature of unergative datives will not capture the full complexity of the data. The phenomena described in this paper have been connected more to the need to posit the existence of a first complement in the argument structure of what we have descriptively called bivalent unergatives, thus equating them with canonical ditransitive configurations in this respect. Throughout this paper, we have made extensive use of microdialectal variation in Basque dialects. Such variation is interesting not only because it supplies us with cases of closely related possible grammatical systems, but also because these very systems often interact in complex ways. We have only considered a very small part of the interactions in a restricted grammatical domain, but it seems they must be taken into account even for sheer descriptive adequacy.

References Aissen, Judith. 2003. “Differential object marking: iconicity vs. economy.” Natural Language & Linguistic Theory 21: 435–483.  doi: 10.1023/A:1024109008573 Albizu, Pablo. 1998. “En torno a la inacusitividad sintáctica de las construcciones impersonales en euskera.” Interlingüística 8: 13–18. Arregi, Karlos and Molina-Azaola, Gainko. 2004. “Restructuring in Basque and the theory of agreement.” In Proceedings of the 23rd West Coast Conference on Formal Linguistics, 43–56. Sommerville, MA: Cascadilla Press. Blake, Barry. 2001. Case. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9781139164894 Blume, Kerstin. 1998. “A contrastive analysis of interaction verbs with dative complements.” Linguistics 36.2: 253–280. Bobaljik, Jonathan D. 1993.“On ergativity and ergative unergatives.” In Papers on Case and Agreement II [MIT Working Papers in Linguistics 19], Colin Phillips (ed.), 45–88. ­Cambridge, Mas.: MIT Press. Bossong, Georg. 1991. “Differential Object Marking in Romance and beyond.” In New Analyses in Romance Linguistics, Douglas Kibbee and Dieter Wanner (eds.), 143–170. Amsterdam/ Philadelphia: John Benjamins.  doi: 10.1075/cilt.69.14bos Bossong, Georg. 1997. “Le marquage différentiel de l’objet dans les langues d’Europe.” In Actance et Valence dans les langues d’Europe, Jacques Feuillet (ed.), 193–258. Berlin: Mouton De Gruyter. Comrie, Bernard, 2012. “Differential Object Marking revisited.” Plenary talk, Syntax of the World’s Languages V, Dubrovnik, October 1–4. Deustuko Hizkuntzalaritza Mintegia. 1989. “Inkorporazioa perpaus kausatiboetan.” In Sintaxi teoria eta euskara, Pello Sala­buru (ed.), 87–119. VII. Udako Ikastaroak. Donostia: UPVEHU. Etxepare, Ricardo and Oyharçabal, Bernard. 2008a. “Dative structures in standard and north-eastern Basque.” Licensing conditions at the interface. International Encounter at the ­Summer Courses of the University of the Basque Country. XXVI. Donostia, June 23–24.

 Jon Ortiz de Urbina & Beatriz Fernández Fernández-Ordóñez, Inés. 1999. “Leísmo, laísmo y loísmo.” In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 1317–1397.  Madrid: Espasa Calpe. Fernández, Beatriz & Milan Rezac. 2010. “Datibo osagarri bitxiak eta Datiboaren Lekualdatzea: ari nai diyot eta kanta egin nazu bidegurutzean.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai, Beatriz Fernández, Pablo Albizu and Ricardo Etxepare (eds.), 113–149. Bilbo: UPV/EHU. Fernández, Beatriz and Rezac Milan. This volume. “Differential object marking in Basque varieties.”. Fernández, Beatriz. 2008. “Quirky dative objects in Basque.” European Dialect Syntax Meeting III, Venice, September 18–21. Fernández, Beatriz and Ortiz de Urbina, Jon. 2012. “Dative (first) complements in Basque.”  Journal of Portuguese Linguistics, 11–1, 83–98.  doi: 10.5334/jpl.93 Fernández, Beatriz, Ortiz de Urbina, Jon and Landa, Josu. 2009. “Komunztadurarik gabeko datiboen gakoez.” In Beñat Oihartzabali gorazarre Ricardo Etxepare, Ricardo Gómez and Joseba A. Lakarra (eds.), 352–380. Bilbo: UPV/EHU. Fernández, Beatriz and Ortiz de Urbina, Jon. 2010. Datiboa Hiztegian. Bilbo: UPV/EHU. Haider, Hubert. 1985. “The case of German.” In Studies in German Grammar, Jindřich Toman (ed.), 65–101. Dordrecht: Foris. Hale, Kenneth L. and Keyser, Samuel Jay. 1993. “On argument structure and the lexical expression of syntactic relations.” In The view from building 20: Essays in linguistics in honor of Sylvain Bromberger, Kenneth L. Hale & Samuel Jay Keyser (eds.), 53–109. Cambridge, MA: MIT Press. Hale, Kenneth L. and Keyser, Samuel Jay. 2002. Prolegomenon to a theory of argument structure. Cambridge, MA: MIT Press. Kayne, Richard S. 1984. Connectedness and Binary Branching. Dordrecht: Foris Keenan, Edward L. and Comrie, Bernard. 1977. “Noun Phrase Accessibility and Universal Grammar.” Linguistic Inquiry 8.1, 63–99. Laka, Itziar. 1993. “Unergatives that assign ergative, unaccusatives that assign accusative.” In Papers on Case and Agreement I [MIT Working Papers in Linguistics 18], Jonathan D. Bobaljik and Colin Phillips (eds.), 149–172. Cambridge, Mass.: MITWPL. Levin, Beth. 1983. On the nature of ergativity. Ph.D. dissertation, MIT. Lazard, Gilbert. 2001. “Le marquage différential de l’object”. In Language typology and language universals: An international handbook, Martin Haspelmath, Ekkehard Köning, Wulf ­Oesterreicher and Wolfgang Raible (eds.), 16–50. Delhi: Bahri Publications. Maling Joan. 2001. “Dative: The heterogeneity of the mapping among morphological case, grammatical functions, and thematic roles.” Lingua 111: 419–464.

doi: 10.1016/S0024-3841(00)00039-5

McFadden Thomas. 2004. The position of morphological case in the derivation: a study on the syntax-morphology interface. Ph.D. dissertation, University of Pennsylvania. Mendikoetxea, Amaya. 1999. “Construcciones con se: pasivas, medias e impersonales.” In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 1631–1722.  Madrid: Espasa Calpe. Michelena, Luis. 1977. “Notas sobre compuestos verbales vascos”. Revista de Dialectología y Tradiciones Populares XXXIII, 245–271. Mounole, Celine. 2012. “Evolution of the transitive verbs in Basque and apparition of datively marked patients.” In Ergativity, Valency and Voice, Gilles Authier & Katharina Haude (eds.), 355–379. Berlin/New York: Mouton de Gruyter.



Datives in Basque bivalent unergatives 

Müller, Stefan. 2008. “Depictive secondary predicates in German and English.” In Secondary Predicates in Eastern European Languages and Beyond, Christoph Schroeder, Gerd Hentschel & Winfried Boeder (eds.), 255–274. Oldenburg: Bis-Verlag. Odria, Ane. 2014. “Differential Object Marking and the nature of dative case in Basque varieties.” Linguistic Variation, 14–2. Amsterdam/Philadelphia: John Benjamins. 289–317.

doi: 10.1075/lv.14.2.03odr

Odria, Ane. In progress. Differential Object Marking in Basque syntax. A microcomparative survey. Ph.D. dissertation, UPV/EHU. OEH. Orotariko Euskal Hiztegia [General Basque Dictionary]. 1987–2005. Luis Mitxelena and Ibon Sarasola. Bilbo: Euskaltzaindia, Desclée de Brower, Mensajero. [Available online at www.euskaltzaindia.eus]. Ormazabal, Javier and Romero, Juan. 1998. “On the syntactic nature of the Me-lui and the Person Case Constraint.” ASJU 32(2), 415–433. Ormazabal, Javier and Romero, Juan. 2001. “A brief description of some agreement restrictions.” In Kasu eta komunztaduraren gainean. On case and agree­ment, Beatriz Fernández and Pablo Albizu (eds.), 215–241. Bilbo: UPV/EHU Editorial Services. Ormazabal, Javier and Romero, Juan. 2007. “Object agreement restrictions.” Natural Language and Linguis­tic Theory 25. 315–347.  doi: 10.1007/s11049-006-9010-9 Ormazabal, Javier and Romero, Juan. 2010. “The derivation of dative alternations.” In Argument Structure and Syntactic Relations: A cross-linguistic perspective, Maia Duguine, Susana Huidobro & Nerea Madariaga (eds.), 202–232. Amsterdam: John Benjamins.

doi: 10.1075/la.158.13orm

Ormazabal, Javier and Romero, Juan. 2013a. “Differential object marking, case and agreement.” Borealis. An International Journal of Hispanic Linguistics 2(2): 221–239.

doi: 10.7557/1.2.2.2808

Ormazabal, Javier and Romero, Juan. 2013b. “Non accusative objects.” Catalan Journal of Linguistics 12: 155–173. Ortiz de Urbina, Jon. 1987. “Interferentzia sintaktiko berezi batez [On a special type of syntactic interference]”. 2nd World Basque Conference. San Sebastian, July 31, 1987. Ortiz de Urbina, Jon. 2003a. “Impersonal clauses.” In A Grammar of Basque, Jose Ignacio Hualde and Jon Ortiz de Urbina (eds.), 572–590. Berlin/New York: Mouton de Gruyter.

doi: 10.1515/9783110895285

Ortiz de Urbina, Jon. 2003b. “Causatives.” In A Grammar of Basque, Jose Ignacio Hualde and Jon Ortiz de Urbina (eds.), 592–607. Berlin/New York: Mouton de Gruyter.

doi: 10.1515/9783110895285

Oyharçabal, Bernard. 1987. Etude descriptive de constructions complexes en basque: propositions relatives, temporelles, conditionnelles et concessives. Ph.D. dissertation, Université de Paris VII Oyharçabal, Bernard. 2003. “Relative clauses.” In A Grammar of Basque, Jose Ignacio Hualde and Jon Ortiz de Urbina (eds.), 762–823. Berlin/New York: Mouton de Gruyter. Oyharçabal, Bernard. 2010. “Basque Ditransitives”. In Argument Structure and Syntactic Relations: A Cross-linguistic Perspective, Maia Duguine, Susana Huidobro and Herea ­ Madariaga (eds.), 233–260. Plank, Frans. 1985. “Prädikativ und Koprädikativ.” Zeitschrift für Germanistische Linguistik 13: 154–18.  doi: 10.1515/zfgl.1985.13.2.154. Preminger, Omer. 2012. “The absence of an implicit object in unergatives: new and old evidence from Basque.” Lingua 122(3): 278–288.  doi: 10.1016/j.lingua.2011.04.007

 Jon Ortiz de Urbina & Beatriz Fernández Pylkkänen, Liina. 2008. Introducing Arguments. Cambridge, Mass.: MIT Press.

doi: 10.7551/mitpress/9780262162548.001.0001

Rijk, Rudolf de. 1972. Studies in Basque Syntax: Relative Clauses. Ph.D. dissertation, MIT. Sáez, Luis. 2009. “Applicative phrases hosting accusative clitics.” In Little Words: Their History, Phonology, Syntax, Semantics, Pragmatics, and Acquisition, Ronald P. Leow, Hector Campos and Donna Lardiere (eds.), 61–75. Washington DC: Georgetown University Press. Schultze-Berndt, Eva and Himmelmann, Nikolaus. 2004. “Depictive secondary predicates in cross-linguistic perspective.” Linguistic Typology 8: 59–131.  doi: 10.1515/lity.2004.004 Sigurðsson, Halldór Ármann. 2004. “Icelandic non-nominative subjects: facts and implications.” In  Non-nominative Subjects, vol. 2, Peri Bhaskararao and Karumuri Venkata Subbarao (eds.), 137–159. Amsterdam: John Benjamins.  doi: 10.1075/tsl.61.09sig Simpson, Jane. 2005. “Depictives in English and Warlpiri.” In Secondary Predication and Adverbial Modification. The Typology of Depictives, Nikolaus P. Himmelmann and Eva SchultzeBerndt (eds.), 69–106. Oxford: Oxford University Press.

doi: 10.1093/acprof:oso/9780199272266.003.0002

Torrego, Esther. 2010. “Variability in the case patterns of causative formation in romance and its implications.” Linguistic Inquiry, 41: 3, 445–470.  doi: 10.1162/LING_a_00004 Zabala, Igone. 1993. Predikazioaren teoriak gramatika sortzailean [Predication Theories in ­Generative Grammar], Ph.D. dissertation, UPV/EHU.

Differential object marking in Basque varieties Beatriz Fernández & Milan Rezac UPV/EHU / CNRS – IKER, UMR5478

This chapter studies Differential Object Marking (DOM) in nonstandard varieties of Basque. DOM in Basque overtly resembles a common DOM pattern of coding direct as indirect objects in both case and agreement, that is absolutives as datives, according to their animacy and specificity, as in Spanish and HindiUrdu. However, Basque DOM is also sensitive to properties of the clause such as tense. In both respects, there is variation across Basque. Some varieties let us probe the nature of DOM datives, and reveal systematic patterning with nonDOM absolutives rather than with indirect object datives. This includes robust diagnostics for structural Case, such as ECM and absolutive-dative alternations. We propose a structural Agree/Case mechanism for DOM datives that brings out their relationship to absolutives and allows sensitivity to properties of both the object and the clause. Keywords:  differential object marking; dative case; dative agreement; animacy; leísmo

1.  Introduction1 Under certain conditions, nonstandard varieties of Basque code objects of transitives as dative rather than as the canonical absolutive. The phenomenon belongs to the family of constructions studied in the typological literature as Differential Object Marking (DOM). In Basque, the morphology of DOM has been described in dialectological studies, but syntactic inquiry is only at its beginning (Rezac 2006;

.  We want to thank many colleagues for comments, suggestions and discussion on some of the topics we will present here: Julen Agirre, Pablo Albizu, Kontxi Arraztio, Ane Berro, Kepa Erdozia, Urtzi Etxeberria, Ricardo Etxepare, Inés Fernández-Ordoñez, Richard S. Kayne, Orreaga Ibarra, Aitor Iglesias, Mikel Lersundi, Jesus Mari Makazaga, Cèline Mounole, Ane Odria, Beñat Oyharçabal, Javier Ormazabal, Jon Ortiz de Urbina, Juan Romero, Pello S­ alaburu, Ibon Sarasola and Igone Zabala. We are particularly grateful to Arantzazu Elordieta for

doi 10.1075/lfab.13.05fer © 2016 John Benjamins Publishing Company

 Beatriz Fernández & Milan Rezac

Fernández and Rezac 2010; Mounole 2012 and Odria 2012, 2014, in p ­ rogress, among others). In this chapter, we investigate syntactic differences between absolutive and dative coding of objects. The study leads us to three conclusions. First, DOM objects are direct objects: they have same thematic and argumental relations as canonical absolutive objects, and not those of dative indirect objects; in particular, they do not use applicative or prepositional structures. In this they contrast with dative internal arguments of bivalent unergatives, which are dative indirect objects. Second, the dative of objects under DOM reflects a structural Case: it is independent of argumenthood relations and so available for Exceptional Case ­Marking. More tentatively, like structural Case bearers and unlike inherent datives, DOM objects need to be licensed by Agree with a clausal locus. Third, Basque DOM is parametrizable by tense and finiteness/agreement. This supports the structural Case analysis because these never play a role in the inherent Case of internal arguments. On this score, Basque DOM differs from most nearly comparable phenomena, such as Spanish and Hindi-Urdu DOM. In Basque, the dative versus absolutive coding of objects has access to the features of T+v as well as those of the object. We propose a theory of DOM Agree/Case that fits these results: dative and absolutive codings are both due to Agree with v, differentiated by a property of v that is reflected as dative case/agreement and may involve object shift, and sensitive to any material in the Basque agreement complex which involves a complex head formation from v to Fin. The chapter is organized as follows. First, we briefly present the general properties of DOM and its theoretical analysis with Spanish leísmo and a-marking as well as Hindi ko-marking. Second, we describe the morphology of Basque DOM and delimit its parametric variation across Basque varieties, as we presently understand it. Third, we study its syntax, which leads us to the foregoing conclusions, building on Fernández and Rezac (2010) and Odria (2012).

sharing with us her intuitions in Lekeitio Basque and to Andoni Garai (and also Oier Garai) for patiently answering our questionnaire in Dima Basque. Dima data have been gathered by us in fieldwork and Araitz-Betelu data by Kontxi Arraztio. We also thank an anonymous reviewer for their suggestions and comments. This work has been partially supported by the Basque Government (IT665-13) and the Spanish Ministry of Economy and Competitiveness (FFI2014-51878-P9). The research leading to these results has also received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement n0 613465.



Differential object marking in Basque varieties 

2.  Differential object marking (DOM) In Basque, objects of plain transitives are canonically coded by absolutive case and agreement. However, some varieties code them as dative, like indirect objects, under certain circumstances, for instance if human. This fits the profile of ­Differential Object Marking in the typological literature (Bossong 1991, 1998; Lazard 2001; Aissen 2003): objects of transitives high on the animacy hierarchy show a marked coding, often identical to that of indirect objects. Such is mostly the situation in Romance languages, where DOM objects have the same marker a and sometimes the same clitic form as indirect objects (Bossong 1991), and frequent cross-liguistically, as in Hindi/Urdu (Mohanan 1994), Guarani (Shain 2008) and Tigre (Xasa) (Raz Shlomo 1980), although DOM can also take forms unrelated to indirect objects, as in Romanian (Dobrovie-Sorin 1994), Hebrew (Aissen 2003), Persian (Lambton, 1993) and Turkish (Kornfilt 1997; Enç 1991). We begin by briefly presenting the DOM of Spanish and Hindi-Urdu, for they are similar to Basque and have been studied for their syntax. 2.1  Spanish DOM Spanish has two DOM phenomena. One is a-marking, whereby the coding of indirect objects by a (1a) extends to certain objects of transitives that are otherwise accusative (1b). DOM a-marking occurs under several conditions, the best studied of which are animacy and specificity: it is obligatory for specific animate objects and impossible for nonspecific animates and mostly for inanimates, in contrast to indirect objects of ditransitives where neither factor plays a role (see Torrego 1998: Chapter 2; Ormazabal and Romero 2013):2 (1) a. Entregaron un libro *(a) un físico gave.they a book   p a physicist ‘They gave a book to a physicist’ b. Vieron (a) un físico saw.they   p a physicist ‘They saw a physicist’

[un físico specific with a]

The other DOM phenomenon is leísmo, by which the clitics representing or doubling indirect objects take the form of those of indirect objects. Spanish direct and indirect object clitics are syncretic in 1st/2nd persons, but in the 3rd person distiguish accusative and dative. In leísmo, some varieties code 3rd person m ­ asculine

.  Glosses to Spanish examples are p for the a-marking preposition or case-marker, cl.dat for the 3sg masc./fem. dative clitic, m/fsg for masc./fem. singular concord.

 Beatriz Fernández & Milan Rezac

animates by dative clitics, namely le in (2a), and others like the Spanish of the Basque Country also feminine animates (2b) (Fernández-Ordoñez 1999; Landa 1995). Leísmo only occurs if DOM a-marking also does. There are different types of leísmo; henceforth we keep to that of Basque Country Spanish, whose syntax has been studied in depth by Ormazabal and Romero (2007, 2013). (2) a. ¿Conoces a Juan? Sí, le conozco hace tiempo  know.you p Juan yes cl.dat know.I does time ‘Do you know Juan? Yes, I’ve known him for a long time’ b. A María hace tiempo que no le veo p María does time that not cl.dat see.I ‘María, it has been a long time since I saw her’  (Fernández-Ordoñez 1999)

Neither DOM coding is found in passives, where the object of the active promotes to an agreeing nominative (3a), unlike the indirect object of a ditransitive (3c) (Ormazabal and Romero 2007). Thus DOM participates in the same alternation as canonical accusative coding, giving active/accusative ~ DOM versus passive/ nominative. Such alternation has been one diagnostic of structural rather than inherent Case (Torrego 1998: Chapter 2). (3) a. María fue vista María was seen.fsg ‘María was seen’ b. *Le fue vista/visto *cl.dat was seen.fsg/msg c. El libro le fue entregado a Juan the book cl.dat was given.msg p Juan ‘The book was given Juan’

Studies of the syntax of Spanish DOM have led to two conclusions. One is that a-marked DOM objects are structurally Case-marked, perhaps in contrast to inherent Case on indirect objects. The other is that both a-marking and leísmo DOM reflect a syntax distinct from that of canonical accusatives, not only a morphological difference. This special syntax has been analysed as object shift, required for certain objects like specific animates (Torrego 1998: Chapter 2; cf. Diesing and Jelinek 1995), or in certain configurations like Exceptional Case Marking configurations (4) where even inanimates show DOM (Ormazabal and Romero 2013): (4) Le veo al avión caer envuelto en llamas cl.dat see.I p.the plane fall surrounded in flames ‘I see the plane fall down enveloped in flames’



Differential object marking in Basque varieties 

Superficially, Spanish DOM bears striking similarities to Basque DOM: in both, DOM affects both object coding (a-marking, dative case) and its cross-referencing (dative clitic, dative agreement), and in both DOM coding is identical to indirect object coding (Rezac 2006; Austin 2006; Fernández 2008; Fernández and Rezac 2010; Mounole 2012; Odria 2012, 2014, in progress). Moreover, both are sensitive to some of the same factors, like animacy and specificity. However, we shall also see important differences: for instance Spanish DOM is never conditioned by tense, and Basque DOM never seems available for inanimates.3 2.2  Hindi-Urdu DOM Hindi-Urdu DOM likewise assimilates the coding of transitive objects to ditransitive indirect objects, by using the suffix -ko for both, and animacy and specificity also play a similar though not identical role (Masica 1982; Butt 1993; Mohannan 1995; Bhatt 2007).4 (5) Mina tum-*(ko)/Tina-*(ko) dekh rahii thii Mina.f you-ko/Tina-ko see prog.f be.pst.fsg ‘Mina was looking at you/Tina.’

Unlike in Spanish, Hindi-Urdu objects can retain their ko-marking in passives (6a,b), like indirect objects (Mohanan 1994; Bhatt 2007). (6) a. Ram-ne is t�ehnii-ko kal kaat�-aa thaa Ram-erg this branch.f-ko yesterday cut-pf.msg be.pst.msg ‘Ram had cut this branch yesterday’ b. is t�ehnii-ko kal kaat�-aa this.obl branch.f-ko yesterday cut-pf.msg gayaa thaa pass.pf.dflt be.pst.msg

‘The branch was cut yesterday’

Bhatt (2007) studies the syntax of Hindi-Urdu DOM. He demonstrates that komarked objects are structurally higher than unmarked ones for c-command into

.  Spanish and Basque also share dative marking of otherwise accusative objects in impersonal clauses (Mendikoetxea 1999; Ortiz de Urbina 2003c). In both languages, the phenomenon is independent of DOM, being found in varieties where there is no DOM. Provisionally, we set it aside. .  Hindi-Urdu glosses are prog progressive, pf perfective, pst past, dflt default, m/f masculine/feminine, sg singular.

 Beatriz Fernández & Milan Rezac

temporal adjuncts, and consequently they are linearized between the subject and ko-marked indirect object rather than after the latter. (7) Ram-ne chitthii-koi Anita-ko ti bhej-aa Ram-erg letter-ko Anita-ko send-pf ‘Ram sent the letter to Anita’

Bhatt concludes that Hindi-Urdu ko-marked objects undergo object shift with respect to unmarked ones, converging with the analysis of Spanish DOM above. 2.3  The syntax of Basque DOM Three chief options seem available for the analysis of the syntax of DOM with respect of that of identically marked indirect objects: a. DOM objects have the syntax of indirect objects, distinct from that of canonical direct objects. This explains the morphological identity of the former two against the latter, and predicts the same grouping for syntactic phenomena like (Bleam 2003 for leísmo, contrast Ormazabal and Romero 2013, and Rezac 2006 for Basque). b. DOM objects are just direct objects and the coding difference is due to morphology alone (cf. Ormazabal and Romero 2007 on leísmo as spellout of animacy). c. DOM objects are generated as direct objects, not as indirect objects, but acquire distinctive syntactic properties, for instance by object shift, which may be sui generis or which may partly assimilate them to indirect objects (see Jelinek eta Diesing 1995 generally, Bhatt 2007 for Hindi-Urdu DOM, T ­ orrego 1998: Chapter 2 and Ormazabal and Romero 2013 for Spanish DOM). In what follows, we examine the syntax of Basque DOM. Though it proves difficult to obtain a fine resolution on its syntactic behavior, it seems clear that assimilation of the syntax of DOM dative objects to that of indirect objects is to be excluded, and that DOM dative objects are substantially akin to absolutive direct objects, with certain properties suggesting a difference such as might be given by object shift. We begin by describing Basque DOM, its conditions, and their parameters in Section 3, and delve into its syntax in Section 4. 3.  Differential Object Marking in Basque 3.1  Canonical and Differential Object Marking Basque is a morphologically ergative and syntactically accusative language. Finite structures indicate the presence or absence of agreement with each of absolutive,



Differential object marking in Basque varieties 

ergative, and dative arguments by agreement morphology, the ‘dative flag’ indicating the presence of dative agreement, and in analytic constructions typical of the language also by choice of the auxiliary root.5 In this section we set out how case and agreement reflect canonical and DOM object marking. In Standard Basque and in most varieties of Basque, the form of a sentence with a bivalent transitive like see is shown in (8):6 (8) Nik zu ikusi zaitut I.erg you.abs see trn (2sgabs-1sgerg) ‘I saw you’

The subject is marked by the ergative case, -k, and the object by absolutive, -Ø. Correspondingly, the auxiliary is one proper to ergative-absolutive agreement combinations, the transitive *edun indicated by the root -u-, rather than intransitive or ditransitive. This auxiliary cross-references the ergative and absolutive arguments through ergative and absolutive agreement markers: the suffix -t for 1st person singular ergative and the prefix z- for 2nd person absolutive respectively.7 The corresponding sentence in the dialect of Lekeitio, a DOM variety of Basque, is in (9): (9) (Nik) suri ikusi dotzut    I.erg you.dat see dtrn (2sgdat-1sgerg) ‘I saw you’ (Hualde, Elordieta and Elordieta 1994: 125–7)

Compared to the canonical object coding in (8), (9) presents two differences. (i) The object su-ri ‘to you’ is marked by dative case -(r)i, not by absolutive case su-Ø. (ii) Along with the dative case, the auxiliary is ditransitive, indicated by -ts-, and agrees with the dative object by dative agreement, the 2nd person singular suffix -zu. Both case and agreement are identical to those of dative indirect objects of

.  In English, overviews of Basque agreement morphology may be found in Laka (1993), Trask (1995), Hualde (2003), and with focus on the dative flag and on root allomorphy, Trask (1981), Albizu (2002), Fernández (2013, 2015) (in French, Rebuschi 1984). .  For Basque, glosses are: erg, abs and dat for ergative, absolutive and dative cases ­respectively; fut future and pst past, d determiner; intrn, trn and dtrn auxiliaries with absolutive, absolutive-ergative, and absolutive-dative-ergative agreement respectively; their agreement is glossed in the manner of 2sgerg or 2sge for 2nd person singular ergative. We follow Etxepare’s (2003) terminology of monovalent (unaccusative, unergative), bivalent (unaccusative, unergative, transitive) and trivalent (transitive) for predicates. Dialectal phonology is kept, such as su ‘you’ in (9) rather than standard zu. .  In Basque, the 2nd person singular is morphologically plural, due to its origin as 2nd person plural, while for 2nd person plural, an additional pluralizer is added (see Hualde 2003: 206).

 Beatriz Fernández & Milan Rezac

trivalent predicates like give, as in (10a), also from Lekeitio, which corresponds to (10b) in standard Basque, with the same morphological analysis: (10) a. (Nik) suri liburua emon dotzut    I.erg you.dat book.d.abs give dtrn (2sgdat-1sgerg) ‘I gave you a book’ b. (Nik) zuri liburua eman dizut    I.erg you.dat book.d.abs give dtrn (2sgdat-1sgerg) ‘I gave you a book’

In Lekeitio, 2nd person singular objects participate in DOM only optionally, and the canonical coding is also available, with an absolutive direct object and an absolutive-ergative transitive auxiliary, with the same morphological analysis as Standard Basque (8). (11) (Nik) su ikusi saittut    I.erg you.abs see trn (2sgabs-1sgerg) ‘I saw you’

(Elordieta, Arantzazu, p.c.)

Since DOM (9) and ditransitive (10a) have the same finite verbal form, dotzut, dialectology has interpreted DOM as the use of ditransitive forms, dotzut in (9, 10a) = S.B. dizut in (10b), instead of transitive ones, saittut in (11) = S.B. zaitut in (9); see Table 1. This is a recurrent description of DOM from Bonaparte (1869: 434) to Zuazo (2003: 109). Descriptively, it has two gaps. One is the omission of case morphology. In DOM, the ditransitive auxiliary with dative agreement for the object of bivalent transitives like see always goes together with dative case on this object. Second, in varieties where DOM is more extensive than in Lekeitio, such as Dima, it has become obligatory for 1st and 2nd person, and so the transitive auxiliary forms for 1st/2nd person absolutives disappear and there is no longer a contrast between transitive and ditransitive auxiliaries (unlike in 3rd person, since DOM never affects inanimates). Table 1. DOM varieties of Basque S.B

Lekeitio

Dima

Transitive auxiliary

zaitut

saittut

—

Ditransitive auxiliary

dizut

dotzut

dotzut

The foregoing examples use the analytic conjugation of the verb, which involves an auxiliary and a nonfinite form of the lexical verb. This is the typical formation of Basque finite clauses. Only a handful of verbs also have synthetic forms, where



Differential object marking in Basque varieties 

the lexical root itself carries agreement and finiteness morphology, and many are restricted to registers where DOM does not occur. However, nothing in principle prevents them from combining with DOM. The following example shows DOM with synthetic forms of eroan ‘bring’ and eduki ‘have’ in Basauri Basque (Arretxe 1993). As in the analytic conjugation, the verb has dative agreement with the object, 1st person singular t(a), preceded by the dative flag -(t)s- that indicates the presence of dative agreement (here the object is pro-dropped, but we shall see overt dative objects with eduki later). (12) a. Etzera daroste house.d.to bring (1sgdat-3plerg) ‘They brought me home’ b. Emen dekostasu here have (1sgdat-2sgerg) ‘You have me here’

(Arretxe 1993: 234)

Different terms have been used in Basque linguistics to refer to DOM, including leísmo adopting the name of the similar phenomenon in Spanish (Rezac 2006), dative overmarking (Austin 2006), and quirky (dative) objects which highlights the combination of direct objecthood with an unexpected case (Fernández 2008; Fernández and Rezac 2010). Here, we use Differential Object Marking (also in Fernández and Rezac 2010 and Odria 2012, 2014, in progress), to place the phenomenon in the wider context of superficially similar phenomena in the typological literature. Before proceeding, it is important to highlight and contextualize the observation above that DOM affects both case and agreement, making the object dative for both. Other phenomena involving agreement in Basque that have been the focus of syntactic inquiry affect agreement only, leading to case-agreement mismatches. Particularly relevant is dative displacement, which seems the mirror image of DOM, but does not affect case (Fernández 2001, 2004; Fernández and Ezeizabarrena 2001; Rezac 2006, 2008a, and Rezac and Fernández 2013, among others). In dative displacement, the indirect object of a ditransitive retains dative case but noncanonically controls absolutive agreement. Compare standard (13a), repeating earlier (10b) to dative displacement in (13b) with the same object case but the auxiliary form of transitive (8): (13) a. (Nik) zuri liburua eman dizut    I.erg you.dat book.d.abs give dtrn (2sgdat-1sgerg) ‘I gave you a book.’ b. (Nik) zuri liburua eman zaitut    I.erg you.dat book.d.abs give trn (2sgabs-1sgerg) ‘I gave you a book’

 Beatriz Fernández & Milan Rezac

Basque dialectology mentioning DOM has focused on auxiliary forms, and thus treats DOM and dative displacement as the two sides of the same coin: on the one side, DOM, i.e. the substitution of transitive auxiliary forms by ditransitive ones and on the other dative displacement, with the opposite substitution (e.g. Arretxe 1994: 227). This perspective needs changing, for DOM is a matter of both agreement and case.8 3.2  Dative objects in alternating verbs The case and agreement pattern of Basque DOM is identical to that of dative objects of bivalent unergatives (Fernández and Ortiz de Urbina 2010, 2012; Ortiz de Urbina and Fernández this volume). An example is bultzatu ‘push’ in (14a), compared to DOM of ikusi ‘see’ in (14b). (14) a. (Nik) suri bultzatu dotzut    I.erg you.dat push dtrn (2sgdat-1sgerg) ‘I pushed you’ b. (Nik) suri ikusi dotzut    I.erg you.dat see dtrn (2sgdat-1sgerg) ‘I saw you’

For verbs like bultzatu ‘push’, the ergative-dative frame is available outside DOM varieties, as in standard Basque (15a), and alternates in them with the ergativeabsolutive frame (15b), without conditions on DOM like animacy (15c). (15) a. (Nik) zuri bultzatu dizut    I.erg you.dat push dtrn (2sgdat-1sgerg) ‘I pushed you.’ b. (Nik) zu bultzatu zaitut    I.erg you.abs push trn (2sgabs-1sgerg) ‘I pushed you.’ c. (Nik) mahaiari bultzatu diot    I.erg table.d.dat push dtrn (3sgdat-1sgerg) ‘I pushed the table.’

.  In many works, DOM is a highly stigmatized phenomenon, typically described as a confusion of transitive and ditransitive agreement patterns and attributed to external influence or language loss: see for instance, Yrizar (1981-II: 360) under the title of Observaciones referentes al empleo incorrecto de algunas flexiones, Bonaparte (1869), and more recently Aurrekoetxea and Txillardegi (1983: 49). This is not only an academic but also a public perception, of which DOM speakers are very well aware (see Austin 2006, Fernández and Rezac 2010 and references therein). Strikingly, Spanish leísmo, similar to Basque DOM, is sometimes judged even in literature as sign of elegance and prestige, as pointed out by Fernández-Ordoñez (1999: 1386–1388).



Differential object marking in Basque varieties 

This alternation in object marking and agreement morphology leads to these verbs being called Alternating verbs; see Etxepare (2003) and Fernández and Ortiz de Urbina (2010) for an exhaustive description. They make for an excellent point of comparison with DOM transitives: both are bivalent, both have an ergative external argument, in both the internal argument is sometimes absolutive and sometimes dative in case and agreement, although under different conditions. The dative internal argument of verbs like bultzatu ‘push’ turns out pattern with the dative indirect object of trivalent ditransitives like eman ‘give’, leading to the conclusion that in the ergative-dative frame, alternating verbs are bivalent unergatives: the ergative is the external argument, the dative is an indirect object, and there is no other internal argument to surface as absolutive (Fernández and Ortiz de Urbina 2012; Ortiz de Urbina and Fernández this volume). By contrast, the dative internal argument in DOM will prove to behave differently, like the canonical absolutive direct object of bivalent transitives. 3.3  Conditions and variation of DOM DOM is governed by two types of factors. One is properties of the object that participates in it: animacy, person, and definiteness/specificity. These have been richly documented for DOM in other languages (Bossong 1991; Aissen 2003). The other is properties of the clause like tense and finiteness. These have been observed for Basque DOM (Rezac 2006; Fernández and Rezac 2010), but far more rarely elsewhere, playing no role -as far we know- for instance in Spanish (Ormazabal and Romero 2013) or Hindi-Urdu (Mohanan 1994), or else deriving from different alignments of the tenses involved, as in Iranian languages (Haig 2008). 3.3.1  Animacy DOM in Basque and crosslinguistically is sensitive to the animacy of the object. The point in the Animacy Hierarchy of Silverstein (1976) where DOM occurs varies across Basque varieties. In Lekeitio Basque (Hualde, Elordieta and Elordieta 1994: 125ff.; Fernández 2008; Mounole 2012), the crucial factor is the human/nonhuman distinction: only human objects accept DOM. The examples in (16) show DOM instead of canonical absolutive for animate objects. DOM is always optional in Lekeitio Basque, so that beside DOM the canonical absolutive pattern is also available.9

.  Independently of DOM, first person datives undergo so-called dative displacement whereby they control absolutive agreement (Hualde, Elordieta and Elordieta 1994: 124–7; on dative displacement, see Rezac and Fernández 2013).

 Beatriz Fernández & Milan Rezac

(16) √ abs case/agreement/trn   √ dat case/agreement/dtrn a. (Nik) su ikusi saittut   a’. (Nik) suri ikusi dotzut    I.erg you.abs see trn    I.erg you.dat see dtrn ‘I saw you’ (2nd person pronoun) b. Peruk Jon ikusi dau   b’. Peruk Joneri ikusi dotzo Peru.erg Jon.abs see trn Peru.erg Jon.dat see dtrn ‘Peru saw Jon’ (human object, proper noun) c. (Nik) neskia ikusi dot    c’. (Nik) neskiari ikusi dotzat    I.erg girl.d.abs see trn   I.erg girl.d.dat see dtrn ‘I saw the girl’ (human object, common noun)

The examples in (17) show non-human objects, animate and inanimate, for which DOM is unavailable: (17) √ abs case/agreement/trn        *dat case/agreement/dtrn a. (Nik) txakurra ikusi dot      a’. *(Nik) txakurrari ikusi dotzat    I.erg dog.d.abs see trn      I.erg dog.d.dat see dtrn ‘I saw a dog’ (non-human/animate object) b. (Nik) telebisiñoia ikusi dot     b’.   *(Nik) telebisiñoari ikusi dotzat    I.erg TV.d.abs see trn      I.erg TV.d.dat see dtrn ‘I watched TV’ (non-human/non-animate object)

This distribution of DOM is replicated in the data gathered by Mounole (2012) in Tolosa (Central Basque), by Arraztio (2010) in Araitz-Betelu (Central Basque, oriental variety), and by Odria (2012, 2014, in progress) in Elgoibar (Western Basque, transitional variety).10 Some exceptional DOM of non-human animates is attested, such as Hurtado’s (2001: 104), gizon batek joyo zakurrai ‘A man hit a dog’ where zakurrai ‘a dog’ is marked by dative and not by absolutive. This and similar examples may be rare and reflect idiolectal variation, as suggested in Arraztio (2010). No DOM of inanimates has been reported for any Basque variety, which play an important role in Odria’s (2012) analysis. In particular, Basque DOM does not seem to extend to inanimates in the way DOM does in Spanish (Torrego 1998: 2.7.2.1; Ormazabal and Romero 2013) or Hindi-Urdu (Mohanan 1994; Bhatt 2007), which have been illustrated above (Ane Odria, p.c.). Animacy alone may be insufficient for some more nuanced patterns. While in Lekeitio or Tolosa, DOM is found with all transitives, Austin (2006: 141–142)

.  We follow Zuazo’s (2003) dialectal designations, using central, oriental and occidental for erdigunekoa, sortaldekoa and sartaldekoa respectively, and transitional varieties for tarteko hizkerak.



Differential object marking in Basque varieties 

points out that DOM is prefered with verbs such as jo ‘hit’ or molestatu ‘bother’, as attested in her study of a spoken corpus. This hits at a role for agentivity of the external argument and affectedness of the internal argument, discussed for ­Spanish a-marking by Torrego (1998: Chapter 2).11 3.3.2  Person Closely related to animacy, person seems to play a significant role in Basque DOM, as likewise cross-linguistically. In Lekeitio Basque, DOM is available and optional for humans of any person. In Arratia Basque, DOM is obligatory for 1st and 2nd person objects, but unavailable for 3rd person even if human: (18) a. (Zuk) (neri) ikusi dostesu  you.erg    I.dat see dtrn (1sgdat-1sgerg) ‘You saw me’ b. (Nik) (suri) ikusi dotzut    I.erg    you.dat see dtrn (2sgdat-1sgerg) ‘I saw you’ c. (Nik) Jon/*Joneri ikusi dot / *dotzat    I.erg Jon.abs/*dat see dtrn /dtrn (3sgdat-1sgerg) ‘I saw Jon’

This role of person is a traditional observation, confirmed for Dima Basque by Mounole (2012) based on data from Iglesias (2005) as well as in our fieldwork. Even in varieties where 3rd person objects can participate in DOM, it is more frequent with 1st and 2nd person than with 3rd (Hualde, Elordieta and Elordieta 1994: 125–127 for Lekeitio Basque and Odria 2012 for Elgoibar Basque). Elsewhere the contrast appears as obligatory DOM in 1st/2nd person versus optional DOM with 3rd person, as in Ultzama Basque (Navarrese, central variety) (Ibarra 1995: 427), or at least admitting of exceptions in Erroibar and Esteribar Basque (Navarrese, oriental varieties) (Ibarra 2000: 152–3) (see further Fernández and Rezac 2010 and references therein).12

.  In emerging phenomena, frequency is a factor (Bybee 2010), but all these are basic, common verbs of the colloquial register. .  Descriptions of paradigms suggest the possibility for far more nuanced fine-tuning of DOM, for instance limiting it to 2nd person singular or even to the presence of a particular ergative (Rezac 2006). This may reflect incomplete grammaticalization and multiple grammars, with not all forms being necessarily captured in any given investigation. Detailed work remains to be done to establish even the most basic of these distinctions. Distinct is sensitivity of DOM to whether or not the allocutive paradigm is being used, which may reflect the role of register.

 Beatriz Fernández & Milan Rezac

When all of 1st/2nd person is obligatory affected by DOM, the ergativeabsolutive agreement paradigm disappears for these persons (though as discussed below, their absolutive case forms may be preserved in nonagreeing contexts). This leads to partial collapse of the ergative-absolutive and ergative-dative-absolutive agreement paradigms. The collapse never affects agreement forms for 3rd person objects, as these always retain absolutive forms at least when inanimate.13 3.3.3  Definiteness and/or specificity Finally among properties of objects, Basque DOM as DOM crosslinguistically is affected by the object’s referentiality. As Mounole (2012) first observes, definite but not indefinite (human) objects participate in DOM in Lekeitio Basque: (19) a. *Eztotzat ezaututen iñori  not.dtrn (3sgdat-1sgerg) knowing anybody.dat ‘I don’t know anybody’ b. *Morroi bateri ikusi dotzat  guy one.dat see dtrn (3sgdat-1sgerg) ‘I saw a guy’

Mounole also provides data from Tolosa Basque, where indefinites and reciprocals are barred from DOM: (20) a. *Nik ez diot iñorrei ikusi    I.erg not dtrn (3sgdat-1sgerg) anybody.dat see ‘I didn’t see anybody’ b. *Jonek neska askori ikusi dio    Jon.erg girl.abs many.dat see dtrn (3sgdat-3sgerg) ‘Jon saw many girls’ c. *Elkarri ikusi diote    each other.dat dtrn (3sgdat-3sgerg) ‘They saw each other’ .  An interesting phenomenon is the interaction of DOM with another agreement phenomenon, dative displacement, where dative objects are coded by absolutive agreement, though keeping their dative case (see Rezac and Fernández 2013 and references therein). When a dialect has both phenomena, DOM may occasionally feed dative displacement, producing opaque forms with dative objects but absolutive agreement, as in Basauri and Pasaia. More common is the tendency for the two to apply to distinct parts of the paradigm, DOM originating in 2nd person of past tense and dative displacement in 1st person of the present tense, up to the creation of a single agreement paradigm, some objects of which use original dative agreement markers (2nd person past), others absolutive agrement markers (1st person present), as in ­Hondarribia (Sagarzazu 2005) (Rezac 2006; Fernández and Rezac 2010; Rezac and Fernández 2013).



Differential object marking in Basque varieties 

In our work, the reflexive anaphora bere buru ‘herself/himself ’ is likewise excluded from DOM in Dima. (21) a. Lurrek bere burue ikusi dau ispiluen Lur.erg her head.abs see trn (3sgabs-3sgerg) mirror.d.in ‘Lur saw herself in the mirror’ b. *Lurrek bere buruari ikusi dotza ispiluen     Lur.erg her head.d.dat see dtrn (3sgdat-3sgerg) mirror.d.in ‘Lur saw herself in the mirror’

Similar restrictions have also been noted for Araitz-Betelu (Arraztio 2010). The specific factors and range of variation remain to be studied, addressing such questions as the semantic characterization of the referentiality involved and its relationship to the morphological marking of definite and presupposed DPs in Basque (for Spanish, see Torrego 1998: Chapter 2, Gutierrez-Rexach 2000, and Ormazabal and Romero 2013). 3.3.4  Properties of the clause: Tense, finiteness, agreement So DOM has been conditioned by properties of the object. Properties of the clause matter as well: tense and finiteness/agreement. These do not usually bear on DOM in other languages, and thus are particularly significant for analysis of Basque DOM. Descriptions of verbal paradigms indicates that in some varieties DOM is restricted to the past tense, as in Azpilkueta zaiztet vs. natzen (Yrizar 1997:  716–750), while in others it covers more ground in the past, as in Hondarribia and Irun (Sagarzazu 2005: 82) (see Rezac 2006). Our investigation has confirmed the role of tense for some speakers in Araitz-Betelu Basque thanks to data ­gathered by Arraztio (2011). For one speaker, whereas DOM is optional in the present, it is obligatory in the past. (22) a. Nik zu ikusi zattut I.erg you.abs see trn (2sga-1sge) a’. Nik zui ikusi dizut I.erg you.dat see dtrn (2sgd-1sge) ‘I have seen you’ b. *Nik zu ikusi zintudan    I.erg you.abs see trn.pst (2sga-1sge) b’. Nik zui ikusi nizun I.erg you.dat see dtr.pst (2sgd-1sge) ‘I saw you’

 Beatriz Fernández & Milan Rezac

In other varieties like Lekeitio Basque, DOM is found in both present and past: (23) a. Su ikusi saittuten. you.abs see trn.pst (2sgabs-1sgerg) Suri ikusi neutzun you.dat see dtrn.pst (2sgdat-1sgerg) ‘I saw you’ b. Koldok dzo saittun/eutzun Koldo.erg hit trn.pst (2sga-1sge)/dtrn.pst (2sgd-1sge) ‘Koldo hit you’ c. Peruk ikusi eban/eutzan Peru.erg see trn.pst (3sga-3sge)/dtrn.pst (3sgd-3sge) ‘Peru saw him/her’

In addition to tense, some varieties seem to condition DOM according to whether a clause is finite/agreeing or nonfinite/nonagreeing: DOM may be reduced in the latter with respect to the former. In Dima, DOM with 1st/2nd person is obligatory for agreeing objects in both plain finite and restructurnig configurations like (24a), but optional for the nonagreeing objects of nonfinite clauses, (24b). (24) a. Seuri eroan gure dotzut you.dat carry want dtrn (2sgdat-2sgerg) ‘I want to bring you’ b. Seu/seuri ikusten etorri nes you.abs/dat seeing come intrn (1sgabs) ‘I am coming to see you’

Other varities have no such condition, at least for some speakers, as Irun, ­Errenteria and Hondarribia Basque (central Basque, transitional varieties). The matter remains to be better studied, but provisionally, it seems that finiteness/ agreement affect DOM. 3.4  Dialectal distribution Basque DOM is attested throughout the Basque-speaking area, save for eastern Basque (see Yrizar 1981-II: 359ff. for a general overview, also Fernández and Rezac 2010; for sources, see the Appendix): i.

Navarrese: DOM is widespread (Zuazo 1998: 18), e.g. Bortzerriak (­occidental), Sakana (south-occidental), Ultzama (central), Esteribar and Erroibar (oriental), and also in Aezkoa and Baztan (transitional varieties), with only occidental Navarrese varieties (also) keeping the canonical absolutive marking of the object (i.e. Bortzerriak, Sunbila, Bertizarana, Malerreka, Basaburua Ttikia and Araitz).



Differential object marking in Basque varieties 

ii. Western Basque: Basauri, Bermeo, Igorre (occidental), Forua (Busturialdea), Lekeitio (oriental) and Elgoibar (transitional varieties). iii. Central Basque: Tolosa, Ordizia and Goierri, Lasarte-Oria; Pasaia, Oiartzun, Hondarribia and Irun (transitional varieties); also oriental Imotz, Basaburua Nagusia and Larrau. There is no evidence for DOM in descriptions of several contemporary central varieties, including Antzuola, Arrasate, Bergara, Eibar, Ermua/Eitza, Leioa, Orio, Otxandio, Sopela and Zegama, with descriptions of some like Antzuola explicitly noting its absence. 4.  On the syntactic nature of Basque DOM objects In this section, we examine the syntax of DOM dative objects in the light of absolutive direct objects and of dative indirect objects, building on the work of Fernández and Rezac (2010) and Odria (2012). Our results indicate that DOM dative objects are direct rather than indirect objects configurationally, bear structural rather than inherent Case, and alternate with absolutive direct objects in such a way as to suggest a single underlying mechanism for both absolutive and DOM dative, sensitive to properties of both the probe and the goal in its outcome. We present the following findings: i. Secondary predication: DOM dative objects license secondary predicates like absolutive direct objects but unlike dative indirect objects, indicating that DOM dative objects use a configuration and interpretation similar to that of direct and not indirect objects, in particular not applicative and prepositional structures. ii. Structural Case: DOM can be used in Exceptional Case Marking, where the DOM dative bears no selectional relationships to the clause that assigns it and agrees with it. Consequently, the DOM dative is not an inherent Case but a structural one, like the absolutive of direct objects. iii. Concomitantly, DOM dative objects mostly require agreement like absolutive objects, suggesting the need to Agree with the clause for Case licensing. iv. DOM objects may by conditioned by clausal properties like tense, again contrasting with inherent Case, and indicating that the DOM dative must be able to take into account properties between v and Fin, ceding to the absolutive when its conditions are not met. In light of these results, we propose that absolutive and DOM dative objects occur in the same structure and participate in the same Agree/Case relation as absolutive direct objects, sensitive to the features of goal and probe which modulate the identity of the outcome as absolutive or dative, possibly reflecting a difference like object shift.

 Beatriz Fernández & Milan Rezac

4.1  DOM objects and depictives Depictive secondary predicate licensing suggests that DOM objects occur in a configuration relevantly identical to that of absolutive direct objects, and at any rate distinct from that of dative indirect objects. In systems with ‘low applicative’ indirect objects, depictive secondary predicates can be controlled by the direct but not the indirect object (Pylkkänen 2002): such are English (Pylkkänen 2002), Spanish and French (Zubizarreta 1995), ­German (McFadden 2003, 2004), and Basque (Zabala 2003, Arregi and Molina-Azaola 2004, Oyharçabal 2007).14 This is illustrated in (25a) versus (25b) for standard Basque, but it obtains in DOM varieties like Dima, Elgoibar, and Lekeitio Basque: (25) a. Jonek haragiai gordiniki jan zuen Jon.erg meat.d.abs raw eat trn.pst (3sgabs-3sgerg) ‘Jon ate the meat raw’ b. *Joneki Joanarij berriak mozkori/*j    Jon.erg Joana.dat news.d.abs drunk eman zizkion give dtrn.pst (3plabs-3sgdat-3sgerg)

‘*Jon gave Joana the news drunk’

Oyharçabal (2007)

The dative internal argument of alternating verbs like begiratu ‘look at’ patterns with dative indirect objects in not licensing depictives (Fernández and Rezac 2010). Here we illustrate this from Elgoibar Basque (Odria 2012: 22); the same obtains in Dima Basque. Spanish behaves in the same way (Zubizarreta 1985: 251). (26) a. Niki Mirenij poziki/*j begiratu nion I.erg Miren.dat happy look.at dtrn.pst (3sgdat-1sgerg) ‘I looked at Miren drunk’ b. Niki Mirenij oinutsiki/*j jarraitxu nion I.erg Miren.dat barefoot follow dtrn.pst (3sgd-1sge) ‘I followed Miren barefoot’

In these varieties, DOM dative objects pattern with absolutive direct objects in licensing depictives (Fernández and Rezac 2010; Odria 2012). This is so in Dima Basque (only for 1st/2nd person objects like (26a), for in Dima DOM is limited

.  The generalization includes indirect objects of all types (goals, possessors, experiencers) but only in structures with a higher thematic/promoted subject, thus excluding e.g. the experiencers of unaccusatives or indirect objects promoted to passives, and only in simple applicative constructions, thus excluding causatives and light verb constructions. For derivation of these conditions, see Pylkkänen (2008).



Differential object marking in Basque varieties 

to them), and in Elgoibar Basque whence the following data (Odria 2012: 22, with DOM optional). The same holds of Spanish for both the a-marking and leísmo DOM (Odria 2012).15 (27) a. Niki zurij mozkortutai/j ikusi dizut I.erg you.dat drunk see dtrn (2sgdat-1sgerg) ‘I saw you drunk’ b. Niki umiarij oinutsiki/j ekarri diot I.erg kid.d.dat barefoot carry dtrn (3sgdat-1sgerg) ‘I carried the kid barefoot’

This gives us a minimal contrast between two bivalent structures with an external argument ergative and internal argument. In one, alternating predicates, the internal argument behaves like a dative indirect object in not licensing depictives, confirming their analysis as bivalent unergatives with an indirect object (­Fernández and Ortiz de Urbina 2010, 2012; Ortiz de Urbina and Fernández this ­volume). In the other, the internal argument is a DOM dative object, and it behaves like a canonical absolutive direct object in licensing depictives. Thus, DOM dative objects pattern with absolutive direct objects and against dative indirect objects in licensing depictives. The interpretation of this finding depends on the theory of depictive licensing, which needs to single out external arguments and direct objects against indirect objects of various kinds, among other contrasts. On Pylkkänen’s (2008) proposal, licensing depends on the possibility of conjoining the main and depictive predicates and sharing the subject of the main predicate as subject of the two. This in turn depends on both interpretation and phrase-structural configuration. The interpretation of low applicative heads and prepositions is such that a depictive cannot combine with them, while that of transitive roots and v is such that it can, as is that of predicates derived by movement. DOM objects belong in the latter rather than former group. They are not, at any rate, low applicative or prepositional arguments, as indirect objects are. They could be in the same configuration as absolutive direct objects, but also differ from it by further A-movement like object shift. We cannot go securely beyond this at present, for we do not have to hand for Basque diagnostics like Bhatt’s (2007) adjunct test that shows Hindi-Urdu DOM to involve object shift. Adjunct subjects in Basque can be anteceded by absolutive and DOM dative objects alike, as shown below, as well as subjects and indirect objects (see Ortiz de Urbina 1989 and Duguine 2012 on whether they use PRO or pro).

.  Thus Juan encontró a Maríai borrachai or leísmo, Juan lei encontró borrachai, beside *Juan le habló a Maríai borrachai (Odria 2014, in progress, with references).

 Beatriz Fernández & Milan Rezac

(27) Nok ikusi dotzu /zaitxu who.erg see dtrn (2sgdat-1sgerg) /trn (2sgabs-1sgerg) urteten/urteterakuan? leaving/leaving.upon ‘Who saw youi while/upon PROi leaving?’  (Lekeitio Basque, Arantzazu Elordieta p.c.)

There is indirect evidence for the object shift analysis of Basque DOM: its sensitivity to animacy and specificity, insofar as this can be derived from object shift in general (Diesing and Jelinek 1995), and goes together with clear object movement in Hindi-Urdu DOM (Bhatt 2007).16 It bears to end on two caveats, one minor and one significant, both raised in Fernández and Rezac 2010). The depictive diagnostic seems robust for the speakers/varieties cited. However, some speakers do allow depictives to be controlled by the dative internal argument of bivalent unergatives of the begiratu ‘look at’ type, and others seem to do so even for the dative indirect objects of ditransitives like eman. These qualifications are familiar in the literature even for English (Himmelman and Schultze-Berndt 2005: 55). For such speakers the diagnostic cannot be used. More troubling is the situation with Lekeitio DOM. In Lekeitio, DOM is optional with 3rd person animates, as in Elgoibar, yet whereas in Elgoibar (20b) is grammatical with DOM, in Lekeitio it is not and the canonical absolutive is used instead, Nik umiai ortosiki ekarri dot ‘I carried the kid barefoot’. This could be a quirk in the data; but nothing in our work ensures that Basque DOM is one and the same phenomenon in all varieties. 4.2  ECM: DOM in transitive predication with eduki ‘have’ In this section, we extend this parallelism between DOM dative objects and nonDOM absolutive direct objects through Exceptional Case Marking. It leads us to conclude that DOM objects receive structural rather than inherent Case, like the absolutive of direct objects. We assume the distinction between structural and inherent Case of Case ­Theory in Chomsky (1986, 2000, 2001), reviewed with respect to Basque in Rezac, Albizu and Etxepare (2014). Inherent Case depends on the selectional relationship between a predicate, such as P, V, Appl, vAgentive and the argument it introduces. .  There remain significant uncertainties: Diesing and Jelinek’s (1995) proposal says nothing about animacy, though see Jelinek and Carnie (2003), while the Germanic languages do not bar reflexive/reciprocal pronouns from object shift, unlike Basque from DOM. It is also not clear whether Basque absolutives undergo DOM: see Vicente (2005) and Rezac, Albizu and Etxepare (2014).



Differential object marking in Basque varieties 

It is thus constrained by selection, notably to being phrase-structurally local and invariant under changes in higher function architecture. Structural Case, on the other hand, reflects the Agree relationship between an Agree/Case locus such as v and T and an argument that need bear no selectional relationship to the locus or the extended projection hosting it. It may occur under phrase-structural distance and change through changes in functional architecture alone. In English, an active clause can assign the structural accusative to the subject of a lower clause in ECM, We consider [him-ACC (to be) clever], and switch it to the structural nominative when passive in raising He-NOM is considered [__ (to be) clever]. On the other hand, the prepositional and Saxon genitive are inherent, barring Our consideration of him (*clever, *to be clever), His being considered (*clever, *to be clever) surprises you? Correspondingly, genitives need a selectional relationship, so there are no genitive expletives and idiom chunks beside nominative and accusative ones: The cat(*’s) seeming to be out of the bat surprises you? (Abney 1987). In Basque, the absolutive and ergative clearly come out as structural (Rezac, Albizu and Etxepare 2014). In this section, we look at the DOM dative in ECM constructions, where its bearer has no selectional relationship to the clause responsible for it. The DOM dative does seem available under ECM, identifying it structural Case, independent of selection.17 The Basque ECM construction we examine is the transitive counterpart of intransitive predication with the copula be. Intransitive predication with be involves a small clause subject-predicate structure, whose subject raises to become the s­ ubject of be (Stowell 1978, 1991, Couquaux 1981, Burzio 1986: 2.7, for Basque Zabala 2003). (28) Xabier mutil azkarra da Xabier.abs boy quick.d.abs be (3sgabs) ‘Xabier is a clever boy’ [Xabieri [SC [SUBJ ti] [PRED mutil azkarra]] da]

Basque has a transitive predication counterpart, studied by Rebuschi (1984), ­Etxepare (2003: 4.1.6.1.1–2), Etxepare and Uribe-Etxebarria (2012: sec. 6), and de Rijk (2008: 675–677) who gives it its name.

.  Aside from the ECM structures we study, Basque also has ECM with perception verbs, but it alternates with surface-identical adjunct control construction, and because DOM ­requires human objects we cannot employ tests like idiom chunks to ensure we are dealing with ECM (see Arteatx 2007, 2012; Rezac, Albizu and Etxepare 2014). Thus in Kantatzen entzun dizut ‘I have heard you.DAT singing’, you can be analysed as the theme of hear controlling into an adjunct gerund, rather than ECM.

 Beatriz Fernández & Milan Rezac

(29) a. Orain datozenak adiskideak ditugu now come.3pla.who.d.abs friends.d.abs have (3pla-1ple) ‘Those who are coming now are our friends’ (de Rijk 2008: 676) b. Nor zaitugu, ba? who.abs have (2plabs-1plerg) ‘Who are you, then?’

(de Rijk 2008: 676)

c. Xabier mutil azkarra duzu/dugu/dute Xabier.abs boy quick.d.abs have (3sgabs-2/1/3plerg) ‘Xabier is an intelligent boy, which benefits/interests you/us/them’  (Etxepare and Uribe-Etxebarria 2012: 323)

In transitive predication, the copula is have. This verb is ordinarily used for possession, with the ergative as possessor and absolutive as possessum. Thus if the predicate mutil azkarra is omitted in (29c), the meaning is ‘We/you/they have Xabier’. In transitive predication, however, have works as a copula: it relates the subject-predicate relation, Xabier-clever boy, to the ergative, whose interpretation recalls of that of applicative datives like experiencers in intransitive predication. There is no entailment to possession, from (29c) to ‘We/you/they have Xabier’. The result resembles English She still has her grandparents alive beside possessive Everyone has grandparents, but it lacks the restrictions that English, French, or Spanish impose on such structures. We adopt Etxepare and Uribe-Etxebarria’s (2012: sec. 6) analysis of transitive predication as ECM. It builds on the raising analysis of intransitive predication, and on the similarity between its applicative datives and the ergative. Thus We have Xabier a clever boy has the structure in (30): (i) a subject-predicate complement to the copula be, (ii) to which is added the applicative head Pexp introducing an experiencer, (iv) the experiencer ends up as ergative and be+P as have (following Kayne 1993). We assume the small clause + ECM core of the analysis, and return to Agree/Case relations of the small clause subject as we proceed.18 (30) BE [PP we [Pexp [SC [SUBJ Xabier] [PRED a clever boy]]]]

.  There are restrictions on transitive predication with *edun ‘have’ compared to intransitive predication with izan ‘be’: for instance both can be used with this – my bed but only izan with this – my book (Etxepare 2003: 415–6). These seem attributable to the relation that must hold between the predication and the ergative subject. The ECM + small clause analysis seems confirmed by the diagnostic for these structures developed in Moulton (2013): (i) Dirua eskatzeko, bi artikulu beharrezkoak ditugu, baina bat bera ere ez dugu ‘To ask for the money, we need (we have necessary) two articles, but we don’t have even one’ (necessary > two)



Differential object marking in Basque varieties 

The question before us is the following: when DOM is independently required, does transitive predication use DOM? Since there is no selectional relation between the matrix and the small clause subject, any case/agreement between them must be structural. If transitive predication is compatible with a DOM dative, the latter is structural; if it is not, a good explanation is that it is inherent, dependent on selection which is absent in ECM. In our examples so far, the transitive predication have verb has been *edun. This cannot be combined with DOM, nor can DOM be suspended when independently required, say for a 1st/2nd person object. In that case, the result is ineffable. However, Basque has another have verb, *eduki. Both *edun and eduki may be used for possession, *edun in eastern varieties, eduki in western and many central ones (Hualde 2003: 221; de Rijk 2008: 307), while some central varieties allow both with the same meaning (Etxepare and Uribe-Etxebarria 2012: sec. 6, R. Etxepare, p.c., in some varieties with nuances, Orreaga 2000: 4.16).19 In transitive predication, the situation is somewhat different. Where *edun is available, eduki does not participate in transitive predication of the sort seen above, e.g. Xabier is clever, a clever boy (Etxepare and Uribe Etxebarria 2012: sec. 6). Even in some varieties where eduki has replaced *edun for possession, *edun is kept precisely in transitive predication (Zuazo 1998: 221). Eduki does occur in a superficially similar use, illustrated in (31) (de Rijk 2008: 677, 679, 25.2.2, 25.5). It is then usually restricted to predicates like zain ‘waiting’, alboan ‘beside’ or eginda ‘done’: predicates that in intransitive predication use copula egon ‘be’ (locational or stage-level) rather than izan ‘be’ (individual-level), and that also occur as secondary predicates or predicate complements with verbs like jarri ‘put = become, turn (tr. and intr.)’, utzi ‘leave, be left’ (Zabala 2003; Artiagoitia 2012; Eguren 2012).20 (31) Bi gizon zeuzkan behean zain two men.abs had (3plabs-3sgerg) downstairs waiting ‘She had two men waiting for her downstairs’ (de Rijk 2008: 677)

Comparable structures are found in English, French, and Spanish, unlike the general transitive predication of *edun: She has friends alive, waiting, near her, at home but not *clever, *clever girls, *necessary, fine in Basque with *edun. The English structures have been analysed much in the same way as Basque transitive predication above: a small clause complement to a copula or empty have, which introduces the subject, which needs to bind a variable in the small clause, including possibly

.  The use of *edun and not eduki as the transitive auxiliary is universal in all varieties. .  The details of which be copula is used when vary with dialects: while izan is present in all dialects as copula, egon is absent from some eastern ones, and its use is more restricted as one moves from west, where it corresponds to Spanish estar, to east.

 Beatriz Fernández & Milan Rezac

silent possessor and experiencer (Sæbø 2009, Ritter and Rosen 1997). Under this analysis, eduki in (31) is transitive predication. If that is so, then we do find DOM in ECM, because DOM is compatible in transitive predication with eduki. This is the situation in Itsasondo Basque. Both eduki and *edun are used for possession have, with eduki more natural, (32a). Transitive predication with *edun is available and normal when the object is absolutive. DOM is required of 1st/2nd person objects, and *edun is then impossible, nor can DOM be suspended (save through switch to Standard Basque with its 1st/2nd person absolutives). However, eduki is available in the restricted transitive predication (32b) described above, excluding predicates like ‘clever’ or ‘be a shepherd (temporarily or permanently)’, (32c). (32) a. Ardi asko edukiko/izango dugu aurten sheep many.abs eduki.fut/be.fut trn (3pla-2ple) this year ‘We will have many sheep this year’ b. Alboan edukiko nauzu/didazu beside eduki.fut trn (2sga-1sge)/dtrn (1sgd-2sge) ‘I will always be beside you, which benefits/interests you’ c. Nevadan artzain edukiko nauzu/*didazu Nevada.in shepherd.abs eduki.fut trn (2sga-1sge)/dtrn (1sgd-2sge) ‘I will be shepherd in Nevada, which benefits/interests you’

Still, it is less clear in this case than it was for *edun that we have to do with ECM, because the predicates available are restricted in a way that is not well understood (but cf. Ritter and Rosen 1997) and similar restrictions on ­secondary predication contribute to the contentiousness of its analysis as raising/ECM (Beaver 2011). This issue can be sidestepped in varieties where eduki spreads onto the terrain of transitive predication elsewhere held by *edun. This is so in Dima Basque. In Dima, DOM is obligatory for 1st/2nd person objects, which are thus dative, and unavailable for 3rd, which are absolutive. Transitive predication is available with *edun for absolutive but not DOM dative objects, thus for 3rd but not 1st/2nd persons: (33) Oier artzain dek(og)u Nevadan Oier.abs shepherd.abs eduki (3sgabs-1plerg) Nevada.in ‘Oier is shepherd in Nevada which benefits/interests us’

Transitive predication with eduki is available for dative 1st/2nd persons under DOM. As in Itsasondo, it includes predicates like beside (34a), which uses the location/stage-level copula egon (34b) in intransitive predication: (34) a. Ondoan edukiko dostezu beti beside eduki.fut dtrn (1sgdat-2sgerg) always ‘I will always be beside you, which benefits/interests you’



Differential object marking in Basque varieties 

b. Beti egongo naz zure ondoan always be.fut intr (1sgabs) your beside ‘I will always be beside you.’ (consultant paraphrase)

However, it extends to predicates like (a) nationalist, which use the individuallevel copula izan, and are impossible with eduki in the varieties considered earlier (U. Etxeberria, p.c. for Itsasondo; R. Etxepare, p.c.): (35) a. Abertzalea edukiko dostezu beti nationalist.d.abs eduki.fut dtrn (1sgdat-2sgerg) always ‘I will always be a nationalist, which benefits/interests you’ b. Beti izango naz abertzalea always be.fut intr (1sgabs) nationalist.d.abs ‘I will always be a nationalist’ (consultant paraphrase)

Likewise contrasting with Itsasondo is eduki with ‘(be) a shepherd’: DOM occurs with 1st/2nd person and not with with 3rd person: (36) a. Ne(r)i artzain dekostesu Nevadan I.dat shepherd.abs eduki (1sgdat-2sgerg) Nevada.in ‘I am shepherd in Nevada, which benefits/interests you’ b. *Oierrei artzain dekotsagu Nevadan    Oier.dat shepherd.abs eduki (3sgdat-1plerg) Nevada.in ‘Oier is shepherd in Nevada, which benefits/interests us’

So in Dima, DOM dative objects occur in transitive predication as an ECM structure with eduki. We conclude that the DOM dative is not inherent Case dependent on selection between assigner and bearer, but structural Agree/Case, like the absolutive. The above examples illustrate another property where Dima differs from Itsasondo: the availability of synthetic forms with dative agreement for eduki. As mentioned in Section 3, the agreement complex of Basque finite clauses normally uses an auxiliary, which combines with a nonfinite form of the lexical verb. For a handful of verbs, synthetic verb forms exist for certain tenses, where agreement and finiteness is borne by the lexical verb itself. *edun and eduki ‘have’ are the most common synthetic transitives, but in most varieties only *edun can bear dative agreement. Dima belongs to a western group that does have synthetic forms with dative agreement for eduki, seen above. Like Dima seems to be Basauri Basque, with transitive predication and synthetic forms with dative agreement.21

.  All forms of *edun as ‘have’ are identical with those as transitive auxiliary, and this identity extends whatever root is used in a given dialect for the ditransitive auxiliary, -i(central), eutsi (west), eradun (east), to make available synthetic dative-agreeing forms ‘have’

 Beatriz Fernández & Milan Rezac

(37) a. Orrek e niri lokatzako dekoste this.erg too I.dat mud.as eduki (1sgdat-3sgerg) ‘This one too slanders me [lit. has me for mud]’ (Arretxe 1994: 190) b. Emen dekostasu here eduki (1sgdat-2sgerg) ‘You have me here’

(Arretxe 1994: 234)

Dima DOM in transitive predication with eduki raises the question of why DOM is not available with *edun, in Dima and elsewhere. We will end up suggesting that DOM occurs when the regular absolutive Agree/Case structure has an extra feature or head, P, that leads to dative rather than absolutive case and agreement, possibly with other consequences like object shift. It can then be stipulated that *edun is a lexical exception to the possession of P. This might be derivable from the relationship between intransitive predication with izan ‘be’ and transitive predication with *edun ‘have’. If the latter is just the former plus an applicative head introducing the ergative, as in (30), then there is no agent-introducing v of regular transitives, and so no P if P depends on this v. Evidence that P does depend, at least partly, on agentive v comes from constraints like agentivity on DOM both in some varieties of Basque and elsewhere. The question then is why eduki in Dima does have P. One possibility is to hark to the morphology. Dima is special in having synthetic forms of eduki with the dative flag -ki- and dative agreement.22 This might be evidence for the learner to postulate P on v, since its surface outcome is dative agreement with the object. Outside Dima, when eduki participates only in restricted transitive predication of the English type, the very restrictions imposed on it might be evidence of a less direct relationship between intransitive and transitive predication, giving have more content, like restrictions on event structure (cf. Ritter and Rosen 1997). The

in meanings like ‘you have it for me’. For eduki, dative-agreeing synthetic forms are not part of the standard language, but occur in both old and modern dialects across the Basque Country, east and west, without it being clear whether they go to a common source (Gaminde 2007). Beside their application to DOM, which is a recent phenomenon, they include the meanings for (Aitak batek semeari diadukon amorioa ‘The love that a father has for his son.’, Axular Guero chapter XXXIII/§234), from (entzunde dekotzat aitxeri ‘I have heard this from my father’, Gaminde 2007: 55), and apparent ethical datives (Orrek estekotzue erremediorik ‘(S)he is hopeless and it affects you.’ Gaminde 2007: 55). For eduki’s dative-agreeing synthetic forms, see also Fernández (2013). .  The dative flag is added to the root -duk-. Historically, -duk-/eduki itself has been analysed as *edun + the dative flag -ki- (Trask 1981), but synchronically at any rate this -ki- is not an indicator of dative agreement: verbs so formed do not need to take dative objects, including eduki (B. Oyharçabal p.c.). See also Fernández (2013).



Differential object marking in Basque varieties 

effect of both synthetic ki-forms and of restrictions on event structure is to sever the simple relationship between the be of intransitive predication and be+P of transitive predication seen in izan -*edun, giving transitive predication with eduki more content and structure, and thus allowing it to bear P.23 4.3  Double dative constructions: DOM objects + indirect objects Case Theory posits that bearers of structural Case need to Agree with clausal Agree/Case loci for Case licensing (Chomsky 2000). In Basque finite/agreeing clauses, this requirement is visible in obligatory person agreement with absolutives. DOM datives behave similarly, suggesting the same mechanics, partly in contrast to indirect object datives. We focus on person agreement, which requires a brief explanation. Absolutive and dative agreement morphemes have distinct form and position. In dative agreement, 1st, 2nd, and 3rd persons each control distinctive morphology, including 3SG -o-, 3PL -o-te- or -e- where -te- is a plural marker. In absolutive agreement, 1st and 2nd person control distinctive morphology, but 3rd person is indicated only by plural agreement in 3PL, and 3SG has no expression. This is one of several pieces of evidence that all controllers of dative agreement group with controllers of 1st/2nd person absolutive agreement as +person, against controllers of 3rd person absolutive agreement (Arregi and Nevins 2011; Etxepare 2012; Rezac 2011). The distinction must be grammaticalized, because inanimate 3rd person dative indirect objects are still +person, while even human 3rd person absolutives are not (DOM datives agree like other datives). The grouping plays a role in the Person Case Constraint of Bonet (1991): dative agreement (even by inanimates) is not compatible with 1st/2nd person absolutive agreement but only with 3rd person

.  Outside DOM, Basque has one verb that seems to take ECM complements and always assign dative to the ECM subject: iritzi ‘deem, consider’, like eduki with synthetic dative-­ agreement forms (Zabala 2003). Iritzi passes Moulton's (2013) diagnostic for ECM in small clauses: in (i), the quantifier in the dative can take scope below the predicate complement, as it should not if it were a matrix argument and anteceded PRO/pro in the predicate complement. Questions arise about the exceptionality of iritzi (why is there no ECM assigning dative in Icelandic or French?), possible links to a/le-marking in Spanish ECM structures (see (4), which Ormazabal and Romero (2013) analyse it through object shift as other DOM), and any relationship between iritzi and Basque DOM (the iritzi construction is historically and dialectally independent of DOM, and includes inanimate datives that do not fall under the purview of DOM).

(i) Hiru ikuskatzaileri/aukerari beharrezkoak deritzegu, baina zaila izango da aurkitzea ‘We deem (3pldat-1plerg) three reviewers.dat/options.dat necessary, but it will be difficult to find them.’ (deem > necessary > three)

 Beatriz Fernández & Milan Rezac

(even if human), or in other words, two +person internal arguments cannot both agree (see Laka 1993; Albizu 1997; Arregi and Nevins 2011; Rezac 2011 for overviews focusing on Basque). We can now probe the licensing of +person absolutives. In finite/agreeing clauses, 1st/2nd person absolutives must control person agreement, whether the outcome is good as in (38a) because a dative goal need not agree, or bad due to the Person Case Constraint, as in (38b) where a dative causee must agree (Rezac 2010). (38c) is given for comparison to show that (38b) is good if the absolutive is not +person. (38) a. Poliziari zu eramango zaituzte / police.d.dat you.abs carry.fut trn (2pla-3ple) / *zaizkiote / *di(zki)ote   dtrn (2pla-3sgd-3ple) /dtrn (3sg/pla-3sgd-3ple) ‘They will bring you to the police’ b. *Pellori zu ezagutaraziko zaituzte /    Pello.dat you.abs know.make.fut trn (2pla-3ple) / diote / dute dtrn (3sg/pla-3sgd-3ple) /dtrn (3sga-3ple) ‘They will made Pello know you’ c. Pellori ikasleak ezagutaraziko dizkiote Pello.dat students.d.abs know.make.fut dtrn (3pla-3sgd-3ple) / *dituzte /    dtrn (3pla-3ple) ‘They will made Pello know the students’  (Rezac 2010, Tolosa, Central Basque)

The obligatory agreement of +person absolutives may be attributed to their need to Agree to satisfy a licensing requirement like Case or a person-specific version thereof (Béjar and Rezac 2009). Dative indirect objects fall into two groups (Albizu 1997, 2001, 2011; ­Elordieta 2001; Etxepare and Oyharçabal 2013; Fernández and Landa 2009; Fernández and Ortiz de Urbina 2010; Fernández 2011; Rezac 2008b, 2011; Rezac, Albizu and ­Etxepare 2014). One type, in (38a) above, is structurally c-commanded by the absolutive and limited to interpretations similar to those of the English prepositional object construction, like goals. These low datives do not need to control dative agreement under certain and dialectally varying conditions. They have been analysed as PPs with inherent Case. The other type c-commands the absolutive, and can in addition have interpretations like possessor, experiencer, and causee in (38b) and (38c). These high datives must control dative agreement, even if doing so would incur the ­Person Case constraint, as in (38b) and (38c). They have been analysed as



Differential object marking in Basque varieties 

a­ pplicatives. Their dative could be viewed as structural with obligatory agreement (cf. Ormazabal and Romero 2013 for Spanish), or as inherent with obligatory agreement attributed to the mechanics of absolutive Agree, cliticizing (clitic-doubling) a high dative because it is on its path (Rezac 2011, building on Anagnostopoulou 2003). Consider now DOM dative objects. Their case and agreement is dative, like that of indirect objects, but their theta-role is that of absolutive direct objects, and depictive licensing shows that they do not use the prepositional or applicative constructions. Their agreement can be examined by combining them with another dative, which has been investigated particularly for combinations with low datives in Albizu and Fernández (2006), Fernández and Rezac (2010), and Odria (2014, in progress). The Basque agreement complex permits only one instance of dative agreement. When it is taken up, for instance by the dative causee-agent of a causativized ditransitive, there is variation in the acceptability of a low dative without agreement (Trask 1981: 294; Ortiz de Urbina 2003b).24 When a DOM dative and a low dative combine, there is similar variation, but one constant: whatever the low dative does, the DOM dative must agree. Albizu and Fernández (2006) investigate a speaker of Markina Basque. Combination of optional DOM and a low dative is unavailable: (39) a. Martak Aneri eraman dio ikastolara Marta.erg Ane.dat carry dtrn (3sgd-3sge) school.d.to ‘Marta carried Ane to school’ b. *Martak Aneri eraman dio amonari    Marta.erg Ane.dat carry dtrn (3sgd-3sge) grandma.d.to ‘Marta carried Ane to (her) grandma’ (Albizu and Fernández 2006)

Instead, DOM is avoided, or the goal is marked allative (we gloss it ‘to’): (40) a. Martak Aneri eraman dio amonarengana Marta.erg Ane.dat carry dtrn (3sgd-3sge) grandma.d.to ‘Marta carried Ane to (her) grandma’ b. Martak Ane eraman dio amonari. Marta.erg Ane.abs carry dtrn (3sga-3sgd-3sge) grandma.d.dat ‘Marta carried Ane to (her) grandma’ (Albizu and Fernández 2006)

.  We do not know to what extent we are dealing with syntactic ungrammaticality versus pragmatic issues, and to what extent with dialectal versus individual characteristics. Trask (1981: 294) contrasts variation in causativizing ditransitives in Milafranga and Oñati as a matter of dialect difference, while Ortiz de Urbina (2003b) highlights the role of factors like object position in speaker variation.

 Beatriz Fernández & Milan Rezac

Both options are independently available: they are not ‘repair’ strategies that exist only in the context of an agreement constraint, such as the Person Case Constraint, but independently available structures, as in English I brought (grandma) a book (to/for grandma). Similar behavior has been found for Lekeitio Basque (Fernández and Rezac 2010). The following example from Dima Basque differs in having obligatory DOM for 1st/2nd person. The DOM dative must agree, and an allative replaces the low dative: (41) *Medikuari /Medikugana eroan dotzu    doctor.d.dat /doctor.d.all carry dtrn (2sgdat-3sgerg) ‘(S)he carried you to the doctor.’

Araitz-Betelu Basque, studied in Fernández and Rezac 2010) and Arraztio (2010), has obligatory DOM for 1st, 2nd and 3rd person humans. This is so even if a 1st/2nd person DOM dative combines with a low dative. However, the low dative has an option not seen above: it can simply not agree, (43a), though it can also be replaced by an allative (43b, c): (43) a. Deabruak nei saldu diate etsaiai demons.erg I.dat sell dtrn (1sgdat-3plerg) enemy.dat ‘The demons have sold me to the enemy’ b. Martak nei eaman dit zugana Marta.erg I.dat carry dtrn (1sgdat-3sgerg) you.to ‘Marta have brought me to you.’ c. Martak eaman dizu zui nigana Marta.erg carry dtrn (1sgdat-3sgerg) you.dat me.to ‘Marta have brought you to me’

The data so far converge on the need of DOM datives to agree when combining with a low dative, with variation in whether a nonagreeing low dative is available. This fits the assimilation of DOM dative to structural Case like the absolutive: +person elements of both must control person agreement. However, the same speakers also have another phenomenon that relativizes this conclusion. In the following Araitz-Betelu example, a 3rd person DOM dative combines with a 1st/2nd person indirect object dative, and it is the latter that agrees. This example involves an understudied phenomenon, the doubling of dative agreement by the allative zugana. Such datives are arguably high datives of interest/affectedness. (44) Martak Anei ekarri dizu zugana Marta.erg Ane.dat carry dtrn (2sgdat-3sgerg) you.to ‘Marta have carried Ane to you’

Generalizing somewhat beyond the data, DOM datives need agreement when combining with low datives, but DOM datives combining with high datives may



Differential object marking in Basque varieties 

cede it to the latter. The paradigms are not complete and more study is needed to ascertain the role of dative height and of person. If the generalization is on the right track, it partly converges with results reached for Spanish DOM, which themselves leave much to be understood. Ormazabal and Romero (2013) find that while DOM is usually obligatory if its conditions are met, it can be suspended precisely when combining with an o ­ bligatorily-agreeing/­cliticized high dative. In (45a), DOM is obligatory as ­a-marking for most of ­Spanish and as leísmo in Basque Country Spanish. In (45b), a high dative is added, identifiable by clitic doubling. As consequence, neither a-marking nor leísmo DOM is available, and instead otherwise obligatory DOM is absent. The examples involve 3rd person DOM, but they extend to 1st/2nd person DOM in the presence of a 1st/2nd person dative for some speakers (Bonet 1991: 203, Rezac 2011: 184 Note 4). (45) a. Enviaron *(a) los enfermos a la doctora sent.they    p the sick p the doctor ‘They sent the sick to the doctor’  [Basque Country leísmo: Les/*los enviaron a la doctora] b. Le enviaron (*a) los enfermos a la doctora cl.dat sent.they    p the sick p the doctor ‘They sent the sick to the doctor’  [Basque Country leísmo: Se los/*les enviaron a la doctora.]

The existence of an otherwise unavailable syntactic coding of an argument is widespread cross-linguistically to avoid the Person Case Constraint, such as a nonagreeing low dative in western Basque where all datives must agree otherwise (Bonet 1991; Albizu 1997; Rezac 2011). However, by and large, +person direct objects cannot suspend agreement to avoid the constraint if accusative or absolutive, as seen earlier for Basque in (38b) above; such suspension only seems available, sometimes, in systems where they are coded in the same way as indirect objects, as in Spanish (45b) above (Rezac 2011 generally and 184 Note  4, 252 specifically). It is not clear what this entails for the hypothesis that DOM objects are like +person absolutives in needing Agree for licensing. As if this were not mysterious enough, Araitz-Betelu (44) is not quite reducible to Spanish (45b): in Spanish both a-marking of the object and leísmo on the verb are suspended, while in Araitz-Betelu the object fails to control dative agreement but remains dative in case. We leave this subject in this presently unsettled state.25

.  We might simply suppose that in Araitz-Betelu the DOM dative agrees using the zero otherwise found for 3rd person singular absolutives. There are isolated data suggesting such a possibility: a different Araitz-Betelu speaker lets a DOM dative do this for 2nd person, (i).

 Beatriz Fernández & Milan Rezac

4.4  Dative-absolutive alternations: The role of clausal properties Our results so far suggest systematic parallelism between DOM datives and canonical absolutives in interpretation, configuration, and Agree/Case. In this section, we look at clausal properties that influence the availability of DOM: tense and finiteness/agreement. They lead to alternations between DOM datives and canonical absolutives that have no counterpart with indirect object datives. We draw two conclusions. One, the DOM dative is again a structural Case, one that must be able to alternate with the structural absolutive. Second, the alternation must be modulable by properties on high clausal heads. One way to work this out is for both the DOM dative and the canonical absolutive to reflect a structural Agree/Case relations sensitive to properties of the entire agreement complex hosting the probe. We begin with tense. In some varieties, DOM is present or obligatory in the past but absent or optional in the present. This is a revealing condition, for there is nothing like it for indirect object datives, nor for similar DOM elsewhere as in Spanish. Tense may condition DOM in Iranian languages like Zazaki, but there it is contingent on an overall alignment split between present and past (Haig 2008). In Basque past and present are equally ergative morphologically and equally accusative in syntax.26 One fairly clear conclusion to draw is that the DOM dative is not inherent Case, one that depends on the selectional relationship between an argument and its predicate. Argument coding in virtue of selection is not affected by tense, at least not for internal arguments: speak to does not lose or change its ­preposition

Possibly to be related is Albizu’s (1997) report of a PCC repair, whereby the 1st/2nd direct object is absolutive in case but dative in agreement, (ii). We have not encountered this ­phenomenon. (i) Martak zui negana ekarri zattu Marta.erg you.dat me.to carry trn (2sgabs-3sgerg) ‘Marta brought you to me’ (ii) Azpisapoek ni etsaiari traitors.d.erg me.abs enemy.d.dat saldu *naute/didate sell   trn (1sga-3ple)/dtrn (1sgd-3ple)

‘Traitors sold me to the enemy’

.  The Basque past tense does exhibits the ergative displacement phenomenon in person agreement (not in case nor in number agreement), which has been viewed as a morphological relic of split ergativity (see Gómez and Sainz 1995 and references there). Even if this is so, it belongs to the prehistoric development of the language with no effect on synchronic alignment, whereas DOM is a recent development.



Differential object marking in Basque varieties 

a­ccording to tense, only in virtue of argument structure operations that give bespeak. This follows from the locality of selection, which occurs under phrasestructural sisterhood, too limited to encompass an influence of T on V (see recently Shlonsky 2006 for cartographic frameworks). Structural Case, on the other hand, reflects the phrase-structurally unbounded Agree relation, so the DOM dative could be sensitive to properties of T. Less clear is what to make of the absence of tense conditioning on comparable DOM outside Basque. A second conclusion to draw is that the mechanism assigning DOM dative must cede to the absolutive if its conditions are not met. In Basque, the absolutive is an assigned structural Case rather than a freely available default, since Basque has the same Case Filter effects as English in analogues of I am afraid *(of) his arrival (Rezac, Albizu and Etxepare 2014). Yet alternations between the DOM dative and the canonical absolutive do have the character of the absolutive emerging whenever the dative is not available, under finely parametrizable conditions such as DOM for 1st/2nd person in past tense and absolutive otherwise. They do not resemble voice-based nominative-accusative alternations, where each case is tied to a distinct functional item, T versus active v. A similar conclusion may be reached from the sensitivity of DOM to finiteness or agreement. Basque finite clauses agree with absolutives, ergatives, and datives, while nonfinite ones do not. The difference between the two clause types can affect DOM; DOM is less available in nonfinite/nonagreeing clauses. Again, this factor never affects the dative marking of indirect objects, nor, as far as we know, DOM in languages like Spanish. In some varieties, DOM occurs under the same conditions in both clause types, as in Araitz Betelu Basque (Arraztio 2010). In (46a), the DOM dative is the object of the participle immediately associated with the agreeing auxiliary, and the auxiliary agrees with the dative. In (46b), it is the object of the infinitival complement of the progressive ari and the matrix auxiliary associated with ari does not agree with arguments of the infinitive (cf. Ortiz de Urbina 2003a). (46) a. (Nik) Jonei ikusi diot    I.erg Jon.dat see dtrn (3sgdat-1sgerg) ‘I saw John’ b. Jon Mikeli ikusten ai da Jon.abs Mikel.dat see prog intrn (3sgabs) ‘Jon is seeing Mikel’ (Arraztio 2011, 3rd speaker)

Others varieties like Dima Basque diminish the extent of DOM in nonagreeing/ nonfinite clauses. DOM is obligatory for 1st/2nd person in agreeing clauses, so that a 1st/2nd person object can only be dative. In the gerund complements of come, however, a 1st/2nd person object can be dative or absolutive. These gerund

 Beatriz Fernández & Milan Rezac

complements are full clauses without restructuring or transparency for agreement (Artiagoitia 2003). (47) Seu/Seuri ikusten etorri nes you.abs/dat seeing come intrn (1sgabs) ‘I have come to see you’

DOM in Dima remains obligatory in the participial complements of gura ‘want’, but these are transparent to agreement and have other properties of full restructuring (Ortiz de Urbina 2003b). They thus form a single clausal architecture with the finite, agreeing auxiliary, which agrees with the DOM dative. (48) Seuri eroan gure dotzut you.dat carry want dtrn (2sgdat-1sgerg) ‘I want to carry you’

It is not yet clear which of finiteness or agreement, or which subcomponent of them, is relevant to DOM. Attributing it to agreement would converge with a traditional perspective: dialectal descriptions often neglect argument marking and speak of DOM as a phenomenon affecting auxiliary agreement patterns. Further insight might be had through a wider range of restructuring configurations, where agreement and finiteness partly divorce (cf. Ortiz de Urbina 2003b; Arregi and Molina-Azaola 2004). The sensitivity of DOM to finiteness/agreement leads to the same conclusion as its sensitivity to tense. First, the DOM dative must reflect a structural Case, since inherent Case is determined too low in the clause for its finiteness/­agreement to be taken into account.27 Second, the structural Case mechanism resulting in the DOM dative must cede to the absolutive when its conditions are not met. We propose an account of these findings in the next section. 5.  A theory of DOM We have reached the following conclusions for the syntax of Basque DOM: A. The DOM dative occurs in a configuration akin to canonically absolutive direct objects but not applicative or prepositional dative indirect objects.

.  The sensitivity of DOM to finiteness/agreement must be kept distinct from the use of different argument coding in virtue of the verbal-nominal difference, which is determined lower (roots or their v vs. n categorizers), and does affect aspects of coding like preposition choice: I greet you vs. My greeting to you.



Differential object marking in Basque varieties 

B. The DOM dative is a structural Case like the absolutive. C. The DOM dative is parametrisable by properties of both the goal and the clausal architecture, at least transitivity (v) and tense (T) as well as being in a finite/agreeing or nonfinite/nonagreeing clause. D. The DOM dative alternates with the absolutive so that the absolutive emerges when conditions for the DOM dative are not met. One way to unify these desiderata is to take the DOM dative to be a variant of the canonical absolutive: the same Agree/Case mechanism underlies both, but its outcome is modulated by additional features, which result in dative instead of absolutive and potentially in other consequences like object shift. Let us make the following assumptions: i. Following analyses of Basque ergativity, the Agree/Case locus for absolutive objects is a probe on v (Rezac, Albizu and Etxepare 2014 and references). ii. Functional architecture from v to Fin amalgamates in a single agreement complex in Basque, surfacing as the agreement complex with ergative, dative, absolutive agreement, mood, tense, and complementizers, but separate from the lexical verb and aspect (Laka 1993; Haddican 2007 and references). To model DOM, we now propose: iii. DOM dative and canonical absolutive case and agreement reflect Agree by the same head, v, differentiated according to the presence or absence of a feature P on v. The presence of P is reflected as dative case and agreement. The way P results in dative case and agreement might be a simple featural matter, whereby v with P values the [uCase] of its goal to a value spelled out as dative, and on v with P valued [uphi] is spelled as dative agreement morphology. However, we also need to make DOM sensitive to properties of the goal, animacy and referentiality. These are typical conditions on object shift (Diesing and Jelinek 1995; Holmberg 1999; Chomsky 2001), and analyses of similar DOM elsewhere do reduce it to object shift or a similar A-movement with respect to canonically marked objects (Torrego 1998; Bhatt 2007; Ormazabal and Romero 2013). Therefore, tentatively: iv. P is a trigger for object shift, associated with interpretive conditions in the way as has been discussed for other types of object shift. This construal of P is agnostic about whether and how it relates to dative agreement. Recent work has argued that Basque dative agreement is clitic doubling of dative arguments in the applicative construction, distinct from absolutive agreement that is the result of phi-Agree perhaps partly accompanied by clitic doubling (Arregi and Nevins 2011; Etxepare 2012; Preminger 2011; Rezac 2011 for overviews). The

 Beatriz Fernández & Milan Rezac

distinctive agreement morphology (clitics), dative flag (perhaps Appl), and root allomorphy (root + Appl) of dative agreement have been attributed to this mechanism. We have seen that DOM datives do not involve the applicative structures of any indirect object datives, so they cannot themselves have an Appl to be realised as the dative flag or condition root allomorphy. Yet a rapprochement might be achieved through object shift, if both applicative and object shift structures can be unified as providing a higher A-position for the objects that participate in them, perhaps licensed by a feature like P on v, perhaps by a special head hosting this feature in the agreement complex. This comes close to the unification of DOM and high dative coding in Spanish by Ormazabal and Romero (2013), with the difference that in Spanish +person objects always participate in this structure and so participation in this structure seems to be require for +person, whereas in Basque they only do so under DOM. Finally, we need to capture the effect of clausal properties on DOM: v. The presence of P on v can be sensitive to any properties in the agreement complex that hosts it, including those of Fin and T. There are various options for the mechanism by which properties like tense on T can affect the presence of P on v thanks to formation of the agreement complex. Local selection might suffice, between say [past] and P, if formation of the complex creates a single minimal domain for all the heads involved (Chomsky 1995). Morphological conditions on well-formedness of feature combinations in the complex might be invoked, as seen in other complex units like clitic clusters (Bonet 1991). More radically, we might extend Agree from being sensitive to features of the terminal that hosts the probe (so that Agree by [uphi] can assign [iCase:ACC] if on agentive v) to sensitivity to features of the entire derived X0 hosting the terminal, namely the Basque agreement complex (for related discussion, see Keine 2010). The formation of the agreement complex is essential in all these options, and might be used to distinguish Basque from other languages where DOM is not sensitive to high clausal material. We might suppose, for instance, that the Basque agreement complex is formed in syntax, whereas in Spanish v and T only amalgamate in morphology (on syntactic versus morphological head movement, see Embick and Noyer 2001).28

.  In the Germanic languages, there are correlations between finiteness and object shift, but due to a factor that cannot play a role in Basque DOM: the height of the verb, which in some systems permits object shift by moving out of its way (Holmberg and Platzack 1995), while in others simply permits it to be seen by staying low (Fischer et al. 2004).



Differential object marking in Basque varieties 

Each element of our proposal above is only one in an array of possibilities. We have spelled out one set of choices concretely to illustrate a possible approach to Basque DOM that seems to us to advance in deriving its properties. However, the choices are highly tentative: in many cases further work is needed to understand the phenomena we have discussed, as in the interaction of DOM and dative agreement, in others diagnostics still remain to be developed if possible, as for height of the object with and without DOM. Even so, our results circumscribe the hypothesis space, eliminating options such as the assimilation of DOM datives to indirect object datives, and revealing conditions on DOM like tense that are remarkable among most nearly comparable phenomena.

6.  Conclusions In this chapter, we have explored the nature of Basque DOM. DOM dative coding of transitive objects has the same morphology as the coding of indirect objects, as does DOM in Spanish, possibly with mutual influence between the two languages, and frequently cross-linguistically, as in Hindi-Urdu. Yet our investigation of the syntax of DOM reveals that the syntax of DOM dative objects is much like that of DOM absolutive objects, without being fully identifiable with it, and leaves only narrow scope for similarities to the syntax of indirect objects. Many aspects of Basque DOM remain to be understood, but contributions as Odria’s (in progress) or ours are beginning to circumscribe the phenomenon.

References Abney, Steven. 1987. The English noun phrase in its sentential aspect. Ph.D. dissertation, MIT. Aissen, Judith. 2003. “Differential object marking: Iconicity vs. Economy.” Natural Language and Linguistic Theory 21: 435–483.  doi: 10.1023/A:1024109008573 Albizu, Pablo. 1997. “The syntax of person agreement.” Ms., University of Southern California. Albizu, Pablo. 1998. “Generalized Person-Case Constraint: A Case for a Syntax-Driven Inflectional Morphology.” In Theoretical Issues on the Morphology-Syntax Interface [Supplements of ASJU 40], Myriam Uribe-Etxebarria and Amaia Mendikoetxea (eds.), 1–33. Donostia/ Bilbo: Gipuzkoako Foru Aldundia/UPV/EHU. Albizu, Pablo. 2001. “Datibo sintagmen izaera sintaktikoaren inguruan: eztabaidarako oinarrizko zenbait datu.” In Kasu eta komunztaduraren gainean. On Case and agreement, Beatriz Fernández and Pablo Albizu (eds.), 49–69. Bilbo: UPV/EHU. Albizu, Pablo. 2002. “Basque verbal morphology: Redefining cases.” In Erramu boneta: Festchrift for Rudolf P. G. de Rijk,  Xabier Artiagoitia, Patxi Goenaga and Joseba A. Lakarra (eds.), 1–19. Bilbo: UPV/EHU.

 Beatriz Fernández & Milan Rezac Albizu, Pablo. 2011. “Micro-variación en las construcciones inacusativas de dativo posesivo del vasco/Unaccusative predicates with possessor datives in Basque.” Handout from Wedisyn’s 2nd Workshop on Syntactic Variation, February 17–18, 2011, Universitat de Barcelona. Albizu, Pablo and Fernández, Beatriz. 2006. “Licit and illicit ERG-DAT pairings.” In Andolin gogoan, Beatriz Fernández and Itziar Laka (eds.), 69–96. Bilbo: UPV/EHU. Anagnostopoulou, Elena. 2003. The Syntax of Ditransitives: Evidence from Clitics. Berlin/New York: Walter de Gruyter. Arraztio, Kontxi. 2010. “Datibo osagarri bitxiak Araitz-Beteluko hizkeran.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai. [Supplements of ASJU 52], Beatriz Fernández, Pablo Albizu and Ricardo Etxepare (eds.), 187–194. Bilbo: UPV/EHU. Arregi, Karlos and Molina-Azaola, Gainko. 2004. “Reestructuring in Basque and the Theory of Agreement.” In Proceedings of WCFFL 23, G. Garding and M. Tsujimura (eds.), 101–114. Arregi, Karlos and Nevins, Andrew. 2011. Morphotactics. Dordrecht: Springer. Arretxe, Jon. 1993. Basauriko euskara. Basauri: Basauriko Udala. Arteatx, Iñigo. 2007. “Euskarazko oharmen-aditzen osagarrietako perpaus jokatugabeak.” Uztaro 63: 31–63. Arteatx, Iñigo. 2012. Euskarazko egitura jokatugabeak : -t(z)en morfemadun perpaus osagarrien azterketa sintaktikoa. Ph.D. dissertation, UPV/EHU. Artiagoitia, Xabier. 2003. “Subordination.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 634–752. Berlin: Mouton de Gruyter. Artiagoitia, Xabier. 2012. “Genitive case and multiple checking in Basque.” In Noun Phrases and Nominalization in Basque, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-­ Etxebarria (eds.), 209–242. Amsterdam: John Benjamins.  doi: 10.1075/la.187.10art Artola, Fernando (“Bordari”). 1981. “Hondarribiko mintzairan bertako berri.” Euskera XXVI. 295–303. Aurrekoetxea, Gotzon and Txillardegi. 1983. Euskal dialektologiaren hastapenak. Bilbo: UEU. Austin, Jane. 2006. “Dative overmarking in Basque.” Euskalingua 9: 136–145. Beavers, John. 2011. “On affectedness.” Natural Language and Linguistic Theory 29: 335–370. doi: 10.1007/s11049-011-9124-6 Béjar, Susana and Rezac, Milan. 2009. “Cyclic Agree.” Linguistic Inquiry 40–1: 35–73.

doi: 10.1162/ling.2009.40.1.35

Bhatt, Rajesh. 2007. “Unaccusativity and Case Licensing.” Handout available at: http://people. umass.edu/bhatt/papers/mcgill-may2007-handout1.pdf. Bleam, Tonia. 2003. “Properties of the Double Object Construction in Spanish.” In Romance perspective on language knowledge and use, Rafael Núñez-Cedeño, Luis López and Richard Cameron (eds.), 233–252. Amsterdam: John Benjamins.  doi: 10.1075/cilt.238.19ble Bonaparte, Luis Luciano. 1869. Le verbe basque en tableaux. Reprinted in: Opera omnia vasconice, vol. 1, 221–242, 1991. Bilbo: Euskaltzaindia. Bonet, Eulàlia. 1991. Morphology after Syntax. Pronominal Clitics in Romance. Ph.D. dissertation, MIT. Bossong, Georg 1991. “Differential object marking in Romance and beyond.” In New analyses in Romance linguistics, Dieter Wanner and Douglas A. Kibee (eds.), 143–170. Amsterdam: John Benjamins.  doi: 10.1075/cilt.69.14bos Bossong, Georg. 1998. “Le marquage differential de l’object dans les langues d’Europe.” In Actance et valence dans les langues d’Europe, Jack Feuillet (ed.), 193–258. Berlin: Mouton de Gruyter. Burzio, Luigi. 1986. Italian syntax. Dordrecht: Reidel.  doi: 10.1007/978-94-009-4522-7



Differential object marking in Basque varieties 

Butt, Miriam. 1993. “Object specificity and agreement in Hindi/Urdu.” In Papers from the 29th Regional Meeting of the Chicago Linguistic Society, Katharine Beals et al. (eds.), 89–103. Chicago: Chicago Linguistic Society. Bybee, Joan. 2010. Language, usage and cognition. Cambridge: Cambridge University Press.

doi: 10.1017/CBO9780511750526

Chomsky, Noam. 1986. Knowledge of Language: Its Nature, Origins and Use. New York: Praeger. Chomsky, Noam. 1995. The Minimalist Program. Cambridge (Mass.): The MIT Press. Chomsky, Noam. 2000. “Minimalist inquiries.” In Step by step, Roger Martin, David Michaels, and Juan Uriagereka (eds.), 89–155. Cambridge: MIT Press. Chomsky, Noam. 2001. “Derivation by phase.” In Ken Hale: A life in language, Michael ­Kenstowicz (ed.), 1–52. Cambridge: MIT Press. Couquaux, Daniel. 1981. “French predication and linguistic theory.” In Levels of Syntactic Representation, Ken Hale and Samuel J. Keyser (eds.), 33–64. Dordrecht: Foris. Diesing, Molly and Jelinek, Eloise. 1995. “Distributing arguments.” Natural Language Semantics 3: 123–176.  doi: 10.1007/BF01249836 Dobrovie-Sorin, Carmen. 1994. The Syntax of Romanian. Berlin: Mouton de Gruyter. Duguine, Maia. 2012. “Basque nominalizations and the role of structural case in the licensing of null arguments.” In Noun Phrases and Nominalization in Basque, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 333–374. Amsterdam: John Benjamins.

doi: 10.1075/la.187.15dug

Eguren, Luis. 2012. “Predication markers in Basque.” In Noun Phrases and Nominalization in Basque, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 243–266. Amsterdam: John Benjamins.  doi: 10.1075/la.187.11egu Elordieta, Arantzazu. 2001. “Double Object Constructions and neutral order in Basque.” Verb movement and constituent permutation in Basque. Utrecht: LOT. Embick, David and Noyer, Rolf. 2001. “Movement operations after syntax.” Linguistic Inquiry. 32–4: 555–595.  doi: 10.1162/002438901753373005 Enç, Mürvet. 1991. “The semantics of specificity.” Linguistic Inquiry 22, 1: 1–25. Etxepare, Ricardo. 2003. “Valency and argument structure in the Baque Verb.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.). Berlin: Mouton de Gruyter. 363–425. Etxepare, Ricardo. 2012. Agreement configurations in Basque: Lessons from distance. Habilitation thesis, CNRS-IKER. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2012. Denominal necessity modals in Basque. In Noun Phrases and Nominalization in Basque, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 283–330. Amsterdam: John Benjamins.  doi: 10.1075/la.187.13etx Etxepare, Ricardo and Oyharçabal, Bernard. 2013. “Datives and adpositions in northeastern Basque.” In Variation in Datives: A Micro-comparative Perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 50–95. New York: Oxford University Press. Fernández, Beatriz. 2001. “Absolutibo komunztaduradun ergatiboak, absolutibo komunztaduradun datiboak: Ergatiboaren Lekualdatzetik Datiboaren Lekualdatzera.” In On Case and Agreement. Kasu eta Komunztaduraren gainean, Beatriz Fernández and Pablo Albizu (eds.), 87–112. Bilbo: UPV/EHU. Fernández, Beatriz. 2004. “Gustatzen nau gustatzen dizu: aditz laguntzaile eta komunztadura bitxiak perpaus ez-akusatiboetan.” In Euskal gramatika XXI. mendearen atarian: arazo zaharrak, azterbide berriak. Pablo Albizu and Beatriz Fernández (eds.), 89–112. VitoriaGasteiz: UPV/EHU/Arabako Foru Aldundia.

 Beatriz Fernández & Milan Rezac Fernández, Beatriz. 2008. “Quirky dative objects in Basque.” Paper presented at European Dialect Syntax Meeting III, September, 18–21, 2008, Venice. Fernández, Beatriz. 2011. “Goi- eta behe-datiboak eta Datibo-komunztaduraren murriztapena: ipar-ekialdeko hizkerak eta euskara estandarra.” In Euskal dialektologia: lehena eta oraina. [Supplements of ASJU 69], Irantzu Epelde (ed.), 37–63. Bilbo: UPV/EHU. Fernández, Beatriz. 2013. “-(k)i hiztegian, eta hiztegiaz harantzago.” ASJU 47 (2): 1–26. Fernández, Beatriz. 2015. “The syntax of -(k)i, a Basque preposition.” In Syntactic variation in Western European languages: From the noun phrase to clause structure [Special Issue V of Dialectologia], Xulio Sousa, Carlota de Benito and Víctor Lara (eds.), 81–107. On-line: http://www.publicacions.ub.edu/revistes/dialectologiaSP2015/. Fernández, Beatriz and Ezeizabarrena, Marijose. 2001. “Itsasaldeko solezismoa, Datiboaren Lekualdatzearen argipean.” In Euskal gramatikari eta literaturari buruzko ikerketak XXI. mendearen atarian: Gramatika gaiak [Iker, 14 (1)], Jesus Mari Makazaga and Bernard Oyharçabal (eds.), 255–278. Bilbo: Euskaltzaindia. Fernández, Beatriz and Landa, Josu. 2009. “Datibo komunztadura beti zaindu, inoiz zaindu ez eta batzuetan baino zaintzen ez denean. Hiru ahoko aldagaia, datu iturri bi, eta erreminta bat: Corsintax.” Lapurdum 13: 159–181. doi: 10.4000/lapurdum.2053 Fernández, Beatriz and Ortiz de Urbina, Jon. 2010. Datiboa hiztegian. Bilbo: UPV/EHU. Fernández, Beatriz and Ortiz de Urbina, Jon. 2012. “Dative (first) complements in Basque.” In Syntactic Microvariation in Westmost European Languages, [Special Issue of Journal of Portuguese Linguistics 11 (1)], Ernestina Carrilho and Beatriz Fernández (eds.), 83–98. ­Lisboa: Ediçoes Colibrí – Universidade de Lisboa. Fernández, Beatriz and Rezac, Milan. 2010. “Datibo osagarri bitxiak eta Datiboaren Lekualdatzea: ari nai diyot eta kanta egin nazu bidegurutzean.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai. [Supplements of ASJU 52], Beatriz Fernández, Pablo Albizu and Ricardo Etxepare (eds.), 113–149. Bilbo: UPV/EHU. Fernández-Ordoñez, Inés. 1999. “Leísmo, laísmo y loísmo.” In Gramática descriptiva de la lengua española I, Ignacio Bosque and Violeta Demonte (eds.), 1317–1394. Madrid: Espasa Calpe. Fischer, Olga, van Kemenade, Hans, Koopman, Willem and van der Wurff, Wim. 2004. The Syntax of Early English. Cambridge: Cambridge University Press. Gaminde, Iñaki. 2007. “Datibodun adizkiak jakin eta eduki aditzekin Bizkaian.” Euskalingua 10: 53–62. Gómez, Ricardo and Sainz, Koldo. 1995. “On the origin of the finite forms of the Basque verb.” In Towards a history of the Basque language, José Ignacio Hualde, Joseba A. Lakarra and Jon Ortiz de Urbina (eds.), 235–274. Amsterdam: John Benjamins.  doi: 10.1075/cilt.131.10gom Gutiérrez-Rexach, Javier. 2000. “The formal semantics of clitic doubling.” Journal of Semantics 16: 315–380.  doi: 10.1093/jos/16.4.315 Haddican, Bill. 2007. On egin: do-support and VP focus in Central and Western Basque. NLLT 25: 735–764. Haig, Geoffrey. 2008. Alignment Change in Iranian Languages: A Construction Grammar Approach. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110198614 Himmelmann, Nikolaus P. and Schultze-Berndt, Eva. 2005. “Issues in the syntax and semantics of participant-oriented adjuncts: an introduction.” In Secondary Predication and Adverbial Modification: The typology of Depictives, Nikolaus P. Himmelmann and Eva Schultze-Berndt (eds.). Oxford: Oxford University Press. 1–68.

doi: 10.1093/acprof:oso/9780199272266.003.0001



Differential object marking in Basque varieties 

Holmberg, Anders. 1999. “Remarks on ‘Holmberg’s Generalization’.” Studia Linguistica 53: 1–39.  doi: 10.1111/1467-9582.00038 Holmberg, Anders and Platzack, Christer. 1995. The Role of Inflection in Scandinavian Syntax. New York: Oxford University Press. Hualde, José Ignacio. 2003. “Verbs: Finite forms.” In A Grammar of Basque, José Ignacio and Jon Ortiz de Urbina (eds.), 205–246. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110895285 Kayne, Richard S. 1993. “Towards a modular Theory of auxiliary selection.” Studia Linguistica 47 (1): 3–31. Keine, Stefan. 2010. Case and Agreement from Fringe to Core. Berlin: Mouton de Gruyter.

doi: 10.1515/9783110234404

Kornfilt, Jaklin. 1997. Turkish. London: Routledge. Jelinek, Eloise and Carnie, Andrew. 2003. “Argument hierarchies and the mapping principle.” In Formal Approaches to Function, Andrew Carnie, Heidi Harley and Mary A. Willie (eds.), 265–296. Amsterdam: John Benjamins.  doi: 10.1075/la.62.20jel Lambton, Ann K.S. 1993. Persian Grammar. Cambridge: Cambridge University Press. Landa, Alazne. 1995. Conditions on null objects in Basque Spanish and their relation to leísmo and clitic doubling. Ph.D. dissertation, University of Southern California, Los Angeles. Laka, Itziar. 1993. “The structure of inflection: A case study in X0 syntax.” In Generative studies in Basque linguistics, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 21–70. Amsterdam: John Benjamins.  doi: 10.1075/cilt.105.02lak Lazard, Gilbert. 2001. “Le marquage différential de l′object.” In Language Typology and Language Universals: An International Handbook. Martin Haspelmath, Ekkehard Köning, Wulf Oesterreicher and Wolfgang Raible (eds.), 873–885. Berlin: Mouton de Gruyter. Masica, Colin P. 1982. “Identified object marking in Hindi and other languages.” In Topics in Hindi Linguistics, vol. 2, Omkar N. Koul (ed.), 16–50. Delhi: Bahri Publications. McFadden, Thomas. 2003. “The underlying structures of German inherent case.” Ms., University of Pennsylvania. McFadden, Thomas. 2004. The position of morphological case in the derivation: A study on the syntax-morphology interface. Ph.D. dissertation, University of Pennsylvania. Mendikoetxea, Amaya. 1999. “Construcciones con SE: medias, pasivas e impersonales.” In Gramática descriptiva de la lengua española, vol. II, Ignacio Bosque and Violeta Demonte (eds.), 1631–1722. Madrid: Espasa Calpe. Mohanan, Tara. 1994. Argument Structure in Hindi. Stanford (CA): CSLI Publications. Moulton, Keir. 2013. “Raising from the dead.” Linguistic Inquiry 44: 157–167.

doi: 10.1162/LING_a_00123

Mounole, Céline. 2012. “Evolution of the transitive verbs in Basque and apparition of datively marked patients.” In Ergativity and Voice, Gilles Authier and Katharina Haude (eds.), 355–379, Berlin: Mouton de Gruyter. Odria, Ane. 2012. What lies behind Differential Object Marking: a Basque dialect survey. Master Thesis, UPV/EHU. Odria, Ane. 2014. “Differential Object Marking and the nature of dative case in Basque varieties.” Linguistic Variation 14: 289–317.  doi: 10.1075/lv.14.2.03odr Odria, Ane. in progress. Differential Object Marking in Basque syntax: A microcomparative survey. Ph.D. dissertation, UPV/EHU. Oyharçabal, Bernard. 2007. Basque ditransitives. Ms., CNRS-IKER. Ormazabal, Javier and Romero, Juan. 2007. “Object agreement restrictions.” Natural Language and Linguistic Theory 25: 315–347.  doi: 10.1007/s11049-006-9010-9

 Beatriz Fernández & Milan Rezac Ormazabal, Javier and Romero, Juan. 2013. “Object clitics, agreement and dialectal variation.” Probus 20: 33–66. Ortiz de Urbina, Jon. 1989. Parameters in the Grammar of Basque. Dordrecht: Foris. Ortiz de Urbina, Jon. 2003a. “Semiauxiliary verbs.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.) 300–321. Berlin: Mouton de Gruyter.

doi: 10.1515/9783110895285

Ortiz de Urbina, Jon. 2003b. “Causatives.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 592–607. Berlin: Mouton de Gruyter. doi: 10.1515/9783110895285 Ortiz de Urbina, Jon. 2003c. “Impersonal clauses.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 572–691. Berlin: Mouton de Gruyter.

doi: 10.1515/9783110895285

Pylkkänen, Liina. 2002/2008. Introducing Arguments. Cambridge, MA: MIT Press. Preminger, Omer. 2011. Agreement as a fallible operation. Ph.D. dissertation, MIT. Raz, Shlomo. 1980. “Tigre syntax and Semitic Ethiopian.” Bulletin of the School of Oriental and African Studies 43 (2): 235–250.  doi: 10.1017/S0041977X00115617 Rebuschi, Georgés. 1984. Structure de l’énoncé en basque. Paris: SELAF. Rezac, Milan. 2006. Agreement displacement in Basque: Derivational principles and lexical parameters. Ms., UPV/EHU. On-line: www.umr7023.cnrs.fr/spip.php?article675. Rezac, Milan. 2008a. “The forms of dative displacement: From Basauri to Itelmen.” In Gramatika jaietan: Patxi Goenagaren omenez. [Supplements of ASJU 51], Xabier Artiagoitia and Joseba A. Lakarra (eds.), 709–724. Bilbo: UPV/EHU. Rezac, Milan. 2008b. “The syntax of eccentric agreement: The Person Case Constraint and Absolutive Displacement in Basque.” Natural Language and Linguistic Theory 26: 61–106.  doi: 10.1007/s11049-008-9032-6 Rezac, Milan. 2010. “On the unifiability of repairs of the Person Case Constraint: French, Basque, Georgian, and Chinook.” In Beñat Oihartzabali gorazarre. [Supplements of ASJU 43 (1–2)], Ricardo Etxepare, Ricardo Gómez and Joseba A. Lakarra (eds.), 769–790. Bilbo: UPV/EHU. Rezac, Milan. 2011. Phi-features and the Modular Architecture of Language. Dordrecht: Springer.  doi: 10.1007/978-90-481-9698-2 Rezac, Milan, Albizu, Pablo and Etxepare, Ricardo. 2014. The structural ergative of Basque and the theory of Case. Natural Language and Linguistic Theory 32: 1273–1330.

doi: 10.1007/s11049-014-9239-7

Rezac, Milan and Fernández, Beatriz. 2013. “Dative Displacement in Basque.” In Variation in Datives: A Micro-comparative Perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 256–282. New York: Oxford University Press. Rijk de, Rudolf. 2008. Standard Basque: A Progressive Grammar. Cambrige (Mass.): The MIT Press. Ritter, Elizabeth and Rosen, Sara T. 1997. “The function of have.” Lingua 101: 295–321.

doi: 10.1016/S0024-3841(96)00024-1

Sæbø, Kjell J. 2009. “Possession and pertinence: the meaning of have.” Natural Language Semantics 17: 369–397.  doi: 10.1007/s11050-009-9047-5 Silverstein, Michael. 1976. “Hierarchy of features and ergativity.” In Grammatical Categories in Australian Languages, Robert M.W. Dixon (ed.), 112–171. Canberra: Australian Institute of Aboriginal Studies. Shain, Cory. 2008. Differential object marking in Paraguayan Guaraní. BA Honors Thesis, Ohio State University.



Differential object marking in Basque varieties 

Shlonsky, Ur. 2006. “Projection étendue et cartographie de SC.” Nouveaux cahiers de linguistique française 2: 83–93. Stowell, Tim. 1978. “What was there before there was there?” In Papers from the Fourteenth Regional Meeting of the Chicago Linguistics Society, D. Farkas et al. (eds.), 457–471. ­Chicago: University of Chicago. Stowell, Tim. 1991. “Small clause restructuring.” In Principles and Parameters in Comparative Grammar, Robert Freidin (ed.), 182–218. Cambridge, MA: MIT Press. Torrego, Esther. 1998. The Dependencies of Objects. Cambridge, MA: MIT Press. Trask, Robert L. 1981. “Basque Verbal Morphology.” In Iker 1: 285–304. Trask, Robert. L. 1995. “On the history of the non-finite forms in Basque.” In Towards a History of the Basque Language, José Ignacio Hualde, Joseba A. Lakarra and Robert L. Trask (eds.), 207–234. Amsterdam: John Benjamins.  doi: 10.1075/cilt.131.09tra Vicente, Luis. 2005. “Licensing objects with and without movement.” Manuscript, Leiden University Centre for Linguistics.  Yrizar, Pedro. 1981. Contribución a la dialectología de la lengua vasca (2 vols.). Donostia: Gipuzkoako Kutxa. Yrizar, Pedro. 1991. Morfología del verbo auxiliar guipuzcoano (2 vols.) Bilbo: Euskaltzaindia. Yrizar, Pedro. 1992. Morfología del verbo auxiliar alto navarro septentrional (2 vol.). Iruñea: ­Euskaltzaindia/Nafarroako Gobernua. Yrizar, Pedro. 1997. Morfología del verbo auxiliar labortano. Iruñea: Euskaltzaindia/UPV/EHU. Yrizar, Pedro. 1999. Morfología del verbo auxiliar bajo navarro occidental. Iruñea: Euskaltzaindia/ UPV/EHU. Zabala, Igone. 2003. “Nominal predication.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 426–447. Berlin: Mouton de Gruyter. Zuazo, Koldo. 1998. “Nafarroako euskal hizkerak.” In Nafarroako hizkerak, Iñaki Camino (ed.), 1–22. Bilbo: UEU. Zuazo, Koldo. 2003. Euskalkiak, herriaren lekukoak. Donostia: Elkar. Zubizarreta, María Luisa. 1985. “Romance causatives.” Linguistic Inquiry 16: 247–289. Zubizarreta, María Luisa. 1998. Prosody, Focus and Word Order. Cambridge, MA: MIT Press.

Appendix: Dialectal grammars Aezkoa: Camino, Iñaki. 1997. Aezkoako euskararen azterketa dialektologikoa. Iruñea: Nafarroako Gobernua. Ahetze: Cabodevilla, Josu. 1991. Aetzen Uskara. Iruñea: Nafarroako Gobernua. Amaiur: N’Diaye, Genevieve. 1970. Structure du dialect basque de Maya. Mouton: The Hague. Antzuola: Larrañaga, Jone. 1998. Antzuolako hizkera. Antzuola: Antzuolako Udala. Arrasate: Elortza, Jerardo, Garro, Eneritz, Garai, Jesus M., Ormaetxea, José Luis and Plazaola, Esteban. 1999. Arrasateko euskara. Arrasate: Arrasateko Udala. Baztan: Salaburu, Pello and Lakar, Maite. 2005. Baztango mintzoa: gramatika eta hiztegia. ­Iruñea/Bilbo: Nafarroako Gobernua/Euskaltzaindia. Basauri: Arretxe, Jon. 1993. Basauriko euskara. Basauri: Basauriko Udala. Bortzerriak: Zelaieta, Edu. 2005. “Bortzerrietako euskara, herriz herri (ez)berdintasunetan ­barrena II.” FLV 99: 287–306.

 Beatriz Fernández & Milan Rezac Bergara: Elexpuru, Juan Martin. 1988. Bergarako Euskara. Bergara: UNED. Bermeo: Egaña, Aitzane. 1984. “Bermeoko aditzaren azterketa lorratzak.” Bermeo 4: 13–43. Laka, Enara, Olondo, Leire and Gaminde, Iñaki. 2012. “Bermeoko gazteen euskararen aditz morfologiaz.” Euskalingua 13: 27–36. Gaminde, Iñaki, Romero, Asier and Legarra, Hiart. 2012. Gramatika eta hizkuntz bariazioa Bermeon. Bermeoko Udala eta Campos Hegaluzea. Beskoitze: Duhau, Henri. 1993. Hasian hasi: Beskoitzeko Euskara. Donibane Lohitzune: Irkus. Deba: Zuazo, Koldo. 1999. Deba Ibarreko Euskeria. Antzuola: Antzuolako Udala. Zuazo, Koldo. 2002. Deba ibarretik euskararen herrira. Eibar: Badihardugu Euskera Alkartia. Eibar: Eibarko Euskara Mintegia. 1998. Eibarko aditza. Eibarko Udala. Etxebarria, Toribio. 1998. Flexiones verbales y lexicon del euskera dialectal de Eibar. Bilbo: Euskaltzaindia. Elgoibar: Makazaga, Jesus Mari. 2007. Elgoibarko Euskara. Elgoibarko berbak, egiturak eta irakurgaiak. Elgoibar: Elgoibarko Udala. Ermua: Aranberri, Fernando. 1996. Ermua eta Eitzako Euskara. Ermua: Ermuko Udala. Erroibar: Ibarra, Orreaga. 2000. Erroibarko eta Esteribarko hizkera. Iruñea: Nafarroako Unibertsitate Publikoa. Erronkari: Angos, Alberto. 2005. Erronkariera I: Gramatika. Donostia: Hiria. Forua: Gaminde, Iñaki. 1992. Foruko euskararen morfosintaxiaz. Forua: Foruko Udala. Hondarribia/Irun: Sagarzazu, Txomin. 2005. Hondarribiko eta Irungo euskara. Irun: Alberdania. Igorre: Iglesias, Aitor. 2005. Igorreko hizkeraren azterketa dialektologikoa. Ms., UPV/EHU. Imoz: Apalauza, Amaia. 2008. Imozko euskara. Iruñea: Nafarroako Gobernua. Lasarte-Oria: Lakaba, Ana, Azurza, María Eugenia and Bereziartu, Juan Inazio. 1996. LasarteOriako euskararen azterketa. Lasarte-Oria: Lasarte-Oriako udala. Laudio: Urkijo Orueta, Natxu. 1994 Zenbait apunte Laudioko euskaraz. Bilbo: UEU. Lazkao: Hurtado, Irene. 2001. Goierriko eta Tolosalde hegoaldeko hizkerak. Tolosa: Goierriko Euskal Eskola Kultur Elkartea, Maizpide Euskaltegia eta Lazkaoko Udala. Leioa: Gaminde, Iñaki. 1989. Leioako euskararen gramatikaz. Leioa: Leioako udala. Lekeitio: Hualde, José Ignacio, Elordieta, Gorka and Elordieta, Arantzazu. 1994 The Basque Dialect of Lequeitio [Supplements of ASJU XXXIV]. Bilbo and Donostia: UPV/EHU and Gipuzkoako Foru Aldundia. Mallabi: Mugarza, Pello 2006. Mallabiko euskara. Mallabiko Udala. Marquina: Rollo, William. 1925. The Basque Dialect of Marquina. Amsterdam: H.J. Paris. Oiartzun: Fraile Ugalde, Idoia and Fraile Ugalde, Ainhoa. 1996. Oiartzungo hizkera. Oiartzun: Oiartzungo Udala. Oñati: Badihardugu. 2005. Deba ibarreko aditz-taulak: Oñati. Ondarroa: Rotaetxe, Karmele. 1977. Estudio estructural del euskera de Ondarroa. Durango: Leopoldo Zugaza. Arregi, Karlos and Nevins, Andrew. 2011. Morphotactics. Dordrecht: Springer. Ordizia: Garmendia, Larraitz and Etxabe, Karmele. 2004. Ordiziako euskara: garai bateko bizimodua eta gaurko hizkuntza-egoera. Ordizia: Ordiziako Udala. Orio: Iturain, Iñaki and Loida, Loren. 1995. Orioko euskara. Orio: Orioko Udala.



Differential object marking in Basque varieties 

Otxandio: Burguete, Xabier and Gaminde, Iñaki. 1991. Otxandioko euskaraz. Otxandio: Otxandioko Udala and Bizkaiko Foru Aldundia. Pasaia: Agirretxe, Joxe Luix, Lersundi, Mikel and Olaetxea, Ortzuri. 1998. Pasaiako hizkera. Pasaia: Pasaiko Udala. Sakana: Erdozia, Jose Luis. 2001. Sakana erdialdeko euskara. Ph.D. dissertation, UPV/EHU. Sopela: Gaminde, Iñaki, Markaida, Bene and Markaida, Eukene. 1993. Sopelako euskaraz. Ugao: Salazar, Belene. 2001. Ugaoko euskara. Ugaoko euskararen azterketa etnolinguistikoa. Bilbo: Ediciones Beta. Ultzama: Ibarra, Orreaga. 1995. Ultzamako hizkera. Inguruko euskalkiekiko harremanak. ­Iruñea: Nafarroako Gobernua. Zaldibia: Etxabe, Karmele and Garmendia, Larraitz. 2003 Zaldibiako euskara: Bertako bizimodua, ohiturak, eta pasadizoak. Zaldibia: Zaldibiako Udala. Zamudio: Gaminde, Iñaki. 2000. Zamudio berbarik berba. Bilbo: Labayru Ikastegia. Zeberio: Etxebarria, Juan Manuel. 1996. Zeberio haraneko euskararen azterketa etno-inguistikoa. Bilbo: Deustuko Unibertsitatea. Zegama: Azurmendi, Joxe Migel. 1998. Zegamako euskara. Lazkao: Goierriko Euskal Eskola Kultur Elkartea.

Gaps and stopgaps in Basque finite verb agreement* Milan Rezac CNRS – IKER, UMR5478

This chapter surveys gaps in Basque finite verb agreement, their sources, and their theoretical consequences. First come interpretive gaps due to Condition B. In Basque they are grammaticalised, and reveal a useful distinction between agreement and clitic systems. Next are gaps due to conditions on syntactic dependencies, two taken up in detail. One is the impossibility of dative agreement across an absolutive, which does not rule out any forms but certain pairings of form and meaning. The second is the Person Case Constraint, a gap with repairs through otherwise unavailable syntactic structures and novel agreeing forms. Last and in most detail are taken up morphological gaps in many western dialects, where some or all 1st person plural + 2nd person combinations are missing. In their place are often found stopgaps, forms that fail to code features of the arguments involved, including forms surface-identical to impersonals. A concern throughout is the potential of gaps and responses to them in helping understand the boundaries and interactions of morphology, syntax and interpretation. Keywords:  agreement; gaps; repair strategies; morphology; syntax

1.  Introduction to gaps Gaps are a curious property of natural language: expressions that one would expect on the basis of others, but missing. Gwen’s tunic might evoke I want one too, but not her trousers *I want ones too; one can get less fruit or fewer vegetables, but not ask for ??less/*fewer oats. These absences are striking in inflectional paradigms.

*  I am grateful to P. Albizu, B. Fernández, M. Jouitteau, and A. Nevins for discussion, to B. Fernández and an anonymous reviewer for helpful comments, and to the editors for their long labours. This research has been partly supported by Spanish Ministry of Economy and Competitiveness grant FFI2014-51878-P9. The research leading to these results has also ­received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement n0 613465.

doi 10.1075/lfab.13.06rez © 2016 John Benjamins Publishing Company

 Milan Rezac

There is underwent and embargoed, and there was forwent, but for many forgo no longer has a past; stridden too is mostly gone, though hid and ridden and glided are with us still. Other domains of language as well riddle the field of expression with gaps. The typical island constraint gives rise to syntactic gaps: *How many mechanics did they ask if __ fixed the cars? “is a fine thought, but it has to be expressed by some circumlocution” (Chomsky 2013). A fascinating and difficult aspect of gaps is how language responds to them, I-language and E-language. It is characteristic of most gaps that I-language does not see them. There is no form, structure or interpretation licensed only by the absence of *forwent or *ones: there are no repairs. E-language is another matter. To get around gaps, speakers resort to circumlocution, to independently available expressions of similar meaning – say I went without dinner or I need a pair too. It seems reasonable that in time, circumlocutions would grammaticalise to fill gaps. Yet though there are grammaticalised stopgaps, nonce or repurposed syntactic structures and morphological formations, whenever we will meet one, it will turn out never to have been available as a circumlocution. These themes of gap and response, their natures and their sources, are the business of this article. The domain is the rich and intricate agreement system of the Basque finite verb. Gaps may be defined by phi-features, and bar agreeing forms, like *amn’t, structures or interpretations, *Either she or I __ lying right now, aren’t __? (Pullum 2013; McCawley 1998), or uses, (*)John gave the book to me, who speak(s) French (Morgan 1972, cf. Sobin 1997), (*)One of my friends’ mother broke a vase (Green 1971, cf. Fodor and Inoue 1994). I take up first interpretive gaps due to Condition B. In Basque, they have become grammaticalised beyond their interpretive source, and that grammaticalisation is a characteristic of agreement system in one sense of the term. Next come gaps due to conditions on syntactic dependencies. One of them is a rare gap with syntactic repairs: the Person Case Constraint. Finally, and in most detail, are taken up morphological gaps in the agreement paradigms of many western dialects, where some or all 1p+2 combinations do not exist, and instead morphological stopgaps similar to aren’t for *amn’t are found: a phenomenon with much potential to shed light on the boundary between syntax and morphology and the ways to find it. 2.  The Basque agreement complex Basque finite verb agreement is part of an agreement complex, a morphological word also reflecting mood and tense (for overviews, see Laka 1993a; Albizu 2002; Hualde 2003; Ariztimuño 2013). In the usual analytic formation, (1), the complex is built on an auxiliary root accompanying a nonfinite form of the verb. The



Gaps and stopgaps in Basque finite verb agreement 

form of the root can depend on just about any information coded in the complex, including the case and phi-features of agreement controllers. In the synthetic formation, the complex is built in a similar manner on a verbal root; it is available with be (mostly syncretic intransitive auxiliaries) and have (syncretic with transitive auxiliaries), and more sporadically with a handful of other verbs.1 (1) a. (Nik zuri haiek) eman d-i-zki-zu-t    I.E you.D they.A given X.PRS-√-pA-2sD-1sE ‘I gave them to you’ 1sE-2sD-3pA b. (Nik zuri haiek) eman n-i-zki-zu-n    I.E you.D they.A given 1sE-√-pA-2sD-PST ‘I gave them to you’ 1sE-2sD-3pA-PST

Each suffix agrees for person and number agreement with the ergative (-t) or the dative (-zu-).2 The prefix agrees in person and number with the absolutive if 1st/2nd person, or else reflects tense and mood (d-). The plurality of the absolutive is indicated by one or more affixes in a variety of positions (-zki-). Some phenomena modify these correlations. In (1b), ergative displacement ED in the past tense shifts control of the prefix to a 1st/2nd person ergative if there is no 1st/2nd person absolutive, but never affects the absolutive plural morpheme. Agreement is usually obligatory if allowed. A morphological breakdown of the agreement complex will often not be necessary. In that case the complex is glossed by agreement controllers alone in the order ((E-)D-))A, as in the translations in (1). 3sA systematically uses the morphology that one would expect if there were no A controller, so in glossing examples, 3sA is bracketed unless there is a 3sA argument, as in 3sE(-3sA). Each of these controller-form relationships is subject to dialectal variation. A description of the Basque dialectal situation and of variation in verbal morphology is found in Hualde (this volume). Table 1 illustrates the variation found in agreement. A quick impression may be had by putting the forms side by side, e.g. nazkitzun – nitizu – dotzutesen, not unlike (you) are – art – be, but pervasive. ­Standard Basque or Euskara batua (EB) has often codified older forms and

.  Glosses for agreement combine 1/2/3 person, s/p number, E, D, A ergative, dative, absolutive case (also glossing case on nouns), save that 20 is the familiar 2nd person singular followed by m/f gender if need be. Other glosses are X default prefix, √ root, TM theme marker, and Leipzig Glossing Rules glosses. Translations distinguish 2s you, 2p ye, 20 thou. Translations of citations and corpus examples are mine. .  In allocutive forms, an additional suffix agrees in gender with the 20 addressee, and in some varieties also in politeness level (Oyharçabal 1993).

 Milan Rezac

f­ ormations, so it is a useful point of reference for dialectal developments. In what follows, unprovenanced examples are EB. Table 1.  Dialectal variation (from Yrizar) Variety†

Traditional analysis

Divergences from EB‡

EB

n1sE-i√-zkipA-zu2sD-nPST

L Urdax 1

n1sE-aTM+Ø√-zkipA-tzu2sD-nPST

√; TM; 2sD form

L Urdax 2

z2sD-inTM-tpD-u√-zkipD-da1sE-nPST

D treated like A; no ED

B Lekeitio 1

n1sE-eTM-u√-tzu2sD-sepA-nPST

√; TM; pA form, pos.

B Lekeitio 2

dX.PRS-oTM+Ø√-tzu2sD-te1sE-sepA-nPST

√; no ED

B Bermeo

-tsu2sD-te1sE-nPST

√; no ED; no pA; affixal

N Esteríbar

n1sE-itpA-i√-zu2sD

pA form, pos.; no PST

†EB ‡√

Standard Basque, L Lapourdian, B Biscayan, N Navarrese root allomorphy; TM theme marker allomorphy; pos. positional variation

Essentially for reference, Table 2 gives the basic agreement prefix, suffix, and absolutive pluraliser forms in EB (Hualde 2003).3 Table 2.  EB agreement morphemes and pronouns Prefix

Suffix

Pronoun (A, E, D)

3s

Ø-‡

E -Ø(-)  D -o(-)

hura hark hari

3p†

Ø-‡

E -te(-)  D -e(-)

haiek haiek haiei

1s

n-

-da-/-t

ni nik niri

20m

h-

-ka-/-k

hi hik hiri

20f

h-

-ka-/-n

1p†

g-

-gu(-)

gu guk guri

2s†

z-

-zu(-)

zu zuk zuri

2p†

z-…-te(-)

-zue(-)

zuek zuek zuei

a pluraliser if absolutive: -(i)t-, -tza(-), -z(-), -zki(-), -de or default prefix: present d-, past z-, hypothetical l-, imperative b-/Ø-

†Controls ‡ED

.  The complexity of the 2nd person is mostly historically straightforward: the 2nd person familiar, 20 or thou, is the old 2s; 2s, you, is the old 2p so it combines with pluralisers; 2p, ye, is a repluralised 2s. The unique presence of gender distinction in 20 remains mysterious.



Gaps and stopgaps in Basque finite verb agreement 

3.  Interpretive phi-gaps: Condition B By and large, interpretation imposes gaps on pairings of form and meaning for the agreement complex, not on the availability of forms. For Hei described herk to them*i+k, there is a candidate agreeing form, deskribatu die 3sE-3pD-3sA, but Condition B bars its use with this meaning. There is nothing here specific to Basque. In one case, however, interpretation appears to rule out the forms themselves. Condition B bars full covaluation in local domains, *I chose me, and reflexives step in, I chose myself. It also bars partial covaluation, I asked us ?(??each) a question, and then there is only a gap, because reflexives need exhaustive antecedence, *I asked ourselves a question. The partial covaluation ban is of some importance, since it cannot be derived from preference for reflexives over pronouns (Lasnik 1981, Reinhart and Reuland 1993: sec. 4, Kayne 2002: 143–6, Safir 2004: 3.3.1, 2013: 15.4.2). It does not matter here whether the partial covaluation ban should be attributed to the same principles as the full covaluation ban, only that it c­ reates gaps. In English, there has been debate about the existence of the partial covaluation ban (Büring 2005: 9.2–4, Safir 2004: 3.3.1). It is known to be far stronger in certain clitic systems, such as French (2) (clitics in italics). (2) a. Vous deux t’avez %(*chacun) choisi pour répondre à cette question Ye two have %(*each) chosen thee to answer that question b. Mai, Mael et Yann lMai’ont %(*chacun) choisi pour répondre à cette question Mai, Mael and Yann have %(*each) chosen herMai to answer that question(Rezac and Jouitteau 2015)

The literature gives such cases as ungrammatical (Blanche-Benveniste 1975: 213; Morin 1978: 347; Kayne 2002: 143–6; Schlenker 2005; Rooryck 2006). Some speakers tolerate similar examples with a collective reading of the plural argument, but not when there is an entailment of reflexivity, as here with chacun ‘each’ (cf. Reinhart and Reuland 1993: 677). The clitic system of French is a useful point of comparison for Basque, as it stands between strong pronouns and agreement systems, descriptively speaking: it looks morphologised in idiosyncratic constraints imposed by one clitic form on another, but it allows clitic combinations that only arise in multiclausal structures and unlikely to be lexicalised (Rezac 2010a). In Basque, both the full and partial covaluation cases of Condition B are irredeemably bad for agreeing arguments. For 3+3 combinations, the forms exist, as du, ditu 3pE-3s,pA, but they cannot get the wholly or partially covalued meaning. For 1+1, 2+2 combinations, there are no forms: between h20A-au-t1sE

 Milan Rezac

1sE-20A and n1sA-au-k20E 20E-1sA, one would expect *n1sA-au-t1sA 1sE-1sA, and so for *naugu 1pE-1sA, *gaitut 1sE-1pA, *didagu 1pE-1sD-3sA. The morphology is transparent and speakers amuse themselves in constructing the forms; but they are sharply ungrammatical, unlike other nonce formations discussed later.4 Instead, partial covaluation is ineffable. Full covaluation is coded by the 3rd person reflexive X’s buru- ‘X’s head’ and reciprocals like elkar ‘each.other’ (Artiagoitia 2003; Etxepare 2003b; Albizu 2001; Hualde, Elordieta and Elordieta 1994; Ortiz de Urbina 1989). It is natural to look to Condition B for the absence of these forms, since Condition B degrades similar combinations of full pronouns and clitics. Oyharçabal (1993: 102) proposes this reduction for full covaluation in Basque, and Rhodes (1993) even for partial covaluation in Ojibwa. However, more seems needed for the agreement system of Basque. Arregi and Nevins (2011: 1.4.5.4) show that Condition B applies in nonagreeing clauses, but less strongly than in agreeing forms (citing X. Artiagoitia p.c.). Adapting their Ondarroa example to EB, their starred nonfinite (3)a is more acceptable than the strictly impossible finite (3)b (mere overlap like guk neu ‘we.E I.A’ is even better). (3) a. *[Nik neu maite izatea] nahi du    I.E I.A love being want 3sE(-3sA) *He wants me to love me5 → a’. [Nik neure burua maite izatea] nahi du    I.E my head.A love being want 3sE(-3sA) ‘He wants me to love myself ’ b. **Nik neu maite nahi naut     I.E I.A love want 1sE-1sA **I want to love me

.  A. Irurtzun and M. Duguine had first made me aware of the recreational use of these ungrammatical forms. B. Fernández points out the Ondarroa poet Leire Bilbao’s use of them in Ezkatak (2006: 22, Zarautz: Susa), nicely encapsulated in the critique “In this Ezkatak we also find poems that play at breaking the laws of the language: Ez daukat ezer, ez naukat [‘I have nothing, I do not have me’ have-1sE-1sA]; Ez daukazu ezer, Ez zauzkazu [‘You have nothing, You do not have you’ have-2sE-2sA]” (Aritz Galarraga, “Buia itsaso erdian”, Gara, 25.3.2006). Cysouw and Fernández (2012: 775) give as playful but ungrammatical 1sE-1pA Eta zain, Etxarrin oroitzen gaitut, eta berdin zait Euskaltzaindiak nire oroitzapenak ez onartzea, nik, zu eta ni, Etxarrin oroitzen gaitut “And wait, I remember us in Etxarri, and the Basque Language Academy not accepting my memories does not matter to me, I, you and me, I remember us in Etxarri” (Xabier Silvera, “Egin topa, haiengatik!”). .  Original Ondarroa [Nik *neu/√neure buru matxe ixati] nai dau. I have had similar contrasts reported for EB.



Gaps and stopgaps in Basque finite verb agreement 

→ b’. Nik neure burua maite nahi dut I.E my head.A love want 1sE(-3sA) ‘I want to love myself ’

Analysing agreement as clitics, Arregi and Nevins propose that the agreeing forms violate “a restriction that is specific to clitic combinations, perhaps related to a similar restriction found in Spanish clitic combinations (Perlmutter 1971: 41–45).” However, the ungrammaticality of the partial covaluation forms exceeds that of Romance clitics. Cysouw and Fernández (2012: 774–7) contrasts the judgments of a native Basque-Spanish bilingual in the two languages for overlapping reference combinations: Spanish a mí me excluimos de la expedición “We excluded me from the excursion” and similar examples “feel a bit strange”, but the Basque counterpart *txangotik baztertzen naugu is “completely wrong”, and so for *izozkiak erosi nigun beside nos compré unos helados “I bought us some icecreams” and likewise for 2nd person. They conclude that there is a contrast between the absence of overlapping reference forms in agreement systems (citing also Belhare, to which Rhodes’s 1993 study of Ojibwa may be added), and their relative acceptability in clitic or weak pronoun systems in (Spanish, French, Dutch, and Serbo-Croatian).6 Thus while Condition B contributes to the absence of *naugu 1sE-1sA, there is more. With Cysouw and Fernández (2012), we seem to have a difference between systems traditionally described as agreement and clitics. It encompasses even those Basque morphemes most frequently viewed as clitics, the dative and ergative suffixes, as in *di-da1sD-gu1pE 1pE-1sD-3sA. This is of great theoretical interest. Much work has viewed some or all Basque agreement morphemes as clitics (Laka 1993a; Rezac 2006; Preminger 2009; Arregi and Nevins 2011: Chapter 2; Etxepare 2012); other work has viewed Romance clitics, especially in clitic doubling, as Basque-like agreement (Suñer 1988, Franco 1993, Ormazabal and Romero 2013 on ­Spanish; Heger 1966, Lambrecht 1981, Miller and Sag 1997 on French). There are interesting arguments one way or the other, like the presence or absence of tense-conditioned allomorphy, but their theoretical grounding is unclear. A sharp difference in the status of overlapping reference forms has potential in this debate – as a touchstone in telling one type of system from another, and as a window on what makes the two types different, that is on one theoretical notion of clitic and agreement among others (cf. Rezac 2010c). This contrast for Condition B between Basque agreement and Spanish clitics could be construed as syntactic or morphological. Basque morphology might simply not have *naugu the way that English does not have *amn’t, *stridden, *forwent.

.  Cysouw and Fernández do not check for reflexive entailment, so their larger conclusion that there is no grammatical ban on overlapping reference at all does not follow.

 Milan Rezac

In a syntactic approach, Basque agreement might involve structures smaller than Romance clitics, and that might bar collective readings of plurals needed to obviate the partial covaluation ban. Here is one way to work this out, using a common theory of plurality (Barker 1992): i Plural pronouns have [plural] whereby they denote pluralities. ii Higher up there is a lexical item G that turns pluralities into group atoms, so that semantically we chose me can be our group chose me. iii Condition B bars reflexive entailments and so access to the me in we, which pluralities but not groups give, as in we/*our group knew each other. iv Basque agreement affixes spell out structures too small to include G, Romance clitics can spell out larger structures and so avoid Condition B. This story needs for agreement affixes to constrain the interpretation of controllers, since partial covaluation is out even when there is a full pronoun controller that can evade the constraint in nonagreeing clauses, as in (3). This would be so if agreement affixes are interpreted as arguments covalued with their controllers.7 One way to tease apart the syntactic and morphological accounts is to examine cases of Condition B for which neither morphology nor interpretation can be responsible. Arregi and Nevins find that Condition B in Basque operates even when English escapes it in (4), under pragmatic coreference rather than semantic covaluation (Reinhart 2006; Heim 1998, 2008; for marginality in English, ­Roelofsen 2008: 46).

(4) Nobody loves me/Jon: Mikel doesn’t love me/Jon, Miren doesn’t love me/ him, and…

a. Neuk be es dot nire buru matxe I.E even not 1sE-3sA my head.A love ‘Even I don’t love myself ’ a’ *Neuk be es nat (ni) matxe 1sE-1sA (I.A) Even I don’t love me

.  Cf. Rooryck (2006) for French, and Den Dikken, Lipták and Zvolenszky (2001) for ­Hungarian, analysing of a subset of overlapping reference combinations that contrast with others in escaping the ban on partial covaluation, discussed below, by positing a complex structure for plurals in question. This is one out of a gamut of options. Suppose alternatively that multiple identical features within a terminal cannot be linearised, e.g. {…[1]…[1]…} (Béjar and Rezac 2009: 58); then Basque *naugu 1sA-1sE would run into this problem if it reflects the Agree-valued features pooled on a single head, while French nous me ‘we me’ would not if displaced pronouns, correlating with other differences (Rezac 2010c).



Gaps and stopgaps in Basque finite verb agreement 

b. Jonek be es dau bera buru matxe Jon.E even not 3sE-3sA his head.A love ‘Even Jon doesn’t love himself ’ b’ Joneki be  es dau (bera*i) matxe 3sE-3sA (he.A) ‘Even Jon doesn’t love him’ (Ondarroa, Arregi and Nevins 2011)

Here there seems to be a problem that is not due to interpretation, given English, or morphology, given the availability of dau 3sE-3sA. That leaves something about the syntax of Basque agreement to explain the problem. Continuing in the same vein as above, perhaps pragmatic coreference needs extra content in a pronoun, and this content is unavailable to agreement affixes. In Romance clitics, the coreference exception is more difficult to obtain than in English (Baauw and Delfitto 2005 for Spanish as mostly out, Hamann 2011 as good, Schlenker 2005: 73–5 is more nuanced and relates the ban on partial covaluation). This suggests the cline English strong pronouns > Romance clitics > Basque agreement. The cline lends itself to reification through increasing syntactic structure and correlated interpretive options, along the lines of Cardinaletti and Starke (1999). There might be correlated clines elsewhere; the syntactic Person Case Constraint discussed later seems to have the strength cline Basque agreement > Romance verb-attached ­clitics > Czech second position clitics or Germanic weak pronouns (Rezac 2010b). Condition B in English and French creates gaps that have no repairs: in the absence of I helped us ?(*each) realise this, one cannot press into service the X-self reflexive, *I helped ourselves. Remarkably, Artiagoitia (2003: 623–4) (5) and Arregi and Nevins (2011: op.cit.) observe that the X’s buru- ‘X’s head’ anaphor does appear in partial covaluation environments. On the other hand, Cysouw and Fernández (2012: 775–6) claim the contrary with an example virtually identical to Arregi and Nevins’s, so there is variation.8 (5) Geure burua aipatu dut our head.A mentioned 1sE-3sA [The problems of the people out there resemble those we have or might have …] I have mentioned ourselves, [and I don’t know if I said it right, …]  (Mitxelena, cited in Artiagoitia 2003: 623)

The availability of X’s burua in partial covaluation looks like the emergence of an otherwise impossible syntax or interpretation, a repair of the sort we will see for

.  The examples are: Arregi and Nevins, Ondarroa Basque, Nik geure buru ikusi dot (>rot) ispillun “I saw us in the mirror”, beside Cysouw and Fernández, EB, *Nik ispiluan geure burua ikusten dut “I see us in the mirror”.

 Milan Rezac

the Person Case Constraint. There are two hints that it is not: one in Basque itself, another in French. First, X’s burua is less grammaticalised as reflexive than X-self. Artiagoitia (op.cit.) and Oyharçabal (2003) find contexts where X’s burua need not have a clausemate antecedent but X-self does. Artiagoitia (2003: 621–3) also observes that the coding of X and the plural marking on buru ‘self ’ vary in ways similar to inalianable possession, cf. They each saw their self/selves, whereas for X-self they are fixed, They each saw themselves. So perhaps in partial covaluation, X’s burua is not an anaphor; Oyharçabal (2003) proposes this for other uses of it. Den Dikken, Lipták and ­Zvolenszky (2001) reach a similar conclusion for Hungarian. It too allows reflexives in SG→PL though not *PL→SG combinations like I represented ourselves. Observing that the reflexive is morphologically core-1p ‘our core’, the authors suggest that “If … syntactic structure of Hungarian reflexives is that of a possessed noun phrase, then [the inclusive use] is syntactically parallel to something like ‘I represent/vote for our friend’” (p. 148).9 The same idea works for Hausa, where “a singular subject can take a plural reflexive if the subject is included in the referential group” in counterparts of Ladi criticised “themselves”, Today I embarrassed “ourselves”, where the reflexives are as in Basque X’s head (Newman 2000: 524; Dixon 2003: 147, 164). Yet Cysouw and Fernández (2012: 772) report an I-ourselves case in Even with a direct object reflexive that seems morphologically simplex. The second hint for what might be going on in (5) comes from French. The perfect auxiliary is ordinarily avoir ‘have’, but être ‘be’ if there is an object clitic reflexive to the subject. Rooryck (2006) finds a class of partial covaluation combinations that are fine when others are not. The most robust is 1s→1p, Je nous ai chacun inscrit “I registered us each”, much better than 1p+1s or 2s+2p.10 Just for 1s→1p, être is sometimes found for avoir, (6). However, this has not been studied for speaker judgments (they are shifty), nor extension beyond 1s→1p (it does not seem to). (6) a. une promesse que je nous suis faite a promise that I usclitic am [= have] made  (Balzac, cited in Haas 1909: 191) b. Je nous suis réservé un taxi I usclitic am [= have] reserved us a taxi  (Queneau, cited in Blanche-Benveniste 1975: 223n1)

.  The generalisation is stated for overlapping context but only illustrated with 1s→1p. .  Even je nous is rejected by Morin (1978: 347) and treated as a jest by Blanche-Benveniste (1975: 45n1); but research leading to Rezac and Jouitteau (2015) revealed speakers confirming Rooryck’s je nous even if nous is distributed by chacun.



Gaps and stopgaps in Basque finite verb agreement 

So there is something more “reflexive” about the 1s→1p combination than about other partial covaluations, and that might bear on (5). The snag is understanding the relevant notion of reflexivity.11

4.  Syntactic phi-gaps: Locality Syntax is the best-studied domain of gaps, and it often affects agreement systems. Even the poor agreement system of English has gaps of the sort in (7), and sometimes syntax has been seen as their source (Pullum 2013: 512; McCawley 1998: 506; Sobin 1997: 320–1).

(7) Either she or I __ lying right now, aren’t __?

This section takes up a gap created by island constraints in the agreement system of Basque. All island violations can be thought of as gaps, ruling out certain form-meaning pairs in paradigms of good pairings. The constraint of interest bars agreement with a dative past an intervening absolutive in ECM configurations, making unavailable certain interpretations for the dative, and lexical items that need those interpretations. The effects of our constraint have been studied in French. Consider first raising configurations. (8a) embeds the small clause α built on the predicate adjectives antipathique ‘antipathic’ with an optional dative experiencer; the dative must cliticise if it is 1st/2nd person. (8b) has an unaccusative VP β with a dative possessor; it too must cliticise. The subjects raise to agreeing nominatives, and the dative clitics attach to the finite verbs. (8) a. Tout le mondei (mek) sera/paraîtra [α ti antipathique tk] Everyone    me will.be/seem antipathic [= to me] b. Tout le mondei mk’ est [β ti tombé dans les bras tk] Everyone me has fallen into the arms. [= my arms]

.  In examples of the type (i) No one chose me; only I chose me or (ii) Ii dreamt that Ii was Gwenk, and that Ik kissed mei, the auxiliary in French must be the reflexive auxiliary être, though in English the pronoun prefers to be disjoint (Roelofsen 2008: 4.7; Heim 1998: n15; Anand 2007: 29n19; Arregui 2007; Safir 2004: 4.3). That makes it seem that reflexivity for auxiliary choice and self-anaphora does not coincide. However, in French (i) at least strongly prefers reflexive to disjoint object pronouns in 3rd person, the sole to make this distinction overtly (Schlenker 2005: 73). The matter is complicated by varieties that seem to make the me/ myself distinction in (i) through auxiliary choice (Morin 1978: 361n8).

 Milan Rezac

When α, β are embedded in an ECM construction, their lower subject, now accusative, cliticises upstairs or remains low as a nonclitic, and the dative cannot cliticise: (9) a. Tout le monde (*me) croit Gwen antipathique Everyone    me believes Gwen antipathic b. Tout le monde (*me) la croit antipathique everyone    me her believes antipathic (10) a. *Tout le monde m’ a vu tomber Gwen dans les bras.  everyone me has seen fall Gwen into the arms b. *Tout le monde me l’ ai vu/cru tomber dans les bras  everyone me her has seen/believed fall into the arms

There is agreement that uncliticisability of the dative in ECM reflects constraints on syntactic dependencies. In generative grammar it has been chiefly attributed to the Specified Subject Condition: in raising the subject does not intervene in dative cliticization because it raises, in ECM it does because it stays low (Kayne 1975: 4.5–6; Emonds 1999; cf. den Dikken 2006, 2007; Gallego 2010 for subjecthood barriers in current frameworks).12 The same pattern seems to characterise the interaction of absolutive and dative agreement in Basque. Basque datives come in “high” and “low” configurations (Albizu 1997a, 2011; Etxepare and Oyharçabal 2013, Fernández, Ortiz de Urbina and Landa 2009; Fernández and Ortiz de Urbina 2010; Fernández 2011; Rezac 2008b, 2011: chap. 5, Rezac, Albizu and Etxepare 2014). High datives are above an absolutive coargument for purposes like anaphor binding, and must control agreement in finite/agreeing clauses. Low datives are below the absolutive. In eastern dialects, they do not agree. In western dialects, they agree when they can, but often permit nonagreement otherwise (cf. for causatives Ortiz de Urbina 2003a, variation in Trask 1981, and for differential object marking Fernández and Rezac this volume). Basque allows only one instance of dative agreement, so low but not high datives are possible when another dative controls agreement. Certain interpretations like experiencers need high datives.13 .  I give a simplified description of the French data and its analysis, which have both been a battlefield: see Rezac (2011: Chapter 4) and literature there, esp. Postal (1983, 1984, 1990). .  Distinct from nonagreing low datives is optionality of dative agreement in multipredicate structures. It affects both high and low datives and reflects restructuring versus richer nonfinite structures (Ortiz de Urbina 2003c). This may be the explanation of alternations like begira egon zaio/da ‘She has been looking her.D’ 3sD(-3sA) (Fernández 2013), where agreement of the dative is free even in western Basque. Some instances of nonagreeing low datives might fall to a similar analysis: for instance, causatives with a dative causee might introduce more structure above the causativised predicate than ones without.



Gaps and stopgaps in Basque finite verb agreement 

Fernández (2011) shows that the two types of datives are found with adjectives. Adjectives like leial ‘loyal’ allow their argument to use the low dative structure, and so do not require dative agreement. Others like atsegin ‘pleasant’ require the high dative structure and agreement. The contrast comes out in contexts where another dative already controls agreement, (11)–(12): the low dative of leial is fine, the high one of atsegin is out, with no “repair” such as otherwise unavailable nonagreement.14 (11) a. Niri Jon [hizkuntzari leiala] iruditzen zait me.D Jon.A    language.the.D loyal seeming 1sD-3sA ‘Jon seems to me to be loyal to the language’ b. *Niri entretamendua [Joni atsegina] iruditzen zait   me.D exercise.A   Jon.D pleasant seeming 1sD-3sA ‘Exercise seems to me to be pleasant to Jon’ (Fernández 2011) (12) a. Niri ikasleak [bere irakasleari leialak] iruditzen zaizkit me.D students.A    their teacher.the.D loyal seeming 1sD-3pA b. Niri ikasketak [(*Joni) atseginak] iruditzen zaizkit me.D lessons.A    (*Jon.D) pleasant seeming 1sD-3pA  (R. Etxepare, p.c.)

Armed with high-dative atsegin and high/low-dative leial, consider dative agreement past an absolutive in ECM. The ECM subject controls matrix absolutive agreement, but bars agreement with the dative argument of the adjective. This is fine with leial but barred for atsegin. There is no repair: atsegin simply cannot have an experiencer in ECM.15 (13) VECM-φABS-(*φDAT) [… DPABS … DATDAT …]

.  Supporting contrasts between the atsegin and leial adjective types are reflexive binding, Mikeli bere burua atsegin/*leial zaio, ‘Mikel.D himself.A has been pleasant/*loyal’ versus Mikel bere buruari leial zaio ‘Mikel.A himself.D has been loyal’ (data thanks to B. Fernández p.c.), and control into nominalizations (Fernández 2011). Etxepare (2003a: 168) contrasts the PCC found with gustatu ‘please’ but not atsegin izan ‘be agreeable’, which makes sense if the theme gustatu is generated below the high dative but that of atsegin izan above it, that is predicated of an AP extended to include the high applicative (Rezac 2008b: 73n7; cf. Diesing 1992). Of relevance may be that while in French dative pronouns need to cliticise, this is weaker with fidèle ‘faithful’ than with antipathique (Rezac 2011). .  I am grateful to P. Albizu and R. Etxepare for discussion of this matter for drafts of Rezac (2013) and Rezac, Albizu and Etxepare (2014).

 Milan Rezac

Basque has a couple of ECM structures. One is nonfinite perception complements (Arteatx 2007, 2012; Rezac, Albizu and Etxepare 2014). Embedding be + adjective contrasts leial and atsegin as described.16 (14) a. Ikasleak zuri leialak izaten antzeman ?ditut / students.the.A you.D loyal being perceive   1sE-3pA / *dizkizut *1sE-2sD-3pA

‘I perceive the students to be loyal to you’

b. Ikasleak zuri atseginak izaten antzeman *ditut / students.the.A you.D pleasant being perceive 1sE-3pA / *dizkizut *1sE-2sD-3pA

‘I perceive the students to be pleasant to you’

A second ECM structure is so-called transitive predication, where Basque have predicates take a small clause complement in ECM and relate it to their ergative subject in a vague relation like an experiencer (Etxepare 2003b; de Rijk 2008; E ­ txepare and Uribe-Etxebarria 2012; Fernández and Rezac this volume). One type, with eduki ‘have’ in EB, seems similar to the English type She has all four grandparents alive/*clever. This type takes stage-level predicates like begira ‘looking’, adi ‘attentive’. Both allow low datives and so are legitimate in ECM, but only without agreement: (15) a. Jon zuri begira egon zaizu/da Jon.A you.D looking be 2sD-3sA/3sA ‘Jon is looking at you’ b. Jon zuri begira eduki {duzu / du Martxelek / Jon.A you.D looking had 2sE-3sA /3sE-3sA Martxel.E / *dizu}    3sE-2sD-3sA

‘You have Jon looking at you’

(A. Elordieta p.c. to B. Fernández)

The other type of transitive predication, with *edun ‘have’ in EB, is immune to the English restrictions. Here we may test the individual-level leial and atsegin. Agreement is out, leaving acceptable a low nonagreeing dative with leial, but not atsegin: (16) a. Mikeli leiala duzu/nauzu (*diozu/*naiozu) Mikel.D loyal have-2sE-3sA/1sA (*have-2sE-3sD-3sA/1sA) ‘She is/I am loyal to Mikel’ (lit. ‘You have me/her loyal to Mikel’)

.  I give this as the simplest illustration of datives under ECM, but unlike the following, even the good example is quite unnatural, perhaps owing to the difficulty of individual-level predicates as perception complements (Felser 1999: 45; Arteatx 2012: 413).



Gaps and stopgaps in Basque finite verb agreement 

b. *Mikeli atsegina duzu/nauzu (diozu/naiozu)    Mikel.D kind have-2sE-3sA/1sA (have-2sE-3sD-3sA/1sA) ‘I am/She is kind to Mikel’ (lit. You have her/me kind to Mikel)  (B. Fernández, p.c.)

Finally, in (17) a perception verb ECM embeds a nonfinite clause based on an unaccusative with a high dative of possession. The dative can agree in a simple clause, because it is a high dative, but not in ECM: (17) a. Mireni giltzak eskuetatik erori zaizkio/* dira / Miren.D keys.the.A hands.the.from fallen 3sD-3pA/* 3sA / *zaio *3sD-3sA

‘The keys fell from Miren’s hands’

b. Giltzak Mireni eskuetatik erortzen entzun ditut / keys.the.A Miren.D hands.the.from falling heard 1sE-3pA / *dizkiot    1sE-(*3sD-)3pA ‘I heard the keys fall from Miren’s hands’ (does not entail #Giltzak entzun ditut ‘I heard the keys’)  (from drafts of Rezac, Albizu and Etxepare 2014)

In all these Basque ECM constructions, a Specified Subject Condition story would suppose that the absolutive is at some point above the dative and a subject in the relevant sense. For apparently bare adjectives in ECM, (14)–(16), this is so if the ECM complement is a small clause predicating the absolutive of the adjective’s AP containing its dative argument, a usual analysis. For the nonfinite clause in (17), evidence for absolutive > dative comes from word order. In a finite clause, the absolutive satisfies the EPP of T by rich agreement, but the dative is higher as an XP, and dative > absolutive is a (or the) neutral word order (Rezac 2008b). In ECM, absolutive > dative is the sole neutral word order. Possibly, the absolutive must satisfy the EPP of nonfinite T in ECM complements phrasally because there is no analogue of rich agreement (cf. Rezac, Albizu and Etxepare 2014). Then the absolutive intervenes for agreement with the dative.17

.  It is not clear whether the dative remains legitimate in the infinitive without agreement because infinitives have a richer structure than the predicate adjectives seen so far, or because we have to hand a possessor dative whose point of origin is lower than the experiencer dative seen with atsegin (cf. Albizu 2011). We cannot compare the two dative types directly, as there are no adjectives with possessor datives, and perception verbs do not embed psych-verb ­infinitives due to constraints on event structure (Felser 1999).

 Milan Rezac

The ban on dative agreement past an intervening absolutive is an example of syntactic constraints that create gaps in available pairings of agreeing forms and interpretations. Other examples include the limit of a single agreeing dative per clause, and constraints on remote agreement studied in Etxepare (2003a, 2006, 2012), Preminger (2009).18 5.  The Person Case Constraint The Person Case Constraint PCC of Bonet (1991, 1994) excludes in Basque combinations of 1st/2nd person absolutive and dative agreement in certain structures: *eraman zaizkiot “I brought you to her 1sE-3sD-2sA”, *gustatzen zatzaizkit “You please me 1sD-2sA”. I set out here briefly the syntactic character of the constraint and its repairs, from Rezac (2008b, 2009, 2011); key literature on the constraint in Basque includes Albizu (1997ab), Laka (1993a), Oyharçabal and Etxepare (2012), Arregi and Nevins (2011). Three aspects of PCC are extraordinary. First, while with transitives the result is the absence of E-D-1/2A agreeing forms, with unaccusatives D-1/2A forms exist but cannot be paired with certain syntactic structures, bespeaking the syntactic rather than morphological character of the constraint. Second, some of the structures incurring the constraint are repaired by otherwise impossible ones. That makes for an apparent case of reference-set computation: one syntactic structure depends for its legitimacy on the badness of another, existing only as last resort. One type of repairs moreover results in a morphological formation that does not

.  The nature of the restriction to one agreeing dative is unclear, but it looks syntactic. Perhaps there is not enough structural space in a single functional architecture for multiple high datives. Morphology allow for multiple dative agreement suffixes if they are controlled by one dative in agreement doubling and tripling like (Niri) erraiten dei-TA(-DA-)zü-T 2sE-1sD(-3sA) ‘You say it to me’ (Iruri, Zuberoa, Fernández and Albizu 2006). In French and Spanish, multiple dative clitics also typically fail to combine: French Gawain me (*lui) semble reconnaissant (à Gryngolet) ‘Gawain seems to me.D to be grateful to *them.D/to Gryngolet’, Merlin te le (*lui) mettera dans les bras ‘Merlin will put him in your.D arms for him.D’ (Rezac 2010a), Spanish Se la/*le permití escribir ‘I permitted him.D to write it.A/*her.D’ (Rivas 1977). Exceptions like French %Ellei me les luii/k fera envoyer ‘Shei will make me send them (the packages) to heri/k’, binding suggests a more complex structure with multiple Condition A/B domains (see Postal 1984: 125–6, Rezac 2011: chap. 4). Basque also suggests that syntax rather than morphology is at issue, insofar as multiple high datives remain out when the foregoing examples are transposed to nonfinite, nonagreeing clauses, *Niri ikasketak Joni atseginak iruditzea… “For exercises to seem to me.D pleasant Jon.D…” (B. Fernández p.c.; for the status of otherwise agreeing arguments in nonagreeing clauses, see Laka 1993a, Rezac 2011: 190n9).



Gaps and stopgaps in Basque finite verb agreement 

otherwise exist. Last, the constraint is not categorical across the history and varieties of the language, so it must be parameterizable. The canonical context of the PCC in Basque are transitives. No living variety is known to have any E-D-1/2A forms, but they did exist (a convenient gathering is Yrizar 1981: 36f.). A rich set of exceptions is found in Leizarraga’s sixteenth century translation of the Bible into his literary Lapurdian, like redimitu garauzcac Iaincoari ‘thou redeemedst us to God’ 20mE-3sD-1pA (Oyharçabal and Etxepare 2008). There are a handful of such forms in nineteenth century eastern writers, like Duhalde’s Eman giotza ‘He gave us to him’ 3sE-3sD-1pA, and another couple in Zavala’s grammar of western Basque, like eskiniten gautsazac ‘Thou offerest us to him’ 20m-3sD-1pA. Most explicitly, there is the grammarian Azkue’s report that “An old woman of Dima and another of Zeanuri gave me as common” forms like Zuk ni berari eroan nautsazu ‘You brought me to him’ 2sE-3sD-1sA. Basque is not alone in a categorical PCC with exceptions in literature (Mohawk, Baker 1996: 238n3) or dialects (Georgian, Haspelmath 2004). They may reflect a ­different syntax, say an agreement systems like Abaza’s immune to the constraint (Rezac 2008a). The PCC ban on *E-D-1/2A makes one set of structures ineffable: those with 1st/2nd person absolutive arguments and high datives like possessors, datives of interest, and causees, since both must control agreement, (18). (18) a. Mireni haurrak besoetara bota dizkiote/*dituzte Miren.D children.the.A arms.the.to thrown 3pE-*(3sD)-3pA b. Mireni zu besoetara bota *zaituzte/*zaizkiote Miren.D you.A arms.the.to thrown 3pE-*(*3sD)-2sA ‘They threw the children/*you into Miren’s arms’  (possessor dative; Rezac 2009)

It is different for low datives. These do not or need not agree in eastern Basque, while in the west they agree if they can, but not if dative agreement is unavailable, as in the previous section when taken up by another dative. 1st/2nd person absolutives are fine with nonagreeing low datives (perhaps with variation in the west, Artiagoitia 2000: 405; Albizu 1997a; Odria in prep.). The result in western Basque is an alternation for datives between (19)a obligatory agreement with a 3rd person absolutive and (19)b no agreement in a 1st/2nd person absolutive.19

.  Related to the availability of nonagreeing with low datives if another dative usurps agreement may be Albizu’s (1997a) unique report of a PCC repair whereby the 1/2 object agrees as dative while absolutive in case: Azpisapoek ni etsaiari saldu *naute/didate, Traitors.the.E I.A enemy.the.D sold *3pE-1sA/3pE-1sD.

 Milan Rezac

(19) a. Mireni haurrak eramango dizkiote/*dituzte Miren.D children.the.A bring.fut 3pE-*(3sD-)3pA ‘They will bring Miren the children’ b. Mireni zu eramango zaituzte/*zaizkiote Miren.D you.A bring.fut 3pE-(*3sD-)2sA ‘They will bring you to Miren’ (nonagreeing dative)  (goal dative; Rezac 2011: 184)

Cross-linguistic parallels suggest that the availability of nonagreeing low datives in western Basque is a repair strategy dedicated to the PCC, one that licenses a syntactic structure distict from the one that allows agreement, say a structure with extra structure in the dative PP (Rezac 2011: chap. 4). However, within western Basque, the suspension of agreement in PCC contexts is not distinguishable from its suspension in the other cases. In unaccusatives, the PCC does not bar any agreeing forms as such, but rather certain pairings of forms and structures. Datives with unaccusatives again occur in high and low configurations, high ones needing agreement (required for say experiencers), low ones not agreeing in eastern Basque and up to possibility in western Basque (available for say goals of motion). The PCC bars D-1/2A combinations for high but not low datives (Albizu 1997b). Apparent exceptions have existed and may exist still, more robustly than for the transitive PCC (see Albizu 2011 for a careful study).20 Neither the absolutive nor the high dative can suspend agreement to escape the PCC.21 (20) a. Mireni hurbiltzen/gustatzen zaizkio Miren.D approaching/liking 3sD-3pA ‘They are approaching Miren’/‘Miren likes them’

.  Contemporary descriptive grammars giving D-1/2A combinations almost always illustrate with motion and never psych-verbs (e.g. Zeberio laguntan yoan ñatzu “I have gone to help you” 2sD-1sA, Etxebarria 1988: 214). Older grammars are another matter. Azkue (1923–5: §885) has the types geu agertu gekiozan “nosotros nos le aparecímos” 3sD-1pA, il nakio “me le he muerto” 3sD-1sA, Lafitte (1979: §538) lakhet hintzaitan “tu m’étais agréable” 1sD-20A (drawn to my attention by B. Fernández p.c.). Yet we do not know a priori if a given dative requires the high structure – cf. low-PP experiencer in She seemed (friendly) to me, possessor Je suis à toi, or variable evidence for the height of causee datives (for Basque in Fernández, Ortiz de Urbina, and Landa 2009, for French in Rezac 2011: 4.5.4) and possessor datives (for Basque in Albizu 2011). .  One source reports omission of absolutive person agreement, reducing 3sD-1pA to 3sD-3pA: Arretxe’s (1994: 250n26) description of Basauri discussed in Rezac (2008b: 100–1). Speakers generally strongly resist such suspension, even when there is no other repair of the PCC; but morphological gaps discussed later have lots of such nonagreement.



Gaps and stopgaps in Basque finite verb agreement 

b. Mireni hurbiltzen/?*gustatzen natzaio Miren.D approaching/?*liking 3sD-1sA ‘I am approaching Miren’/‘?*Miren likes me’

Two repair strategies are reported in the literature. Both avail themselves of the absence and thus the availability of ergative agreement in unaccusative constructions. One strategy, absolutive displacement or absolutive promotion, deploys ergative agreement to code the absolutive, with speaker variation on whether the pronoun is ergative or absolutive in case (Arregi 2004; Arregi and Nevins 2011; Rezac 2008b, 2011: 5.6).22 This is impossible outside PCC contexts, in D-3A combinations: (21) a. Itxasori liburuak/*liburuek gustatzen zaizkio/*diote Itxaso.D books.the.A/*E liking 3sD-3pA/*3pE-3sD(-3sA) ‘Itxaso likes the books’ b. Itxasori (%zuk/%zu) gustatzen *zatzaizkio/diozu Itxaso.D     you.%E/%A liking    3sD-2sA/2sE-3sD(-3sA) ‘Itxaso likes you’ c. (Zu/*zuk) Itxasori etortzen (*)zatzaizkio/*diozu    you.A/*E Itxaso.D coming (*)3sD-2sA/*2sE-3sD(-3sA) ‘You are coming to Itxaso’  (Tolosa, Rezac 2008b; zatzaizkio learned form) (22) a. Niri su/suk ondo jauste stasu me.D you.A/E well liking 2sE-1sD(-3sA) ‘I like you’ b. Jonei gu/guk es dotzau gustaten Jon.D we.A/E not 1pE-3sD(-3sA) liking ‘Jon does not like us’ c. Niri Jon ondo jausten gasta me.D Jon.A well falling 3sD-3sA

*Niri Jon/Jonek ondo jauste sta *me.D Jon.A/E well falling 3sE-3sD(-3sA)



‘I like Jon’

(Ondarroa, Arregi and Nevins 2011: 3.2)

This strategy appears to be a new phenomenon. I know of no mention prior to Arregi (2004), and it seems to characterise speakers born after 1970; many, old and young, sharply reject it. Yet it may be ever more widespread in western Basque, where it has been found in independent inquiries (Rezac 2008b, Arregi

.  In Tolosa, the variation does not seem to reduce to optionality of ergative marking on 1/2 pronouns, as Arregi and Nevins (2011: 3.2) propose for Ondarroa.

 Milan Rezac

and Nevins 2011; K. Arregi p.c. of a class questionaire; B. Fernández p.c. of a class questionnaire).23 The specificity of this repair strategy to the PCC is highlighted by its unavailability to fix other agreement gaps in unaccusatives. Many dialects, especially in western Basque, are subject to a partial or total loss of D-1/2A forms, whether the dative is high or low. This is a morphological loss. Often it gives rise to a spotty D-1/2A paradigm. These morphological gaps are not repairable by ergative agreement of the absolutive, unlike the PCC. Even if a speaker lacks (*)etortzen zatzaizkio in (21c), and allows gustatzen diozu for *gustatzen zatzaikzio in (21b), she cannot use the latter to fix the former as *etortzen diozu, (21c).24 Both the limitation of the PCC to high datives, and the limitation of PCC repairs to high datives even where low dative D-1/2A forms are missing, indicate the dependency of the PCC and its repairs on syntactic structure rather than surface morphology. Absolutive displacement is distinct from another repair strategy, call it absolutive shift. It takes the root of D(-3sA) forms, and tacks onto it a 1/2 suffix, as if the absolutive were agreeing in the manner of an ergative. Thus for 2sA-1pD-PST in (23), the EB form is z2s-inTM-tzai√-zkipA-gu1pD-nPST, with the absolutive controlling the prefix and absolutive pluraliser. Instead, we get zX.PST-iTM-tzai√-gu1pD-zu1sA-nPST, that is 1pD(-3sA)-PST zitzaigun plus 2sE/D suffix -zu- attached in its usual place before past tense -n, but controlled by the absolutive. In the few examples I have, expected absolutive argument is pro-dropped or in absolutive case, not ergative.25 .  This strategy combines with other developments, such as dative displacement in nekatuta iruditzen didazu → nazu ‘you seem tired to me’ 2sE-1sD → 2sE-1sA and the loss of ergative displacement nekatuta iruditzen zenian → ziazun ‘you seemed tired to me’ 2sE-1sD. Often the phenomena belong to the same register and reinforce each other’s naturalness. .  I do not know whether Western Basque morphological D-1/2.A gaps can avail themselves of nonagreeing low datives, as they can when agreement problems arise in transitives; speakers seem to use circumlocutions such as the allative for goals (Egaña 1984: 11; so also B. Fernández p.c.). This transitive-unaccusative difference may relate to why low datives do not create the PCC with intransitives, (*)etortzen natzaizu ‘I am coming to you’ 2sD-1sA (only a morphological gap in some varieties), but do with transitives, *ekarri naiozu ‘You brought me to him’ 2sE-3sD-1sA (universally out, with repair by a nonagreeing dative). Distinct but possibly related is the tendency in some western varieties to replace all datives in unaccusatives by alternatives like the allative, effectively eliminating the D–A paradigm, which again does not affect transitives or the E-D-A paradigm (see e.g. Etxabe and Garmendia 2003: 166 for younger speakers in Zaldibia). .  For pro-drop, cf. (i), where zitzaidazun is like zitzaiguzun but with 1sD -da-. (i) Ezagutu ginenean, ikusi zintudan eta mutilik ederrena iruditu zitzaidazun (orain ere) When we knew each other, I saw you and you seemed to me the most beautiful of boys (now too) (hitzarentxokoa.blogspot.com/2013/02/ez-dakit-zergatik.html)



Gaps and stopgaps in Basque finite verb agreement 

(23) Anai-odolez gorri zan gure mendia,/etxe ta baso sutan, negarrez erria,/ta arantza-gaiñean Zu, Goi-pakezko izarra,/ agertu zitzaiguzun, uxatzen negarra you.A … appeared 2sA-1pD-PST ‘Our mountain was red with blood of brothers,/house and wood on fire, the country in tears/and above the sloe, you, star of high peace, appeared to us, chasing the tears’ (Jaurtakol, “Beti zure erri”, Olerti 1961 I)

Aramaio (2001) is the sole report known to me of what might be absolutive shift, in the Basque of Berriatua and Ondarroa. Beside the D–A forms expected for this Bizkaian area, Aramaio reports new forms for D-1/2A combinations. They are shown in Table 3. Table 3.  Berriatua D–A present indicative paradigm† 3sD

3pD

1sD

1pD

2sD

2pD

3sA

(j)akoα

(j)akueα

(j)ateα

(j)akuα

(j)atzuα

(j)atzueα

3pA

(j)akosα

(j)akuesα

(j)atesα

(j)akusα

(j)atzusα

(j)atzuesα

1sA

nakoα

nakueα

–

–

*

*

1pA

atzauβ

*

–

–

(g) atzuauβ

*

2sA

atzasuβ

*

(g)astesuβ

askusuβ

–

–

2pA

atzasueβ

*

(g)astesueβ

askusueβ

–

–

†,αold

D–A formation,

βnew

D–A formation, – reflexive, *gap

All the forms are distinct from E-D-A forms, including from E-3sD-3sA otzat1sE/u1pE/su2sE/sue2pE/Ø3sE, 3pE-3sD-3sA (o)tze. Aramaio views them as Berriatua adaptations and extensions of parallel Ondarroa formations. In Ondarroa, they are restricted to 1sD-2sA (g)astasu, 1sD-2pA (g)astasue, transparently formed in the manner of zaidazu: 2s su, 2p sue suffix on 1sD-3sA gasta. ­B erriatua 2sD-1pA is built up in this manner too. So are Berriatua 3sD-1/2A, 1pD-2A forms, save that they use stem atza, asku, which look like a cross between D–A and E-D-A stems. There are two ways of looking at absolutive shift forms. One is as described, with 1/2 suffixes otherwise controlled by D/E tacked onto a D-3A stem. Absolutives almost never control suffixes across Basque dialects, but they do so in absolutive displacement. Alternatively, the forms look like absolutive pronouns tacked onto the stem of D-3A forms, like zaida-, though not onto the D-3A form itself, in this case zait. Absolutive pronouns are identical to suffixes for 1p and 2s/p. Just these are the forms attested in absolutive shift; I know of no 2sD-1sA zaizut – as

 Milan Rezac

yet. Either way, absolutive shift gives rise to forms not used for anything else, while absolutive displacement always yields extant E-D-A forms.26 Almost all remains to be learned about absolutive shift. Aramaio does not give any examples of the forms in use, so it is not even clear whether her forms are used as PCC repairs as in (23), or whether they are found outside PCC contexts with low-dative unaccusatives. In both Ondarroa and Berriatua, the forms are a new phenomenon: We need to keep in mind that these new forms, both in Ondarroa and in B ­ erriatua, are only heard from some young people, and so neither old people nor most speakers use them in daily life. It seems that the lack of certain verbal forms lead them to the invention of new verb forms unconsciously, and perhaps by confusion or analogy, these strange forms were created. (Aramaio 2001: 15n16)

Absolutive shift in any form is sharply rejected by many speakers, including those with absolutive displacement.27 Most speakers have no repairs for the PCC. They resort to circumlocutions, to expressions available independently of gaps. A typical circumlocution for ­Gustatzen zait/*zatzaizkit, lit. liking 1sD-3sA/*2sA, is the transitive predication of a derived adjective, Gustuko(a) dut/zaitut, lit. pleasant have-1sE-3sA/2sA (“I have you pleasant”). The relationship of gap to circumlocution is the same as between any two expressions of similar meaning, say I am on her side and She has me on her side. In this example at least, existence of the gap is keenly felt and the circumlocution frequent. Stories about gaps that look to function have work to do.28 .  That includes any vagaries an E-D-A form may have in a given idiolect, say loss of ergative displacement in 2s/pE-D-A but not 1pE-D-A. .  I am grateful to B. Fernández for providing me with Aramaio’s work, and pointing out that Aramaio does not describe the phenomenon as identical to the use of E-D-A, unlike what is reported in Arregi and Nevins (2011: 3.2). I am also grateful to those who have swiftly and firmly rejected gustatzen zaidazu as wholly unknown, whether they have gustatzen didazu, nazu or nothing: K. Erdozia, U. Etxeberria, A. Irurtzun, and J. Manterola. .  Ideal speakers in secure relationships readily confess awareness of gaps and volunteer circumlocutions to be studied. Otherwise one has Orr’s experience, only imagine the grocer being pressed by one who wants to hear him say I like you. This is further supported by the data reported by Orr concerning nine eggs. Orr first found it impossible to surreptitiously elicit neuf ‘nine’ directly followed by oeufs ‘eggs’; the grocer would consistently say neuf beaux oeufs [nœfbozø] ‘nine beautiful eggs’. When Orr finally asked the grocer directly why he would not say neuf immediately followed by oeufs, the grocer’s answer was extremely revealing: ‘My god sir, with neuf, I always say: [nœfbozø]. [nœfœf], you see, that clashes. [nœfzø], ultimately, that would maybe be more correct.’ Tranel (1981: 214)



Gaps and stopgaps in Basque finite verb agreement 

6.  Morphological gaps The last set of gaps seems to belong to morphology, the best-explored domain of agreement gaps (for overviews, see Rezac 2011: chap. 2 and the papers in B ­ aerman, Corbett and Brown 2010). In drawing out the theoretical consequences, I will view morphology as the realization of syntactic structures by pieces, as in Distributed Morphology DM (Embick and Noyer 2007; on morphological theories, see Stump 2001). Morphological gaps are best known from derivational morphology, for instance -ity nominalizations, but inflection is rife with them, and their sources and behaviours are diverse. One type of gap is illustrated by the absence of the past of forgo, beside underwent and embargoed, or of the past participle of stride, beside hid(den), ridden, glided. Such gaps may be due precisely to the multiplicity of available formations and arise when there is not enough evidence to acquire one (Yang et al. 2012; Albright 2006). The result is ineffability: one simply cannot express the past of She forgoes dinner save by circumlocution. I will refer to this as the stridden type of gap. A second type of gap, the amn’t type, differs on two scores. There is no *amn’t in English beside aren’t, isn’t, wasn’t, weren’t, couldn’t, though the formation should be transparent and unambiguous. The origin of the gap may have to do with phonology, and for long it has been positively reinforced, helped perhaps by markedness, since am is the sole form coding person agreement in English. The gap also has something that the stridden gap does not: a form unexpectedly used in its place. This form is aren’t, and is unexpected insofar as are is not otherwise a 1s form of be. Indeed in some varieties aren’t is not available to plug the gap, while in others it is available only in contexts like inversion. I will refer to such unexpected forms for expected ones as as stopgaps (see Hudson 2000; Broadbent 2009; Nevins 2012 on the origins and distribution of amn’t and its aren’t stopgap). The nature of the relationship between gaps and stopgaps is of great interest. If stopgaps respond to gaps in order to repair them, then they rely on ­reference-set computation whereby one structure refers to the goodness of another. Alternatively, aren’t might be quite independent of amn’t. It might be that are is a default, pre-empted by am ordinarily, but not when am is unavailable due to the gap, which lets are emerge. It might also be that are has simply become grammaticalised as a 1s form in the context of negation and whatever else it is restricted to like inversion (Embick and Marantz 2008; Hudson 2000; Bresnan 2001; Broadbent 2009; Nevins 2011).29

.  I should say that there are gaps that look like a problem with form but whose statement goes beyond what one expects of a morphological condition. The problem has been well

 Milan Rezac

Before turning to gaps, something needs to be said about the productivity of Basque verbal agreement morphology. There is evidence that speakers can construct and use as grammatical novel forms built with pieces found in familiar ones. The most striking are K. Mitxelena’s “morphological monsters”. In the canonical paradigms of the language – agreement forms for all combinations of ergative, absolutive and dative coarguments – only one absolutive controller is found. However, there are cases where a singular absolutive argument of a verb controls absolutive person morphology, and a remote plural absolutive controls absolutive number morphology, giving othewise impossible combinations like 1sA + pA agreement (Ortiz de Urbina 2003c: 293–4; Oyharçabal 2003: 784, 2005). Mitxelena’s example (24)a involves a relative operator as the remote absolutive. The standard form of the finite verb would be n1sA-enTM-bil√-enPST.REL walk-1sA-PST.REL, but instead n1sA-enTM-bil√-tzapA-nPST.REL walk-1sA-pA-PST. REL may be preferred, with the pA morpheme tza seen in canonical forms like g1pA-enTM-bil√-tzapA-nPST.REL walk-1pA-PST.REL. The pA agreement is controlled by the relativised object of bila ‘in search of ’. Similar monsters were identified by P. Lafitte where the remote absolutive is the object of a restructuring infinitival complements, (24)b. The standard form would be nindoakon 3sD-1sA, but the pluraliser z is added to agree with harmak ‘the arms.sA’ in the complement hartzera ‘to take’.30 (24) a. Nik ez ditut aurkitu … [[__ bila] nenbiltzan]-ak    in_search_of walk.1sA.pA.PST.REL-the.Ap ‘I have not found … what I was looking for’  (Mitxelena, cited in Ortiz de Urbina 2003: 293–4) b. etxaiari harmak hartzera nindoazkon enemy.D arms.pA take.to go.3sD.1sA.pA.PST ‘I was going to take the arms from the enemy’ (Lafitte 1979: §531)

studied for the English Double-ing Filter, *It’s starting raining (Ross 1972; Pullum and Zwicky 1999; Richards 2010: 23; Nevins 2011). Here is a case from Gipuzkoan Basque, as described in de Rijk (2008: 24.6.1): the potential formed by the particle EB ahal, Gipuzkoan al, is blocked precisely when it can be interpreted as the question particle al, namely the simple present and past of matrix clauses, and independently available but otherwise less common alternatives like the potential mood must be used. .  Ortiz de Urbina (2003c: 294) observes that pA tza in Lafitte's forms is not the tzi given for canonical forms like tzi in doatzi ‘go-3p’, bagoatzi ‘go-1p’ (but Lafitte does give canonical forms like doatza, §512, goatzak, §584, so the matter needs more study).



Gaps and stopgaps in Basque finite verb agreement 

Construction of such forms in both relativization and remote agreement needs study, but it may be relatively common. Examples can be found in well-edited publications and are acceptable to speakers on reflection:31 (25) a. Nik ez dakit ematen ari natzaizkizuen arrazoiak gutxi ala asko diren, baina beste bat emango dizuet ‘I do not know whether the reasonsi that I am giving ye 2pD-1sA-3pAiREL are few or many, but I will give ye another’  (Eusko legebiltzarra IX N0. 45 zk., 25/03/2010, p. 101) b. Hala ere, dezibelak neurtzen ari natzaizkizu, eta duela 48 urte Caracasera joan zinenean baino ozenago hitz egiten duzu ‘However, I am measuring decibelsi on you 2sD-1sA-3pAi, and you are louder than 48 years ago when you went to Caracas’  (Berria, “Fidel zutik eta onik” by Amagoia Mujika)

These forms suggest that Basque verbal morphology is not to be entirely analysed as a set of monolithic memorised forms whose pieces are not part of I-language though accessible to analogical generalization in language change, as argued by Hualde (2002). There is a certain productivity. The existence of a productive formations does not preclude gaps: past, perfect, and contracted negation all have productive formations in English. Yet Basque morphology can also seem like a lexicalised mess, as in the gradual loss of ergative displacement motivating Hualde’s proposal. Here is one example: the placement of ergative and dative agreement morphemes ka/na 20m/f with respect to other suffixes in the dialect of Guernica (Yrizar 1992a: 245ff.). Usually, pA s precedes 20m/fE ka/na, as in the variety of Ajánguiz 20m/fE-1sD-3pA deusta1sD-sapA-k/na20m/fE. In the neighbouring variety of Navárniz, however, the order depends on the gender of 20E, deu-sta1sD-sapA-k20mE but deu-sta1sD-na20fE-spA. Even in Ajánguiz, this is the pattern if the 1sD dative is changed to 1pD, 20m/fE-1pD3pA: deu-sku1pD-sapA-k20mE but deu-sku1pD-na20fE-spA. In 20m/fE-3pD-3pA, the order in Ajánguiz has evolved from this latter pattern, deu-tze3pD-sapA-k20mE, deutze3pD-na20fE-spA, at the beginning of the twentieth century, to dxau-tza1D-sapA-k/ na20m/fE at the end. Turning to dative 20m/fD ka/na, usually it precedes ergative 1s/pE da/gu in Ajánguiz as in Basque generally, 1s/pE-20m/fD-3sA deu-a/na20m/ fE-t/gu1s/pD. But changing from 3pA to 3sA, there is both variantion on this usual

.  Both examples are acceptable to R. Etxepare (p.c.). Other reactions exist: P. I­ribertegi’s review in Berria 2003-08-16 (kritikak.armiarma.com/?p=854) heaps little but praise on K. Navarro’s translation of I. B. Singer’s Zortzi kontakizun (Elkar: Alberdania, 2002), but asks “Is there a verbal form natzaizkizu?” See also the discussion in Oyharçabal (op.cit.).

 Milan Rezac

order, 1sE-20m/fD-3pA deua-na20fE-da1sD-spA beside deua-da1sD-sapA-na20fE, and exceptions to it, 1pE-20m/fD-3pA deua-gu1pD-na20fE-spA from deu-na20fE-gu1pDspA a century earlier. There is a quaint term comes to mind: higgledy-piggledy. Variety after variety in Yrizar’s fifteen volumes of the morphology of the most commonly used paradigms of the auxiliary has something like this, often a great deal of it. Variations in agreement morphology can be very neat and serve as isoglosses, but there always seem to be varieties they become a higgledy paradigm (including all the dialectal markers in Zuazo 1998). From form to form for a single speaker, from farm to farm, from generation to generation, unpredictability is common, though the rule only when a paradigm stands on the threshold of loss.32 One can see how this complexity might give rise to gaps: if forms are memorised, one cannot predict a missing one; if they are composed, there are many modes of formation. Yet gaps are not usually associated with such complexity. There are no gaps in the E-D-A paradigms of Ajánguiz: no absolute gaps, that is ineffable E-D-A combinations, nor any combinations realised by a form impoverished for agreement phi-features. Rather, there are different ways to map phifeatures to form, like pA + gender and gender + pA, but together they supply one or more forms for every feature combination. English past is like that: the weak pasts have tended to replace the strong ones an item-by-item basis, crow: crew → crowed, and this is the productive mode for wugged, but there is the odd reversal, sneak: sneaked → snuck, and no synchronic predictability, know: knew, streak: streaked. Yet gaps like *forwent are almost absent. A nice counterpart in a clitic system is Iberian variation in the ordering of se and me/te (Heap 2008). Gaps of the stridden type, with no stopgap, and amn’t, with the aren’t stopgap normally realising a different structure, do occur in Basque agreement. They are most frequent where an entire agreement paradigm is being lost. In the Basque auxiliary, this is common and largely restricted for D-1/2A forms (Rezac 2013). .  I had long wondered whether such microvariation was an artefact of Yrizar’s massive study; but exhaustive studies of the grammars of and by individual native speakers explicitly highlight it (e.g. Zeberio, Etxebarria 1988: 210; the source for late 20th century Ajánguiz forms seems to be among them, Yrizar 1992a: 245). My qualms have been allayed by sharp judgments I have met myself: for EB zenion and genion 2sE/1pE-3sD-3sA-PST (2sE/1pE coded by prefixes z-, g- under “ergative displacement”), a central Gipuzkoan speaker with ziozun but giñion (ergative displacement lost in favour of a suffix for 2E but not 1pE, the most common pattern) smiles at the ziñiozun for zenion of a friend a short afternoon walk’s away (ergative displacement doubled by a suffix, still for 2E only) or the gendiogun for genion of another over the hill (same but extended to 1pE) and knows nothing of 1pE-3sE nugun for EB genuen at the end of a frequent trainride (with prefix otherwise used for 1sE and suffix for 1pE). For another example, see Rezac (2006: 3.2.1).



Gaps and stopgaps in Basque finite verb agreement 

The full set of D-1/2A forms is present in EB, and – so it seems – in every variety of which there are records at the beginning of the 20th century or earlier, including varieties where they are now mostly gone (Azkue 1923–5), though even then there are hints of incipient loss, as in Markina (Rollo 1925). Today, a relatively full set of forms is found in eastern varieties, like Ahetze (Cabodevilla 1991) or Maya (N’Diaye 1970). In many western and central varieties, D-1/2A forms are disappearing, leaving only D-3A. D-1/2A forms are gone in Basauri (Arretxe 1994), Pasaia (Agirretxe et al. 1998), and Ondarroa (Rotaetxe 1978). Only vestiges, particularly for present indicative D-1sA and chiefly among older speakers, remain in Ultzama (Ibarra 1995) and Bermeo (Gaminde, Romero and Legarra 2012). More robust is Larresoro (Epelde 2003), with 3sD-1sA/20A present and past, but 1pA and 2s/pA lost. Still more so Zeberio (Etxebarria 1998), with intact and used present indicative save for D-2pA, but otherwise only D-1sA. In Zaldibia young speakers even seem to be giving up the whole the D–A paradigm (Etxabe and ­Garmendia 2003). The loss of these forms is absolute: nothing takes their place, and independently available circumlocutions are used instead: thus Nekaneri joan natzaio “I  went to Nekane”, lit. Nekane.D gone 3sD-1sA, with an agreeing dative, vanishes, but is expressed well enough by Nekanerengana joan naiz 1sA, lit. Nekane.to gone 1sA, with an animate allative.33 .  Contemporary descriptive grammars of western and central dialects where D-1/2A forms are partly or almost wholly lost mention relics among older speakers, typically for D-1sA and the imperative (presumably due to their frequency, see Hualde 2002 for 1s); when examples are given, they involve low datives like the goal for etorri ‘come’ and not high datives like the experiencer of iruditu ‘seem’: Arrasate, Bergara (imperative only), Bermeo, Ermua, Leioa, Mallabi (imperative only), Oñati, Otxandio (imperative only), Sopela, Ultzama, Zaldibia, Zamudio, Zeberio, Zegama. Only D-3A forms are mentioned for Antzuola, Basauri, Eibar, Foru, Lekeitio, Oiartzun, Ondarroa, Ordizia, Orio, Pasaia. Azkue (1923–5) quoted below is clear on the existence of D-1/2A in dialects where they are now lost. He is corroborated by careful descriptions like Rollo (1925) for Markina, who goes to such nuances like the availability of etorri nakisula “that I come to you” 2sD-1sA beside the incipient loss of D-1/2A forms in the allocutive conjugation, and rarity and hesitation in less common tense-mood combinations. Among localities where D-1/2A forms are now mostly gone, the morphophonology of relics seem to presuppose an internal history (e.g. Zeberio (laguntan yoan) ñatzu, natxatzu, nayatzu 2sD-1sA, given as living forms, Etxeberria 1988: 214). This all suggest the classical story of loss I go with in the text. B. Fernández (p.c.) informs me of a widespread current view that D-1/2A forms never existed where they are now absent. It needs very ­Galfridian a reading of Azkue: The inflections … nakio [3sD-1sA] and nakizu [2sD-1sA], which in imperative and subjunctive are said without deformation and are still heard as such in Bakio, Urduliz, Baŕika and Maruri (at least in the past*), the rest of us Bizkaians pronounce in one of the following ways: nađako (Aŕatia), nadxako (Lekeitio), natxako (Markina and ­Mondragón), najako (Soraluze). [Note *] Even aged persons of these villages,

 Milan Rezac

The deterioration of the D–A auxiliary does not typically affect other paradigms of the auxiliary, at least not in the indicative, not until a dialect is on the threshold of death (as the variety Olazagutia-Ciordia of Burunda, Yrizar 1991). It is, however, echoed in the paradigms of strong or synthetic verbs: those that may themselves carry agreement inflections. Their number and range of forms has been decreasing throughout Basque history, and survivors in current spoken varieties mostly present the moth-eaten aspect of typical western D-1/2A paradigms. 7.  1p+2 gaps and stopgaps In this and the remaining sections, I turn to particular set of gaps identified by Arregi and Nevins (2006, 2011).34 They are of particular interest because of the nature of their stopgaps. I will agree with Arregi and Nevins’s analysis: the gaps and stopgaps result from morphological operations that eliminate phi-features in the agreement complex on the way to its realization. Basque agreement morphology combines one of absolutive A agreement, dative D and ergative E agreement in A, D-A, E-A, and E-D-A- paradigms. Each paradigm has characteristic allomorphy, such as the choice of auxiliary root: in EB root u for E-A, tzai for D-A, and i for E-D-A.35 In the +E paradigms, namely those agreeing with E, western dialects sometimes have gaps in 1p+2 combinations. They differ on which combinations are gapped and in how they are expressed. In this section these gaps and their stopgaps are introduced in general; the next looks at one particular variety up close.36

consulted by me in May of 1924, no longer use nakio in indicative, but rather, following the general trend, say nađako (Urduliz and Baŕika), nadxako (­Bakio), naixako (Maruri). The inflections of the past are said in their pristine purity: neu agertu nekion I appeared to him [3sD-1sA], geu agertu gekiozan [3sD-1pA] we appeared to him.  (Azkue 1923–5: §885; all unglossed auxiliary forms are 3sD-1sA) .  The gaps have long been noted in descriptions such as that of Egaña (1984), and briefly in generative literature by Fernández (2001: 156). .  3sA has zero exponents and does not condition allomorphy, as if it were absent, as discussed later. .  The phenomenon has only been discussed for western dialects, but occurs elsewhere. In the southern High Navarrese subvariety of southern Erro, at the locality of Aincioa, the transitive indicative present paradigm uses differential object marking forms for 2-1, e.g. ikusi dirazu neri … karriken “you saw me 2sE-1sD … in the street”, but for 1sE-2sA there is the i­mpersonal-type stopgap elizan ikusi zaizu (!) “te he visto”, lit. church.in seen 2sD(-3sA), instead of 1sE-2sD as in neighbouring Loizu nik zuri ikusi dizut oiania “I saw you in the forest” (Yrizar 1992b: 303; (!) is his).



Gaps and stopgaps in Basque finite verb agreement 

Consider first Zeberio Basque (Etxebarria 1998: 147ff., 211–2). In otherwise intact E–A and E-D-A indicative paradigms of the auxiliary, the combinations 1pE-2s/p/fA and 1pE-2s/p/fD-3A are missing, vestiges aside. Two strategies are reported to make do without them. One is circumlocution, here described for 1pE-2pA: [W]hen a form is almost lost, things are expressed by a sentence of some other sort: For example, instead of saying atzo bakarrik ikusi zinduguzan [“We saw you alone yesterday”, 1pE-2pA-PST], the following are said: atzo bakarrik ibili sintzen [“Yesterday you were walking alone”, 2sA-PST] … se iten saindun ba bakarrik atzo [“What were you doing alone yesterday”, 2sE-3sA-PST] or e(s)aunken lagunik atzo ala [“You did not have friend yesterday then?”, 2sE-3sA-COND]?  (Etxebarria 1998: 212, 215)

The other strategy is stopgaps, forms that exist independently but not with the syntax and meaning of the gap, like aren’t for *amn’t. In Zeberio, the foregoing gaps recruit forms identical to those used as impersonals of transitives. Impersonals of transitives eliminate any overt expression of the ergative argument, including its agreement. Likewise for stopgaps: +E forms are replaced by ‑E ones, 2A for 1pE-2A and 2D-3A for 1pE-2D-3A. I return to this below; here impersonals are translated as passives: Despite some forms not being used, the flow of communication is not lost, ­because it is completed by a filler verb-form. I will mention an example that I have often heard: [A:] belu deta, banoa etzera [‘It’s late, I am going home’] [B:] nasai ba, eroango sara ra [‘Don’t worry, you’ll be brought’, 2sA] That is, instead of eroango zaitugu [‘We’ll bring you’, 1pE-2sA], an impersonal case is used as filler. [A:] ser ingu ba aurten buskentzakas [‘What will we do then this year with the blood sausages?’] [B:] es ikeratu ba, eroango yakos da [‘Fear not, they’ll be brought to him’, 3sD-3pA] Here also, instead of using eroango deutsaguz [‘We’ll bring them to him’, 1pE-3sD3pA], the impersonal is used. (Etxebarria 1998: 211–2)

Bermeo presents similar gaps but a different sort of stopgap (Egaña 1984; Laka, Olondo and Gaminde 2012; Gaminde, Romero and Legarra 2012). 1p+2 combinations are wholly missing in the E(-D)-A paradigms, and there are two sorts of stopgaps, both affecting the 1p argument. 1pE-2A and 1pE-2D-3A behave as in Zeberio: 1pE is obliterated, resulting in 2A and 2D-3A forms respectively, identical to those used in impersonals. In 2E-1pA, on the other hand, the stopgap is 2E(-3sA), which can be viewed as obliteration of 1pA agreement or its impoverishment by deletion of both person and number. I return to this dichotomy later.

 Milan Rezac

What do we do when we need to use these missing forms? Well, the ABS, DAT, or ERG is put before or after the verb, as needed. Let us see with one example:   *Eroango gaitusu [‘You will bring us’, 2sE-1pA] →  Erungosu gu [‘You will bring it’, -2sE.3sA we.A]   … In some other cases, to overcome gaps, the impersonal form is used:    *Esan geuntsun ez erosteko ezer [‘We said to you not to buy anything’, 1pE-2sD(-3sA)-PST] →  Esan dxatsun es erosteko eser [‘It was said to you not to buy anything’, 2sD(-3sA)-PST](Egaña 1984)

Circumlocution is also used: It can happen that using neither, then seizing the sense, what we want to say ­appears in another way. That is: Instead of saying Erungosu gu [see above], to say this: Bakosu guretsako lekorik ela? [“Have you room for us?”] (Egaña 1984)

The circumlocutions look like circumlocutions do: independent expressions with a close enough meaning to the gapped expressions to be usable in their stead. Far less clear, and potentially far more interesting, is the nature of the stopgaps. Consider first those syncretic with the impersonals of transitives. Their description as impersonals suggests that they might be circumlocutions. On this view, instead of saying We’ll bring you home, one says You’ll be brought home, as one could independently of the gap, with all the syntax and the semantics of an impersonal. At most the gap would influence pragmatics, in the way the use of We’ll meet twelve/#one day from today is affected by the availability of tomorrow (for such effects in impersonals, cf. Zribi-Hertz 1982, 2008). Alternatively, the impersonallike stopgaps might be unusual realisation of regular transitive syntax. This is the analysis of Arregi and Nevins (2011) as morphological deletion of ergative agreement. On their proposal, the stopgaps share no aspect of the syntax of impersonals to the exclusion of transitives. The Basque impersonal is understood well enough that these alternatives can be probed (Albizu 1998, 2001; Ortiz de Urbina 2003b, 2006; for the closely similar Spanish se-impersonals, Mendikoetxea 1999, 2008). The impersonal is illustrated in (26) and described below. (26) a. Transitive (Haiek) ateak irekitzen dituzte    they.E doors.the.A opened 3pA-3sE ‘They open (the) doors’



Gaps and stopgaps in Basque finite verb agreement 

b. Impersonal and Anticausative (*Haiek) ateak irekitzen dira     they.E doors.the.A opening 3pA ‘The doors are opened’ and ‘The doors open’

I. The impersonal of a (di)transitive has the realization of an absolutive-subject intransitive, identical to the anticausative when there is one: there is no overt ergative and agreement is (D-)A rather than E-(D‑)A. II. The syntax of the impersonal differs from that of both the anticausative and transitive. The external argument, interpreted as ‘people’ or ‘someone’, cannot be realised as agreeing or nonagreeing or pro-dropped ergative, and is partly syntactically active (for reciprocals and control but not reflexives). III. If the absolutive of a transitive is 1st/2nd person, E-1/2A, it resists this impersonal treatment to 1/2A, but may appear for many speakers as dative in case and agreement, 1/2D(-3sA).37 Using these properties to probe impersonal-looking stopgaps, it is clear that while they have the same realization (I), they differ in syntax (II–III). On (II), stopgaps unlike impersonals have an overt albeit nonagreeing ergative, (27)–(28). On (III), impersonal-looking stopgaps regularly fix 1pE-(D-)2A gaps with (D-)2A forms, (28), flouting impersonals’ restriction to 3rd person absolutives.38 (27) orain geuk kantako dxatxu suri now we.E sing.fut 2sD(-3sA) you.D   cf. EB orain geuk kantatuko dizugu zuri now we.E sing.fut 1pE-2sD(-3sA) you.D ‘Now we will sing to you’(Bermeo; Gaminde, Romero and Legarra 2012: 144) (28) [I]nstead of using zaitugu [1pE.2sA.AUXEA] and zaituztegu [1pE.2pA. AUXEA], the intransitive ABS-subject paradigm is used, for instance: – guk ikusi sara [we.E seen 2sA] ([for] guk zu ikusi saitugu [“We saw you” we.E you.A seen 1pE-2sA]).(Foru; Gaminde 1992: 92) .  This seems independent of whether a speaker allows differential object marking of a transitive high-animacy object as dative (Fernández and Rezac this volume), but the matter has not been studied. Spanish seems to show the same behavior. .  The situation in Bermeo at first sight suggests a more nuanced picture: whereas older speakers plug *1pE-2A gaps with 2A stopgaps, younger speakers use 2D(-3sA), as if making the stopgaps conform to impersonals. However, younger speakers have also introduced differential object marking of plain transitive 1st/2nd person direct objects by the dative, and this is expected to extend to stopgaps if they are just plain transitives with morphological suppression of 1pE (cf. Laka, Olondo and Gaminde 2012: 30).

 Milan Rezac

As far as (II–III) go then, impersonal-like stopgaps look syntactically like ­transitives, not impersonals. Only their realisation is anomalous. To be sure, the evidence is not conclusive. Taking (II), for instance, one could suppose that the stopgaps have the syntax of impersonals plus additional structure to license  the external argument. However, this syntax would have to be parametrised by idiosyncratic phi-­ combinations of two arguments, subsets of 1p+2; and that may be beyond the bounds of syntax, while morphology does manipulate individual phi-features according to other phi-features in local contexts (Arregi and Nevins 2006, 2011; Rezac 2011: 2.2; I return to this in the next section). So (II) and (III) support a realizational analysis. There is something that stopgaps share with impersonals beyond form. Impersonal argument often acquire “pseudospecific” and “specific” uses, notably a use similar to and sometimes replacing we (Cinque 1988). Among stopgaps for 1p+2, those that look like impersonals are only found when the obliterated E is 1p: there is for instance zaitugu 1pE-2sA → zara 2sA, but no gaituzu 2sE-1pA *→ gara 1pA, only → dozu 2sE(-3sA). It is tempting to think of impersonal-looking stopgaps as grammaticalisations of impersonals in (pseudo)specific use. Yet it is opaque how to get a 2A stopgap for 1pE-2A when impersonals are unavailable for E-1/2A combinations. So this might be a dead end. Indeed, 1p is the preferred target of agreement obliteration or reduction not only when the result ends up like the impersonal, but also otherwise: for instance, in Bermeo and Ondarroa, not only in 1pE-2A/D→2A/D, but also in 2E-1pA/D→2E(-3sA). This brings us to the nature of the second type of stopgap, which is not syncretic with the impersonal, but also reduces the phi-features coded in the agreement complex. A well-studied example is for 2E+1p gaps in Ondarroa (Arregi and Nevins 2011: 172–4). In (29)–(30) with 2E-1pD(-3A), 1pD is wholly obliterated to give 2E(-3A), distinct from 2pE-3s/pD-3A. Yet despite lack of agreement, a 1pD pronoun remains available. This is in partial tension with the situation outside the gapped forms: in Ondarroa as in western Basque generally, a low dative must agree if possible.39 (29) a. Suk guri liburu emon dosu/*doskusu you.E we.D book given 2sE-(*1pD-)3sA cf. b. Berak guri liburu emon dosku/*dau s/he.E we.D book given 3sE-*(1pD-)3sA  (Ondarroa; Arregi and Nevins 2011: 172–175) (30) Suek guri lagundu *doskusuen → senduen ye.E we.D helped *2pE-1pD(-3sA) → 2pE(-3sA)  (Zamudio; Gaminde 2000: 376, in Arregi and Nevins 2011: 177) .  Note that 2sE-3sA dosu ≠ 2sE-3s/pD-3sA dotzasu (Arregi and Nevins 2011: 309), in both (29) and (31).



Gaps and stopgaps in Basque finite verb agreement 

The tension is more striking with absolutives. The gap at 2E-1pA surfaces as the 2E(-3sA) stopgap in (31), with 1pA removed by obliteration or reduction to 3sA. Again the nonagreeing 1pA argument can appear as an overt pronoun. The same goes for the impersonal in Bermeo above (erungosu gu). Yet 1st/2nd person absolutives must agree otherwise, even when it leads to ineffability as in the PCC. Arguably, they need to Agree for licensing (Béjar and Rezac 2009). This suggests that in (31) syntactic Agree licenses the overt pronoun. Agree may also license pro-drop when the absolutive is not overt (cf. Arregi and Nevins 2011: 172). (31) Suk geu ikusi dosu/*gatxusu you.E we.A seen 2sE(-3sA)/*1pA  (Ondarroa; Arregi and Nevins 2011: 172–175)

Both sorts of stopgaps seen here then, those that end up looking like impersonals and those that do not, seem to reflect plain transitive syntax with an anomalous realization, reducing or obliterating agreement phi-features. In neither case do they seem to have arisen from circumlocutions, since the putative circumlocutions are ungrammatical: 1/2A impersonals or nonagreeing datives and absolutives. The last section returns to this puzzle. But first, the next section takes up the system of Bermeo, where the mechanics of 1p+2 gaps and stopgaps can be studied in detail. 8.  Morphological 1p+2 gaps in Bermeo This section describes one system of 1p+2 gaps and stopgaps, that of Bermeo, and in its light the suitability of morphological analyses like that of Arregi and Nevins (2011). The agreement morphology of Bermeo is set out in Laka, Olondo and Gaminde (2012), Gaminde, Romero and Legarra (2012), to which may be compared earlier stages and adjacent varieties in Yrizar (1992a). Bermeo has drastically simplified its agreement complex, and the rise of gaps may be related to it. Moods have been pruned to the indicative and the past mostly refashioned as present + -n.40 Gaps

.  Of the indicative, imperative, subjunctive, hypothetical, and conditional, differing in mood affix, root, and ergative displacement, only the indicative survives, and the imperative limited to 3A objects. The subjunctive has been replaced by nominalizations, the hypothetical and the conditional are formed on the indicative by the invariant particle leki. The past has been remodelled on the present + -n, losing ergative displacement and the default past prefix s-, save in a handful of optional outliers: thus old past 3s san, 2s siñen, 2sD-3pA dxatxus, new da-n beside san, sara-n, dxatxus-en. Also lost are 2nd person familiar and allocutive forms and D-1/2A forms, as commonly elsewhere.

 Milan Rezac

cover the 1p+2 combinations of the E–A paradigm, Table 4, and E-D-3A paradigm, taken up later in Table 5.41 Table 4.  Bermeo E–A paradigm† 3sE

3pE

1sE

1pE

2sE

2pE

3sA

dau sauen

dauie sauien

dot nauen, doten

du +n

su +n

suye +n

3pA

daus, txus sausen

dauie(s) sauien

dotas, dotes nausen, tasen

dus +en

sus +en

suye(s) suyen

1sA

nau +en

nauie +n

–

–

nosu +n

nosuye +n

1pA

gaitxus gaitxu(se)n

gaitxusie gaitxuyen

–

–

→E‑3sA

→E‑3sA

2sA

saitxus +en

saitxusie +n

saitxut +en

→A

–

–

2pA

saitxusie +n

saitxusie +n

saitxutie +n

→A

–

–

†Present

and past on separate lines, with +(e)n attached to present form.

In the E–A paradigm 1p+2 gaps, namely 1pE-2A and 2E-1p, stopgaps eliminate 1p. Arregi and Nevins (2006, 2011) distinguish two ways of eliminating 1p: obliteration, which wholly removes its agreement, and impoverishment, which reduces it to 3s. Obliteration occurs in 1pE-2A combinations and obliterates E agreement. It does not just replace the 1pE suffix -(g)u- by the 3sE zero suffix, but also eliminates allomorphy that depends on the presence of E agreement. So for instance 2sA present (past) sara(n) replace 1pE-2sA saitsugu(n) (recorded for Bermeo). Young speakers use 2D(-3A) forms, e.g. 2sD(-3A) dxatzu, because of differential object marking of 2nd person objects as dative rather than absolutive, independently of the stopgap.42 For 2E-1pA combinations, Arregi and Nevins propose impoverishment: deletion of both person and number of 1pA to give 3sA. So for instance 2sE(-3sA) present su is used for such forms as 2E-1pA *gaitxusu or *gosus (not recorded for Bermeo) – and not for instance 2sE-3pA sus, preserving number. We will see below that obliteration and impoverishment canbe distinguished, but not in these .  Data from Laka, Olondo and Gaminde (2008); Gaminde, Romero and Legarra (2012) differ only in the interaction of the s and (y)e pluralisers. .  Younger speakers have also introduced 1pE-2A forms from EB.



Gaps and stopgaps in Basque finite verb agreement 

forms. Impoverishment rather than obliteration would be required if forms like su (EB dosu) were necessarily 2sE-3sA and not 2sE alone. However, 3sA agreement forms are also found where the presence of even an expletive 3sA argument has been doubted.43 Thus forms glossed E-3sA might reflect E alone, and stopgaps apparently impoverishing 1pA to 3sA can be analysed as obliterating 1pA.44 Shorter shrift can now be given to the E-D-3A paradigm in Table 5. Table 5.  The E-D-3A paradigm† 3sE

3pE

1sE

1pE

2sE

2pE

3sD

sA pA

txo(n) txos(en)

txoye(n) =

txat(en) txates(en)

txagu(n) =, txagusen

txasu(n) txasus(en)

txasuye(n) =

3pD

sA pA

txoye(n) txoyes, =

txoye(n) =

txatie(n) txaties, =

txaguye(n) =

txasuye(n) =

txasuye(n) =

1sD

sA pA

dost(en) dostes(en)

dostie(n) =, dostiesen

–

–

stasu(n) status(en)

stasuye(n) =

1pD

sA pA

dosku(n) doskus(en)

doskuye(n) =

–

–

→E-A

→E-A

2sD

sA pA

txu(n) txus(en)

txuye(n) =

txut(en) txutes, =

→D-A

–

–

2pD

sA pA

txuye(n) =

txuye(n) =

txutie(n) =

→D-A

–

–

†=

indicates that a pA form identical to the sA form.

Gaps again cover 1p+2 combinations. Stopgaps are all clearly of the obliteration type. 1pE-2D-A eliminates 1pE to give 2D-A forms, e.g. present 1pE-2sD-3sA → 2sD-3sA dxatxu for -tsugu (recorded in Bermeo). 2E-1pD-3A eliminate 1pD to give 2E-3A forms, distinct from 2E-3sD-3A: e.g. present 2sE-1pD-3sA → 2sE-3sA su(n) for ‑skusu (recorded for Bermeo; general Bizkaian deuskusu). Let us turn in more detail the morphological analysis of stopgaps in Arregi and Nevins (2011: 4.6, 4.8 cf. 2006). It occurs in post-syntactic realizational morphology. Gaps are barred by markedness constraints on two clitics each with

.  For instance: E(-A?) lanean ematen dute ‘they seem at work’ 3sE(-3sA?), E–D(-A?) jarraitu diete ‘they follow them’ 3pE-3pD(-3sA?), D(-A?) damutzen zait honetaz ‘I.D regret this. INS’ 1sD(-3sA?) (Laka 1993b, 2000; Rezac 2011: 5.5; Preminger 2012; cf. Fernández and Ortiz de Urbina 2010: 3.3.2). .  Symmetry of E–A and E-D-3A paradigms might support this obliteration treatment: as to 1pE-2A→2A corresponds 1pE-2D-A→2D-A, so to 2E-1pA→2E(-3sA?) corresponds 2E-1pD-3A→2E-3A, and the latter clearly has obliteration as seen next.

 Milan Rezac

[+participant]. Stopgaps are derived by language-specific rules deleting the person and number of one of the clitics (impoverishment to 3s) or the clitic itself (obliteration). At the rules’ point of application, morphology is distinct from syntax: the computation has branched off from the mapping to LF so there is no effect on meaning, and it is characterised by distinctive properties such as restriction to the morphological word (cf. Embick and Noyer 2001). However, linearization and exponent insertion has not yet occurred, so linear order and phonological form are irrelevant. Four issues may be taken with the analysis: restriction to 1p+2 forms, distinction between obliteration and impoverishment, degrees of impoverishment, and implementation before linearization. The first three are raised here, and I suggested that the proposal is on the right track for each, while the last is left for the next section. The markedness ban on multiple [+participant] elements would be expected to apply to 1s+2 combinations as well as 1p+2, but it does not, in Bermeo or elsewhere in Basque. The major discussion of similar gaps outside Basque is Heath (1991, 1998), who finds that regular expression of 1+2 combinations widely avoided, and attributes it to social conventions. That should include 1sE-2A par excellence, and does in other languages. Frequency might be the culprit in Basque. Basque 1p+2 gaps and stopgaps seem to have arisen in recent remodellings of agreement paradigms, and these are often resisted by 1s as the most frequent form. Bermeo shows this independently. 1sE is the sole ergative in Bermeo to retain the older mode of past formation, beside the newer one of tacking on -n to the present, and Hualde (2002) attributes it to frequency: The persistence of the etymological form nozan ‘I V it to him, past’ in the face of a complete restructuring of the morphological system must undoubtedly be due to the greater frequency of use of this form. The conservative nature of forms for a first person singular subject due to their great frequency is also pointed out in Bybee & Brewer (1980) … (Hualde 2002)

It is in this perspective that Arregi and Nevins’s coding of the 1s exception might be seen. They impoverish 1s [+participant, +author, +singular] by deleting [+participant], excluding it from the ban on multiple [+participant] clitics. It is a stipulation, but one suitable to capturing the consequences of the frequency of 1s, by making it less complex than other [+participant] elements through deletion of redundant specification. Turning to the distinction between obliteration and impoverishment, it has been seen that forms subsumed under impoverishment of 1p to 3s in ­Bermeo might actually be analysed as obliteration. The Bermeo pattern seems to be the most frequent one, and is found in all the stopgaps seen so far. Yet it is not



Gaps and stopgaps in Basque finite verb agreement 

­ niversal. Two divergences from it are found: clear impoverishment of 1p to 3s u rather than its obliteration, and impoverishment or obliteration of 2. Both are found in Albóniga, a subvariety of Bermeo for which Yrizar (1992a: 466ff.) gives two sources, Azkue (henceforth α) in the first third of the 20th century and Olazar (henceforth ω) in the last. In Albóniga 1pE-2A combinations are mostly retained, present 1pE-2sA α saittugu ω saitxugu. However, for 1pE-2pA, where α has the regular saittuegu, ω uses 3E-2pA saitxusie, not 2pA sarie. The difference is the presence of 3E agreement controller, which conditions root allomorphy, even though it has no overt suffix. So impoverishment of 1pE-2pA yields a 3E-2pA stopgap distinct from the 2pA impersonal look-alike. There is nothing inevitable about stopgaps’ surface identity to impersonals, not even when 1pE is the element targeted by reduction. The same is true of present 2s,pE-1pA combinations, retained by α as gosus, gosues but turned to 3s,pE-1pA in ω gaitxus, gaitxues.45 These are of independent interest for two reasons. First, it is the 2nd person argument that is affected, not as hitherto 1p. This, Arregi and Nevins point out, shores up their analysis, which stipulates which phi-features are reduced. Second, the singular/plural distiction of the 2nd person argument is retained, as 3s/p: we have partial impoverishment, of person only (Antzuola below furnishes more examples). I would advance this as support for the full scope of Arregi and Nevins’ impoverishment mechanics: it is in principle capable of deleting person and number separately, and here we see it for person. Number alone can apparently be partially impoverished as well: α has regular past 1pE-2sA sendugusan? [sic], 1pE-2pA senduegusan, but ω replaces 1pE-2sA by 1sE-2sA saitxuten, impoverishing the number of 1pE, and 1pE-2pA by 3sE-2pA saitxusie [sic, no -n] (again Antzuola will have another example).46

.  20E-1pA forms are regular in α and not used or recorded in ω. .  In past 2s,pE-1pA, α retains gendususan, gendusuesan, ω impoverishes 3s/pE in gendusen, the 3s/pE distinction here not being marked either in the stopgap or in regular 3s/ pE-1pA past. Corresponding 20 forms are regular for α when extant, else not used or recorded. In the E-D-A paradigm of Albóniga, 1p+2 are not affected save when 20 is involved. This also support arbitrarily stipulating reductions: 1p+2 combinations behave differently in the E–A and E-D-A paradigms, and in both can be finely microparametrised. The details are 20m, fE-1pD-3sA;3pA α doskuk, doskunaa; doskusak, doskunas? but ω dok, dona; dok, dosuena, with 1pD obliteration save in the opaque last form, 1pE-20m,fD-3s;pA α duagu, deunagu?; deuagus, deunagus but ω dxak, dxana = 20D-3sA and thus 1pE obliteration, with no 3sA-3pA distinction which does exist for 20D-3s/pA. The corresponding past forms with 20 are mostly not in use or not recorded.

 Milan Rezac

Other variations on this theme are found in the Plentzia varieties of Maruri, Gatika and Butron, differing in minor matters such as whether and when the E-D-A paradigm is included and whether past and present behave the same (Yrizar 1992a: 648ff.). Arregi and Nevins (2011: 4.6.3) observe the relevance of these varieties; I have done no more than add support for the finer nuances their mechanics make available. We may conclude with Arregi and Nevins that any or all 1p+2 combinations may be transformed by “deleting the ergative, deleting the absolutive, deleting the dative, deleting the first plural, deleting the second person”, adding that person and number can delete wholesale or piecemeal. This proposal is utterly stipulative about deletions; the data fit. There are, to be sure, things to say about frequency – the preference of affecting 1p in any capacity to 2 for instance – but these seem external to the mechanics of morphology. This capriciousness in gaps and stopgaps is not unique to Basque. The Chukchi spurious antipassive is a good example (Bobaljik and Branigan 2006: 68, 77n20): (32) Chukchi spurious antipassive: deletion of object agreement phi-features and insertion of a voice-like morpheme in subject→object contexts 3SG→1SG, 2→1 in non-participial tenses, and 1→2, 1/2→3, 3SG→3 in participial tenses.

Such arbitrariness fits well theories of morphology like that of Bonet (1991), Noyer (1992), Embick and Noyer (2001), Arregi and Nevins (2011), some explicitly devised to model it. It contrasts with the profile of syntactic mechanisms (Rezac 2011: chapt. 2). Agreed-on syntactic phenomena never seem to depend on arbitrary combinations of phi-features from multiple arguments: there is no system where phrasal object shift or a by-phrase agent is available only in structures with 1p nominative and 2s dative. Moreover, there is a contrast between 1p+2 gaps/ stopgaps, which reduce surface agreement without consequences for syntax, and more clearly syntactic gaps like the PCC that cannot reduce agreement and do affect syntax. To the extent that this is systematic, a line is to be drawn between syntax and morphology, even if its precise course remains in doubt. 9.  At the sources of gaps in Antzuola 1p+2 gaps in Basque go with idiosyncracies seen elsewhere in 1+2 agreement combinations, which often fail to be the cumulations of exponents expected on the basis of 1/2(+3) forms (Heath 1991, 1998). Deletions and transfers of phi-features, obfuscation of their realization, and resort to impersonal-looking forms, are commonplace. Rhodes (1993: 145) aptly calls such opaque agreement in Algonquian inflectional idioms.



Gaps and stopgaps in Basque finite verb agreement 

Heath (1991, 1998) gives 1+2 gapping a pragmatic explanation: [T]he correct cross-linguistic generalization is a negative one, namely, that transparent 1 ↔ 2 combinations are avoided (Heath 1991). In other words, maximally transparent ‘I saw you’, ‘you saw me’, etc., tend to form negative or taboo targets and are often replaced by more opaque surface forms. This preference for opacity resembles pragmatic restrictions in many languages on the use of transparent 2sg pronominals (thou, tu, etc.), which may become bluntly “familiar,” hence inappropriate in polite discourse, where they are replaced by impersonal, third-person, or morphological “2pl” forms. … In languages with rich pronominal agreement, pragmatic restrictions … focus on the verbal agreement system, and particularly (I suggest) on the transitive 1 ↔ 2 combinations. … [In Mississippi Choctaw] I had difficulties eliciting 1 ↔ 2 forms of transitive verbs (even though they turn out to be formally regular in that language). My first informant … when it came to ‘you hit me’ he balked, saying “We Choctaws don’t talk like that; it sounds like I’m accusing you.” (Heath 1998: 84–5)

Heath also observes that some of the avoidance strategies reduce the surface transitivity of the morphology in various ways, “disguising or deleting a subject or object marker, or fusing the two into a portmanteau”, or resort to impersonals like Basque. The situation in Chinook is worth giving in detail, since as in Basque, different strategies target apparently arbitrary subsets of 1+2 combinations: Chinook (Boas 1911: 580–84, cf. Silverstein 1976: 132) has a basically well-­ behaved transitive agreement system with ergative and absolutive prefix slots. The irregularities occur in 1 → 2 combinations. In 1sg → 2 forms, instead of the regular 1sg prefix n- we get a suppletive ya- … In combinations of 1st nonsingular subject on 2d object, the usual 1st nonsingular ergative markers (e.g., 1pl exclusive ergative ntc-k-, including ergative case marker -k-) are replaced by qa-, which seems to be an indefinite subject (ergative) marker q- plus a linking vowel. Boas further indicates (1911: 584) that a 1st-person agent is omitted in ditransitive verbs when a 2d-person dative object is present: Ø-t-am-l-ō’t–a ‘I will give them to you’ (Ø-3pl-2sg-to-give-Future). (Heath 1998: 91; my italics)

Heath’s proposal speaks to the origin of gaps. But it does not help understand how synchronically certain combinations but not others are gapped, say Ondarroa 2→1p but not 1p→2, Albóniga ω 2s/p-1pA, 1pE-2pA not 1pE-2sA. And it is hard to apply it to stopgaps, for insofar as these look like other forms, they look like forms that would be ungrammatical in their context, say impersonals with 1st/2nd person absolutives. Differently, Arregi and Nevins (2011: 6.4) see in a Zamudio gap and its stopgap the grammaticalization through acquisition of a near-homophony created

 Milan Rezac

by regular sound-change. Their example is 1pE-2sD dotzugu > dotzuu > dotzu, acquired as a variant of 2sD(-3sA) datzu. True or not for this particular form, it does not generalise to other 1p+2 gaps. One might seek causes of gaps in complexity. One type of complexity is featural, and it is inherent in Arregi and Nevins’s (2011) characterisation of gaps as involving two [+participant] elements.47 Other sorts of complexity might be looked at as well. In Bermeo, gaps occur in a system that has recently drastically simplified the agreement complex, but it’s not clear how that would exclude 1p+2 and not say 1p+3p. Other dialects have 1p+2 gaps without remodelling the agreement complex, and some could even be said to have added complexity. So it is in Antzuola (Larrañaga 1998). I set it out in some detail, for gaps in it often emerge where the system gets particularly warped, bringing the discussion back to the stridden and amn’t gap types.48 Antzuola Basque keeps all the intricacies typical of the Basque agreement complex. Particularly relevant here are: (i) ergative displacement ED in hypothetical and past (+ED) but not present (-ED) paradigms, whereby 1/2E is coded as prefix rather than suffix if 1/2A does not control the prefix; (ii) the use of different roots for the auxiliaries, one for indicative-conditional paradigms, *edun, and one for the subjunctive-potential paradigms, egin in western and central dialects like Antzuola for EB *ezan. These two parameters design the system in Table 6.49 Antzuola has undergone several developments that render this system more complex and that are relevant to gaps. All are related to ED.50 EB and the anteced-

.  It would be tempting to differentiate 1s-2 and 1p-2 by complexity of number, since morphologically 2s is plural in Basque from older 2p, while 2p is doubly plural; thus 1p-2s/p would be singled out against 1s-2p by having two [+plural] features. However, in varieties where the older morphologically singular 2s, now used as 20, is in living use, it too falls prey to the 1p-2 gap, as in Zeberio and Zamudio and Albóniga, and 1p-20 is not more complex in number 1s-2s/p. .  The complexity of Antzuola is forbidding, and my prose does not help. The section ends on the larger conclusions that are accessible without the details. .  Antzuola along with many western and central varieties lacks several combinations: the potential present daiket [dezaket] has been replaced by originally hypothetical neike [nezake], and the counterparts of [duket] and [baneza] are not used. .  There are other complexities with respect to EB, notably the differentiation, rare in the west, of the E-D-A paradigm by different roots according to as D is 3rd or 1st/2nd person; simplifications, notably the common western loss (vestiges aside) of 3s/pA distinction if there is a D argument; and other developments, such as redeployment of pA s for the 2s/p distinction in some cases. I omit a handful of odd gaps, like 1pA missing in egin past potential 20E-1pA ekian/eikienan (EB gintzakek/gintzaken).



Gaps and stopgaps in Basque finite verb agreement 

Table 6.  The Basque tense-mood system† *edun (-eu-/-o-) [*edun (-u-)]

egin (-ai-, -ei-) [*ezan (-eza-)]

Indicative

Conditional

Subjunctive

Potential

-ED

Pres.

doT [duT]

– [(du-ke-T)

daiDA-n [dezaDA-n]

→ Nei-ke [deza-ke-T]

+ED

Hyp.

ba-Neu [ba-Nu]

Neu-ke [Nu-ke]

– [ba-Neza]

Nei-ke [Neza-ke]

Past

Neb-an [Nu-en]

– [Nu-ke-en]

Neix-en [Neza-n]

Nei-ki-en [Neza-ke-en]

†1sE

agreement morpheme in caps, EB forms in brackets.

ents of all current Basque varieties are characterised by the following generalization about the position of agreement affixes coding person: 1/2A control prefixes, 1/2E suffixes, save in +ED paradigms, where if there is no 1/2A to control the prefix, 1/2E does so. So ED disrupts the A:prefix-E:suffix pattern by 1/2E:prefix. This disruption is extended in Antzuola through three developments in the +ED paradigms. They are more common in the less frequent potential egin-based paradigms than in the indicative and conditional *edun-based paradigms.51 The developments may be introduced by a monster they create when they join forces. Potential nonpast 1pE-2sA should be s-einke-gu-s (EB z-intzake-gu), with 2sA prefix s and 1pE suffix gu. The form conforms to the even in Antzuola predominant coding of 1/2A by prefix and 1/2E by suffix outside ED. Instead, we find the reverse, g-einke-su, with 1pE prefix and 2sA suffix. Arguably not coincidentally, this very feature combination turns out to be frequently gapped, and a 1pE-3sA stopgap is used instead. Here is what seems to be involved, and how it may have come about.52

.  The subjunctive and imperative are based on egin but are not affected by these developments. The subjunctive is riddled with gaps, but some survivors are surprising given where the rest of the gaps are, e.g. 2sE-1pE gaittusun and 2sE-1pD-3sA daigusun; so also imperative 2pE-1pA gaitzue, 2sE-1pD-3sA igusu. Their irregularity suggests they are not felt as part of the system; presumably the subjunctive is rare as elsewhere in the west. I set them aside henceforth when speaking of egin paradigms. .  The morphemes involved are in bold: prefixes 1s n, 1p g, 2s/p s, 20 Ø, default prefixes past s (EB z) and hypothetical l; suffixes 1s t(a), 1p gu, 2s/p su, 20m/f (k)a/na.

 Milan Rezac

Development 1: ED is partly lost for 2s/pE but not 1s/pE.53 It is a common dialectal development to double the 1/2E prefix of ED by suffix, or to lose ED so that 1/2E is coded by suffix alone, as if ED did not take place, and the prefix is filled by a default morpheme. This dialectal development can be sensitive to the presence, phi-­features, and exponents of E, D, A agreement, tense/mood, and root choice; usually it starts with 2s/pE, perhaps because its prefix z/s- is identical with the past tense default prefix z/s-.54 In Antzuola precisely this has taken place. For *edun, the E–A indicative past shows ED, but the E-D-A indicative past contrasts ED in 1pE-2sD-3A gontsun and ED loss in 2pE-1pD-3A soskusun. For egin, compare potential nonpast 1pE-3sA ED geinke but also ED+suffix geinkiu; 2sE-3pA ED sei(n)ske or seinke, but also ED+suffix seinkesu and sei(n)ketzu, as well as nonED leikesuke, where only the suffix expresses 2pA and l- is the default prefix of the hypothetical (in EB forms, as if zenitzake > zenitzakezu > litzakezu). The outcome is the pattern 2E:suffix-3A versus 1E:prefix-3A, robust in the +ED paradigms both of egin (for E–A, E-D-A) and of *edun (for E-D-A but not E–A). Development 2: The +ED paradigms, especially those of egin, show a tendency to code 2A by suffix rather than prefix. For 20A, this occurs frequently outside Antzuola only as the doubling of the 20A prefix h- > Ø- by the gender suffixes (k)a M, na F. The 20A prefix can be detected in such cases even if Ø-, because it does not allow ED of 1/2E into the prefix slot. In Antzuola, we also get 20A controlling gender suffixes in the +ED paradigms of both *edun and egin, e.g. egin potential nonpast 3sE-20A eikek/n. Remarkably, some such forms allow ED of 1/2E into the prefix slot, as will be seen in Development 3. Except for 20A, coding of A by suffix is extremely rare elsewhere, but in Antzuola it does occur for 2s/pA in the +ED egin paradigm, as in potential nonpast 3pE-2sA seikisue (EB zintzakete), alongside regular forms like 1sE-2sA seinket (EB zintzaket). This extraordinary development is perhaps due to Development 1: the latter yields the opposition 2E:suffix-3A – 1E:prefix-3A, on which might have been restructured older E-1A:prefix – E-2A:prefix as E-1A:prefix – E-2A:suffix. Development 3: ED codes 1/2E by prefix rather than suffix in 1/2E-3A of +ED paradigms. In Antzuola, there is a tendency in +ED paradigms to code 1pE by prefix always and by suffix never. The tendency avails itself of ­Development 2, whereby

.  For 20E, one cannot tell, since the prefix is Ø- from older h-, and use of suffixes to express gender of 20E, type (h)unan (EB huen), is a widespread and independent of ED loss (Rezac 2006: appendices and 3.2.1). .  Thus Pasaia Donibane 2sE-3sD-3sA ED seniyon, ED+suffix seniyosun, ED loss siyosun, but with 1pE only the former two options, geniyon and geniyogun (Agirretxe, Lersundi and Olaetxea 1998: 117).



Gaps and stopgaps in Basque finite verb agreement 

2A is coded by suffix. This is clearest in the potential paradigm based on egin, where this tendency characterises 1pE-2A: for instance nonpast 1pE-20A geinkek/ geinken (EB hintzakegu), 1pE‑2sA geinkesu~geinkisu (EB zintzakegu). It extends even to 1sE-20A, as in nonpast neikek/neiken (EB hintzaket). In the +ED paradigm of *edun, namely the indicative past, only 1pE-20A combinations are affected, giñdduan/giñddunan. There a further twist appears: when 1pE is not coded by a prefix for 1pE-20A, it still resists being coded by a suffix, and 3sE-20A is used, iñdduan/ iñddunan – a gap and its stopgap. Likewise in the remaining +ED paradigms of *edun. Sometimes 1pE is properly suffixal, so 1pE-2sA in conditional protases has the expected form basiñusteu with the 1pE suffix u. But alongside it, there is also found as stopgap 3sE-20A basiñus, and remarkably, 1sE-20A basiñusket, which impoverishes 1pE for number alone (these two stopgaps are replicated in the indicative past, so they are not freak accidents). The restriction of this development to +ED paradigms shows that the coding of 1p(s)E by prefix through ED plays an important role in it. It is restricted to 1E and does not affect not 2E because it is bled by Development 1 which codes 2E by suffix.55 The combination of Developments 2 and 3 leads to such spectacular reversals of the usual E:suffix-A:prefix coding as potential nonpast 1pE-20A geinkek/geinken (EB hintzakegu), 1pE-2sA geinkesu~geinkisu (EB zintzakegu). The usual Basque pattern of A:prefix-E:suffix remains typical even of Antzuola +ED paradigms, e.g. *edun indicative past 1sE-2sA siñustan~siñustan or *egin potential nonpast 1sE-2sA seinket (EB zintzaket).56 The developments increase the ­complexity of

.  Though 1pE here undergoes ED to the prefix left vacant by zero expression of 20A, it cannot be said of Antzuola that 20A never controls the Ø- prefix, though this seems to be true of Zeberio. In Antzuola, ED of E with 20A occurs only sometimes, and 20A never permits the insertion of default tense-mood prefixes that appear with 3A if there is no ED, past s- as in 3sE-3sA seban (contrast 3sE-20A iñdduan) or hypothetical l- as in 3sE-3sA leike (­contrast 3sE-20A eike). 2s/pA coded by suffix does end up using the past s-, but that is because of the diachronic pathway identifying 2s/pA and past prefixes s-. In Zeberio, on the other hand, 20A Ø- has been more generally interpreted as absence, so we regularly get default prefixes, 1sE-20A present indicative doat/donat (EB haut), and ED, 1sE-20A past indicative noan (EB hindudan) (cf. 3sE-20A oan (EB hinduen)), conditional protasis banoa (EB bahindudan) (cf. 3sE-20A baloa (EB bahindu)) (unless 20A forms are gapped, as in ke-forms). The existence of this pattern and its absence elsewhere is relevant to theories of ED (Laka 1993a; Fernández 2001; Hualde 2002; Rezac 2003, 2006; Arregi and Nevins 2011; see also Rebuschi 1983 for ED of allocutives). The prefix wants to be filled, comes to dislike what was originally only a phonological zero due to h- > Ø-, and so copies E-3A forms where ED plugs an originally true zero. .  The result of the reversal is accidentally identical to E-D-A 1pE-2D; but it does not reflect differential object marking where 2A would be dative in case, which the dialect does not have (Etxebarria 1988: 99).

 Milan Rezac

form-based generalizations: rather than keeping prefix-coding of E to 1/2E-3A combinations, it differentiates 1E from 2E and so on. Let us turn now to gaps in Antzuola, and their possible relationship to these developments. As in other varieties, the gaps are concentrated in 1p+2 combinations, and are replaced by stopgaps obliterating or impoverishing one argument. In the E-A paradigm, for egin both 2s/pE-1pA and 1pE-2s/pA are gapped in the way we have seen for other dialects, with partial or full impoverishment. For egin potential present/hypothetical and past 2s/pE-1pA, there is 2s/pE-3sA, and its analogues were seen in the last section.57 For egin potential present/­hypothetical 1pE-2s/pA, the paradigm presents the above-discussed warped forms like geinkisu~geinkesu/geinkesue, with inversal of the usual prefix-suffix coding. However, they are in fact usually gapped and stopgapped by 1pE-3sA: 1pE-2sA geinkisu/geinkesu, very few say. Guk su ekarri geinke [we.E you.A brought 1pE-3sA] is said. Thus, the verbal form does not have zu [2sA] or ABS morpheme. … The verbal form corresponding to 3rd person ABS is used, and, to obtain comprehensibility, instead of the verbal form having the ABS morpheme, the pronoun corresponding to this ABS is mentioned in the sentence … The same for the 1pE-2pA form. (Larrañaga 1998: 99f.; 119)

This description does not reveal whether for 1pE-2pA we get 1pE-3sA geinke or 1pE-3pA geinske, that is whether number is retained. Here however the potential past paradigm provides crucial evidence: number is retained: 1pE-2sA uses 1pE-3sA geinkien, but 1pE-2pA uses 1pE-3pA geinskien. This is confirmed by a comment to the paradigm: “In the verbal forms for 1pE-2sA and 1pE-2pA pronouns, no trace appears of the ABS phrase, thus, they have the form of 1pE-3sA and 1pE-3pA.” (Larrañaga 1998: 120, my italics) This retention of number has already been seen in Albóniga and constitutes partial impoverishment, of number only. I have so far eschewed 1pE-20A forms, which might be expected to be gapped like 1pE-2s/pA are (so in Zeberio, for instance). However, 1pE-20A is not gapped with egin, and instead the geinkesu-type inversion is actually used: potential present geinkek/geinken, past geinkien/geinkienan. This is not surprising. To get geinkesu, we need the extraordinary coding of 2s/pA by suffix (­Development 2) to make the prefix available for 1pE (Development 3). In geinkek, the 20 ka, na

.  Thus for present/hypothetical 2sE-1pA “Zuk-gu ekarri seinke… [you.E us.A brought 2sE-3sA] is used; in this verb form there is no morpheme corresponding to us. Zuk hura [s/he.A] ekarri seinke = zuk gu ekarri seinke.” (Larrañaga 1998: 99–100) For potential past presumably the same happens; the description gives a lacuna. In fact, the relevant 2s/pE-3sA forms are also each found for 2s/pE-3pA, though for the latter there are also distinct forms.



Gaps and stopgaps in Basque finite verb agreement 

suffixes were originally gender markers that in many Basque varieties combine with the 20A prefix h- > Ø- regardless of how well ED is retained. Thus Development 2 was not needed here, only the extension of ED of 1pE to the Ø of 20A (as happens more generally in Zeberio). This yields geinkek – which then makes for a plausible starting point for Development 2 in the coding of 2s/p, specific to Antzuola. So much for 1p+2 E-A combinations with egin. Matters are different for 2A with *edun. The details have been described under Development 3. In 1s/pE-2A, 1pE is sometimes retained as suffix, sometimes undergoes ED which replaces the Ø- of 20A, sometimes is fully impoverished to 3sE, sometimes partly to 1sE. 2s/p/ fE-1pA is fine for *edun throughout, unlike for egin. In E-D-A, 1p+2 gaps fit the pattern seen elsewhere and are independent of Antzuola developments. In the egin (potential) paradigms, we have gaps for 2s/ pE-1pD-3A but not 1pE-2s/pD-3A. The regular form is found for 2pE-1pD-3A seinkigusue (EB zeniezagukete but with ED+suffix, as if *zeniezagukezue). Optionally in this combination, and obligatorily for 2sE-, 1pD is obliterated, so that 2s/pE-3A forms are used like 2sE-1pD → 2sE-3sA seikesu, and overt pronouns disambiguate. For E-D-A of *edun, gaps likewise target 2s/pE-1pD-3A and not 1pE-2s/pD-3A. This starts with the +ED paradigms, but among younger speakers extends to the -ED indicative present and for some to the conditional. Stopgaps obliterate 1pD: suk guri karamelua emon dosu (younger speakers), you.E we.D karamel.A given 2sE-3sA, suk guri karamelua emon doskusu (older speakers), 2sE-1pD-3sA, but both suek guri erregalua ekarri siñuen, ye.E we.D gift.A brought 2sE-3sA. Again, overt pronouns disambiguate. Interestingly enough, gapping by obliteration of D seen in E-D-A is also used for a feature combination that is not 1p+2: 3pE-1sD-3A-20ALLOC (the allocutive conjugation adds gender suffixes controlled by the addressee to forms where the addressee is not an argument). It is well past time to take stock of this remarkable system. Antzuola strengthens earlier conclusions but raises the relevance of the complexity of form to the origin of gaps. Overall, the character of gaps and stopgaps in and outside Antzuola fits Arregi and Nevins’ (2011) language-specific morphological rules, capable of targetting any phi-feature in any context in the morphological word, and similar approaches, as seminally developed by Bonet (1991) for Catalan clitics. There is a synchronically arbitrary selection of 1p+2 combinations for gapping, but also s­poradic gaps elsewhere. There are stopgaps that differ arbitrarily from each other in obliterating person, number, both, or agreement entirely, without repercussions on syntax. For both Bonet (1991) and Arregi and Nevins (2011), morphology manipulates phi-features prior to linearization and

 Milan Rezac

insertion of ­morphophonological m ­ aterial – that is independently of form, of the mapping of morphosyntactic features to morphological positions. That too seems to fit Basque.58 Form does seem key to the emergence of gaps in Antzuola. There are regions of the paradigm where ergative and absolutive coding has been bent out of its earlier and still predominant shape, E:suffix – A:prefix, through a new pattern, ­2:prefix – 1p:suffix. Where the two tendencies cross, gaps can arise and are plugged by stopgaps that resolve the conflict by giving up some of the features involved. Antzuola lets us see this because it has undergone Developments 1–3. Yet these realise tendencies inherent in a range of Basque agreement systems, and it may be that they play a role in the origin of 1p+2 gaps elsewhere. How form gives rise to gaps remains dark: through parsing garden paths, through too many formations to acquire (cf. *stridden), through stigmatisation of certain forms or formations (cf. *amn’t). I will end on a gap in Antzuola that has nothing to do with the foregoing ones, but where such role of form shines through: 1pE-3A gein(s)kien they say easily and frequently, 3sE-1pA geinkien on the other hand with difficulty. It is not a form used a lot. (Larrañaga 1998: 120) Table 7.  1pE-3A and 3E-1pA potential past in Antzuola† Antzuola

EB

1pE-3sA

geinkien

g

en

1pE-3pA

geinskien

g

en

3sE-1pA

gei(n)skien

g

in

3pE-1pA

geinskien

g 1p

†ke

eza

ke

en

it

za

ke

en

t

za

ke

en

in

t

za

ke

te

n

TM

pA

√

KE

3p

PST

is a mood marker

The relevant part of the paradigm is in Table 7. The EB forms are distinguished by material between the 1p prefix g- and the root -za-: the so-called theme marker TM ‑en‑ versus ‑in‑ and the absolutive pluraliser ‑it‑ versus ‑t‑. Their form is sensitive to matters like ergative displacement, and varies greatly across Basque dialects (a survey for *edun is given in Rezac 2006). Their allomorphs differentiate ­feature .  Gaps and stopgaps differ between auxiliary roots, like egin and *edun, but it might suffice to refer to syntactic conditions on root allomorphy, if it is never sensitive to the particular form of agreement morphemes, or roots might have their morphophonological identity even in syntax (Harley and Noyer 1999; Embick 2000; Chomsky 2001).



Gaps and stopgaps in Basque finite verb agreement 

combinations that would collapse otherwise, as EB 3sE-1pA gintzakeen and 1pE-3pA genitzakeen. In some dialects, they sometimes do collapse. The western egin root is particularly susceptible to the collapse of 3E-1/2A (‑ED) with 1/2E-3A (+ED) forms. In Antzuola this ocurs with geinskien. The ambiguous form is fine for expressing one combination of features, 3E-1pA, and gapped for its inverse, 1pE-3A.

References Agirretxe, Joxe Luix, Lersundi, Mikel and Olaetxea Ortzuri. 1998. Pasaiako hizkera. Pasaia: ­Pasaiko Udala. Albizu, Pablo. 1997a. The Syntax of Person Agreement. Ph.D. dissertation draft, University of Southern California. Albizu, Pablo. 1997b. “Generalized Person-Case Constraint: A case for a syntax-driven inflectional morphology.” In Theoretical Issues on the Morphology-Syntax Interface, Myriam Uribe-Etxebarria and Amaya Mendikoetxea (eds.), 1–33. Donostia: UPV/EHU. Albizu, Pablo. 1998. “Impersonal constructions of Basque revisited.” Handout from Mayfest ‘98, July 6–8 1998, University of Maryland. Albizu, Pablo 2001. “Sobre la distribución sintáctica de las formas finitas del verbo vasco.” ASJU 35: 65–106. Albizu, Pablo. 2002. “Basque verbal morphology: Redefining cases.” In Erramu Boneta, Xabier Artiagoitia, Patxi Goenaga, and Joseba A. Lakarra (eds.), 1–19. Bilbao: UPV/EHU. Albizu, Pablo. 2011. “Micro-variación en las construcciones inacusativas de dativo posesivo del euskera.” Handout from Wedisyn 2, University of Barcelona, February 17–18, 2011. Albright, Adam. 2006. “Lexical and morphological conditioning of paradigm gaps.” Ms., MIT. On-line: http://web.mit.edu/albright/www/papers. Anand, Pranav. 2007. De de se. Ph.D. dissertation, MIT. Aramaio, Itziar. 2001. Berriatuako aditz laguntzailea. BA Thesis, Deustuko Unibertsitatea, Bilbo. Arregi, Karlos. 2004. “The have/be alternation in Basque.” Ms., University of Illinois. On-line: home.uchicago.edu/~karlos. Arregi, Karlos and Nevins, Andrew. 2006. “Obliteration vs. impoverishment in the Basque g-/zconstraint.” In Penn Linguistics Colloquium 30: 1–14. Arregi, Karlos and Nevins, Andrew. 2011. Morphotactics: Basque Auxiliaries and the Structure of Spellout. Dordrecht: Springer. Arregui, Anna. 2007. “Being me, being you: pronoun puzzles in modal contexts.” Sinn und Bedeutung 11: 31–45. Arretxe, Jon. 1994. Basauriko euskara. Basauri: Basauriko Udala. Arteatx, Iñigo. 2007. “Euskarazko oharmen-aditzen osagarrietako perpaus jokatugabeak.” Uztaro 63: 31–63. Arteatx, Iñigo. 2012. “Perception verb complements in Basque.” In Noun Phrases and Nominalizations in Basque Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 397–436. Amsterdam: John Benjamins.  doi: 10.1075/la.187.17art Artiagoitia, Xabier. 2000. Hatsarreak eta parametroak lantzen. Vitoria-Gasteiz: UPV/EHU and Arabako Foru Aldundia.

 Milan Rezac Artiagoitia, Xabier. 2003. “Reciprocal and reflexive constructions.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 607–631. Berlin: Walter de Gruyter. Ariztimuño, Borja. 2013. “Finite verbal morphology.” In Basque and Proto-Basque, Mikel ­Martinez-Areta (ed.), 359–428. Frankfurt am Main: Peter Lang. Azkue, Resurrección María de. 1923–1925. Morfología vasca. Bilbo: Editorial Vasca. Baauw, Sergio and Delfitto, Denis. 2005. “New views on reflexivity: Delay effects in Romance.” Probus 17: 145–184.  doi: 10.1515/prbs.2005.17.2.145 Baerman, Matthew, Corbett, Greville G. and Brown, Dunstan (eds.). 2010. Defective Paradigms. Oxford: Oxford University Press. Baker, Mark. 1996. The Polysynthesis Parameter. Oxford: Oxford University Press. Barker, Chris. 1992. “Group terms in English: Representing groups as atoms.” Journal of Semantics 9: 69–93.  doi: 10.1093/jos/9.1.69 Béjar, Susana and Rezac, Milan. 2009. “Cyclic Agree.” Linguistic Inquiry 40: 35–73.

doi: 10.1162/ling.2009.40.1.35

Blanche-Benveniste, Claire. 1975. Recherches en vue d’une théorie de la grammaire française: Essai d’application à la syntaxe des pronoms. Paris: Champion. Boas, Franz. 1911. “Chinook.” In Bureau of American Ethnology Bulletin 40: 559–677. Bobaljik, Jonathan D. and Branigan, Phillip. 2006. “Eccentric agreement and multiple Case checking.” In Ergativity: Emerging Issues, Alana Johns, Diane Massam and Juvenal ­Ndayiragije (eds.), 47–77. Dordrecht: Kluwer.  doi: 10.1007/1-4020-4188-8_3 Bonet, Eulàlia. 1991. Morphology after Syntax. Ph.D. dissertation, MIT. Bonet, Eulàlia. 1994. “The Person-Case constraint: A morphological approach.” MITWPL 22: 33–52. Broadbent, Judith M. 2009. “The *amn’t gap: The view from West Yorkshire.” Journal of Linguistics 45: 251–284.  doi: 10.1017/S0022226709005696 Bresnan, Joan. 2001. “Explaining morphosyntactic competition.” In Handbook of Contemporary Syntactic Theory, Mark Baltin and Chris Collins (eds.), 1–44. Oxford: Blackwell. Büring, Danel. 2005. Binding Theory. Cambridge: Cambridge University Press.

doi: 10.1017/CBO9780511802669

Bybee, Joan and Brewer, Mary A. 1980. “Explanation in morphophonemics: changes in Provençal and Spanish preterite forms.” Lingua 52: 201–242.  doi: 10.1016/0024-3841(80)90035-2 Cabodevilla, Josu. 1991. Aetzen Uskara. Iruñea: Nafarroako Gobernua. Cardinaletti, Anna and Starke, Michael. 1999. “The typology of structural deficiency.” In Clitics in the Languages of Europe, Henk van Riemsdijk (ed.), 145–233. Berlin: Mouton de Gruyter. Chomsky, Noam. 2001. “Derivation by phase.” In Ken Hale: A Life in Language, Michael ­Kenstowicz (ed.), 1–52. Cambridge, MA: MIT Press. Chomsky, Noam. 2013. “Problems of projection.” Lingua 130: 33–49.  doi: 10.1016/j.lingua.2012.12.003 Cinque, Guglielmo. 1988. “On si constructions and the theory of Arb.” Linguistic Inquiry 19: 521–581. Cysouw, Michael, and Fernández Landaluce, Javier. 2012. “On the (im)possibility of partial argument coreference.” Linguistics 50: 765–782.  doi: 10.1515/ling-2012-0024 den Dikken, Marcel. 2006. Relators and Linkers. Cambridge, MA: MIT Press. den Dikken, Marcel. 2007. “Phase extension: A reply.” Theoretical Linguistics 33: 133–63.

doi: 10.1515/TL.2007.022

den Dikken, Marcel, Lipták, Anikó and Zvolenszky, Zsófia. 2001. “On inclusive reference anaphora: New perspectives from Hungarian.” WCCFL 20: 137–149. Diesing, Molly. 1992. Indefinites. Cambridge, MA: MIT Press. Dixon, Robert M.W. 2003. Basic Linguistic Theory 3. Oxford: Oxford University Press.



Gaps and stopgaps in Basque finite verb agreement 

Egaña, Aitzane. 1984. “Bermeoko aditzaren azterketa lorratzak.” Bermeo 4: 13–43. Embick, David. 2000. “Features, syntax, and categories in the Latin perfect.” Linguistic Inquiry 31: 185–230.  doi: 10.1162/002438900554343 Embick, David and Noyer, Rolf. 2001. “Movement operations after syntax.” Linguistic Inquiry. 32–4. 555–595.  doi: 10.1162/002438901753373005 Embick, David and Noyer, Rolf. 2007. “Distributed Morphology and the syntax-morphology interface.” In The Oxford Handbook of Linguistic Interfaces, Gillian Ramchand and Charles Reiss (eds.), 289–324. Oxford: Oxford University Press. Embick, David and Marantz, Alec. 2008. “Architecture and blocking.” Linguistic Inquiry 39: 1–53.  doi: 10.1162/ling.2008.39.1.1 Emonds, Joseph E. 1999. “How clitics license null phrases.” In Clitics in the Languages of Europe, Henk van Riemsdijk (ed.), 291–367. Berlin: Mouton de Gruyter. Epelde, Irantzu. 2003. Larresoroko Euskara. Ph.D. dissertation, UPV-EHU. Etxabe, Karmele and Garmendia, Larraitz. 2003. Zaldibiako euskara: Bertako bizimodua, ohiturak, eta pasadizoak. Zaldibia: Zaldibiako Udala. Etxebarria, Juan Manuel. 1988. Zeberio haraneko euskararen azterketa etno-linguistikoa. Bilbo: Deustuko Unibertsitatea. Etxepare, Ricardo. 2003a. “Urruneko komunztadura eta menpeko infinitiboak euskaraz.” Lapurdum 8: 167–206.  doi: 10.4000/lapurdum.1097 Etxepare, Ricardo. 2003b. “Valency and argument structure in the Basque verb.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 363–384. Berlin: Mouton de Gruyter. Etxepare, Ricardo. 2006. “Number long distance agreement in (substandard) Basque.” In Studies in Basque and Historical Linguistics in Memory of R.L. Trask – R.L. Trasken oroitzapenetan ikerketak euskalaritzaz eta hizkuntzalaritza historikoaz [Supplements of ASJU 40], Joseba A. Lakarra and José Ignacio Hualde: 303–350. Bilbo: UPV/EHU. Etxepare, Ricardo. 2012. Agreement Configurations: Lessons from Distance. Habilitation thesis, IKER, CNRS. Etxepare, Ricardo and Oyharçabal, Bernard. 2013. “Datives and adpositions in northeastern Basque.” In Variation in Datives: A Micro-comparative Perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 50–95. New York: Oxford University Press. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2012. “Denominal necessity modals in Basque.” In Noun Phrases and Nominalization in Basque, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 282–330. Amsterdam: John Benjamins.  doi: 10.1075/la.187 Felser, Claudia. 1999. Verbal Complement Clauses. Amsterdam: John Benjamins.  doi: 10.1075/la.25 Fernández, Beatriz. 2001. “Absolutibo komunztaduradun ergatiboak, absolutibo komunztaduradun datiboak: Ergatiboaren Lekualdatzetik Datiboaren Lekualdatzera.” In On Case and Agreement, Beatriz Fernández and Pablo Albizu (eds.), 147–165. Bilbo: UPV/EHU. Fernández, Beatriz. 2011. “Goi- eta behe-datiboak eta datibo-komunztaduraren murriztapena.” In Euskal dialektologia: lehena eta oraina [Supplements of ASJU 69], Irantzu Epelde (ed.), 37–64, Bilbo: UPV/EHU. Fernandez, Beatriz and Albizu, Pablo. 2006. “Agreement doubling in Basque dialects.” Slides from Workshop on Syntactic Doubling, March 2006, Amsterdam. Fernández, Beatriz, Ortiz de Urbina, Jon and Landa, Josu. 2009. “Datibo komunztaduraren gakoez.” In Beñat Oihartzabali gorazarre. Festschrift for Bernard Oyharçabal [Supplements of ASJU 43], Ricardo Etxepare, Ricardo Gómez and Joseba A. Lakarra (eds.), 352–380. Bilbo: UPV/EHU. Fernández, Beatriz and Ortiz de Urbina, Jon. 2010. Datiboa hiztegian. Bilbao: UPV/EHU.

 Milan Rezac Fodor, Janet Dean and Inoue, Atsu. 1994. “The diagnosis and cure of garden paths.” Journal of Psycholinguistic Research 23: 407–434.  doi: 10.1007/BF02143947 Franco, Jon. 1993. On Object Agreement in Spanish. Ph.D. dissertation, University of Southern California at Los Angeles. Gallego, Ángel. 2010. Phase Theory. Amsterdam: John Benjamins.  doi: 10.1075/la.152 Gaminde, Iñaki, Romero, Asier and Legarra, Hiart. 2012. Gramatika eta hizkuntz bariazioa ­Bermeon. Bermeo: Bermeoko Udala eta Campos Hegaluzea. Gaminde, Iñaki. 1992. Foruko euskararen morfosintaxiaz. Forua: Foruko Udala. Green, Georgia. 1971. “Unspeakable sentences: Book 2”. Linguistic Inquiry 2: 601–2. Haas, Josef. 1909. Neufranzösische Syntax. Halle an der Saale. Hamann, Cornelia. 2011. “Binding and coreference: Views from child language.” In Handbook of Generative Approaches to Language Acquisition, Jill de Villiers and Tom Roeper (eds.), 247–290. Dordrecht: Springer.  doi: 10.1007/978-94-007-1688-9_7 Harley, Heidi and Noyer, Rolf. 1999. “Distributed Morphology.” Glot International 4.4: 3–9. Haspelmath, Martin. 2004. “Explaining the ditransitive Person Case Constraint: A usage-based approach.” Constructions 2. [http://www.constructions-online.de/articles/35] Heap, David. 2008. “Morphosyntactic theory needs Iberian linguistic geography (and vice versa).” Dialectologia 1: 45–64. Heath, Jeffrey. 1991. “Pragmatic disguise in pronominal-affix paradigms.” In Paradigms: The Economy of Inflection, Frans Plank (ed.), 75–89. Berlin: Mouton de Gruyter. Heath, Jeffrey. 1998. “Pragmatic skewing in 1↔2 pronominal combinations in Native American languages.” International Journal of American Linguistics 64: 83–104.  doi: 10.1086/466351 Heim, Irene. 1998. “Anaphora and semantic interpretation: A reinterpretation of Reinhart’s approach.” MITWPL 25: 205–246. Heim, Irene. 2008. “Forks in the road to Rule I.” NELS 38: 339–358. Heger, M. Klaus. 1966. “La conjugaison objective en français et en espagnol.” Persée 1: 19–38. Hualde, José Ignacio. 2002. “On the loss of ergative displacement in Basque and the role of analogy in the development of morphological paradigms.” In The Linguist’s Linguist: A Collection of Papers in Honour of Alexis Manaster Ramer, Fabrice Cavoto (ed.), 219–230. München: Lincom Europa. Hualde, José Ignacio. 2003. “Finite forms.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 205–245. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110895285 Hualde, Jose Ignacio, Elordieta, Gorka and Elordieta, Arantzazu. 1994. The Basque Dialect of Lequeitio. Bilbao: UPV/EHU. Hudson, Richard. 2000. “I amn’t.” Language 76: 297–323. Ibarra, Orreaga. 1995. Ultzamako hizkera. Inguruko euskalkiekiko harremanak. Iruñea: ­Nafarroako Gobernua. Kayne, Richard S. 1975. French Syntax. Cambridge, MA: MIT Press. Kayne, Richard S. 2002. “Pronouns and their antecedents.” In Derivation and Explanation in the Minimalist Program, Samuel David Epstein and T. Daniel Seely (eds.), 133–166. Oxford: Blackwell.  doi: 10.1002/9780470755662.ch7 Lafitte, Pierre. 1979 [1995]. Grammaire Basque (Navarro-Labourdin littéraire). Donostia: Elkar. Laka, Itziar. 1993a. “The structure of inflection.” In Generative Studies in Basque Linguistics, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 21–70. Amsterdam: John Benjamins.

doi: 10.1075/cilt.105.02lak

Laka, Itziar. 1993b. “Unergatives that assign ergative, unaccusatives that assign accusative.” MITWPL 18: 149–172.



Gaps and stopgaps in Basque finite verb agreement 

Laka, Itziar. 2000. “Thetablind Case: Burzio’s Generalisation and its image in the mirror.” In Arguments and Case, Eric Reuland (ed.), 103–129. Amsterdam: John Benjamins.

doi: 10.1075/la.34.08lak

Laka, Enara, Olondo, Leire and Gaminde, Iñaki. 2012. “Bermeoko gazteen euskararen aditz morfologiaz.” Euskalingua 13: 27–36. Lasnik, Howard. 1981. “On two recent treatments of disjoint reference.” Journal of Linguistic Research 1: 48–58. Lambrecht, Knud. 1981. Topic, Antitopic, and Verb Agreement in Non-standard French. ­Amsterdam: John Benjamins.  doi: 10.1075/pb.ii.6 Larrañaga, Jone. 1998. Antzuolako hizkera. Antzuola: Antzuolako Udala. McCawley, James D. 1998. The Syntactic Phenomena of English. Chicago, Ill.: Chicago University Press. Mendikoetxea, Amaya. 1999. “Construcciones con se: medias, pasivas e impersonales.” In Gramática de la lengua española, vol. 2, Ignacio Bosque and Violeta Demonte (eds.), 1631–1722. Madrid: Espasa-Calpe. Mendikoetxea, Amaya. 2008. “Clitic impersonal constructions in Romance.” Transactions of the Philological Society 106: 290–336.  doi: 10.1111/j.1467-968X.2008.00210.x Miller, Philip H. and Sag, Ivan A. 1997. “French clitic movement without clitics or movement.” Natural Language and Linguistic Theory 15.573–639.  doi: 10.1023/A:1005815413834 Morgan, Jerry L. 1972. “Verb agreement as a rule of English.” CLS 8: 278–86. Morin, Yves-Charles. 1978. “Interprétation des pronoms et des réfléchis en français.” In Cahier de linguistique 8: 337–76.  doi: 10.7202/800071ar Nevins, Andrew. 2011. “Marked targets versus marked triggers and impoverishment of the dual.” Linguistic Inquiry 32: 413–444.  doi: 10.1162/LING_a_00052 Nevins, Andrew. 2012. “Haplological dissimilation at distinct stages of exponence.” In The ­Morphology and Phonology of Exponence, Jochen Trommer (ed.), 84–116. Oxford: Oxford University Press.  doi: 10.1093/acprof:oso/9780199573721.003.0003 Newman, Paul. 2000. The Hausa Language. New Haven: Yale University Press. N’Diaye, Genevieve. 1970. Structure du dialect basque de Maya. Mouton: The Hague. Noyer, Rolf. 1992. Features, Positions, and Affixes in Autonomous Morphological Structure. Ph.D. dissertation, MIT. Odria, Ane. in prep. Differential Object Marking in Basque Syntax. Ph.D. dissertation, UPV-EHU. Ormazabal, Javier and Romero, Juan. 2013. “Object clitics, agreement and dialectal variation.” Probus 25: 301–344. Ortiz de Urbina, Jon. 1989. Parameters in the Grammar of Basque. Dordrecht: Foris. Ortiz de Urbina, Jon. 2003a. “Causatives.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 592–607. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110895285 Ortiz de Urbina, Jon. 2003b. “Impersonal clauses.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 572–591. Berlin: Mouton de Gruyter.

doi: 10.1515/9783110895285

Ortiz de Urbina, Jon. 2003c. “Periphrastic constructions.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 284–299. Berlin: Mouton de Gruyter.

doi: 10.1515/9783110895285

Ortiz de Urbina, Jon. 2006. “Old Basque (medio)passives.” ASJU 40: 763–786. Oyharçabal, Beñat. 1993. “Verb agreement with non arguments: On allocutive agreement.” In Generative Studies in Basque Linguistics, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 89–114. Amsterdam: John Benjamins.  doi: 10.1075/cilt.105.04oyh

 Milan Rezac Oyharçabal, Beñat. 2003. “Lexical causatives and causative alternation in Basque.” In Inquiries into the Lexicon-Syntax Relations in Basque. [Supplements of ASJU 46], Beñat Oyharçabal (ed.) 223–253, Bilbo: UPV/EHU. Oyharçabal, Beñat. 2005. “Tokiz kanpoko numero komunztadura.” Iker 17: 395–408. Oyharçabal, Beñat and Etxepare, Ricardo. 2012. “The absence of Person-Case Constraints in early Lapurdian.” In Second Conference of the Luis Michelena Chair, Joseba A. Lakarra, Joaquín Gorrochategui and Blanca Urgell (eds.), 149–170. Vitoria-Gasteiz: UPV-EHU. Perlmutter, David. 1971. Deep and Surface Constraints in Syntax. New York: Rinehart and ­Winston Inc. Postal, Paul M. 1983. “On characterizing French grammatical structure.” Linguistic Analysis 11: 361–417. Postal, Paul M. 1984. “French indirect object cliticisation and SSC/BT.” Linguistic Analysis 14: 111–172. Postal, Paul M. 1990. “French indirect object demotion.” In Studies in Relational Grammar 3, Paul M. Postal and Brian D. Joseph (eds.), 104–200. University of Chicago Press: Chicago. Preminger, Omer. 2009. “Breaking agreements.” Linguistic Inquiry 40: 619–666.

doi: 10.1162/ling.2009.40.4.619

Preminger, Omer. 2012. “The absence of an implicit object from unergatives: New and old evidence from Basque.” Lingua 122: 278–288.  doi: 10.1016/j.lingua.2011.04.007 Pullum, Geoffrey. 2013. “The central question in comparative syntactic metatheory.” Mind and Language 28: 492–521.  doi: 10.1111/mila.12029 Pullum, Geoffrey and Zwicky, Arnold. 1999. “Gerund participles and head-complement inflection conditions.” In The Clause in English, Peter Collins and David Lee (eds.), 251–272. Amsterdam: John Benjamins.  doi: 10.1075/slcs.45.17pul Rebuschi, George. 1983. Structure de l’énoncé en basque. Paris: SELAF. Reinhart, Tanya. 2006. Interface Strategies. Cambridge, MA.: MIT Press. Reinhart, Tanya and Reuland, Eric. 1993. “Reflexivity.” Linguistic Inquiry 24: 657–720. Rezac, Milan. 2003. “The fine structure of cyclic Agree.” Syntax 6: 156–182.

doi: 10.1111/1467-9612.00059

Rezac, Milan. 2006. “Agreement displacement in Basque.” Ms., UPV/EHU. On-line: www.iker. cnrs.fr/rezac-milan.html. Rezac, Milan. 2008a. “Phi-Agree and theta-related Case.” In Phi theory: Phi-features across Interfaces and Modules, Daniel Harbour, David Adger, and Susana Béjar (eds.), 83–129. Oxford: Oxford University Press. Rezac, Milan. 2008b. “The syntax of eccentric agreement: The Person Case Constraint and Absolutive Displacement in Basque.” Natural Language and Linguistic Theory 26: 61–106.  doi: 10.1007/s11049-008-9032-6 Rezac, Milan. 2009. “On the unifiability of repairs of the Person Case Constraint: French, Basque, Georgian, and Chinook.” In Beñat Oihartzabali gorazarre. Festschrift for Bernard Oyharçabal [Supplements of ASJU 43], Ricardo Etxepare, Ricardo Gómez and Joseba A. Lakarra (eds.), 769–790. Bilbo: UPV/EHU. Rezac, Milan. 2010a. “Ineffability through modularity: Gaps in French clitic clusters.” In Defective Paradigms: Missing Forms and What They Tell Us, Matthew Baerman, Greville G. Corbett, and Dunstan Brown (eds.), 151–180. Oxford: Oxford University Press. Rezac, Milan. 2010b. “Review of Agreement Restrictions.” Language 86: 348–353. Rezac, Milan. 2010c. “Phi-Agree vs. movement: Evidence from floating quantifiers.” Linguistic Inquiry 41: 496–508.  doi: 10.1162/LING_a_00007



Gaps and stopgaps in Basque finite verb agreement 

Rezac, Milan. 2011. Phi-features and the Modular Architecture of Language. Dordrecht: Springer.  doi: 10.1007/978-90-481-9698-2 Rezac, Milan. 2013. “Person restrictions in Basque intransitives.” Lapurdum 13: 305–322. Rezac, Milan and Fernández, Beatriz. 2013. “Dative displacement in Basque.” In Variation in Datives: A Micro-comparative Perspective, Beatriz Fernández and Ricardo Etxepare (eds.), 256–282. New York: Oxford University Press. Rezac, Milan and Jouitteau, Mélanie. 2015. “The ways of referential deficiency: French impersonal on and its kin.” Ms., CNRS. Rezac, Milan, Albizu, Pablo and Etxepare, Ricardo. 2014. “The structural ergative of Basque and the theory of Case.” Natural Language and Linguistic Theory 32: 1273–1330.

doi: 10.1007/s11049-014-9239-7

Rhodes, Richard. 1993. “Syntax vs. morphology: A chicken and egg problem.” BLS 19: 139–147. doi: 10.3765/bls.v19i2.1559 Richards, Norvin. 2010. Uttering Trees. Cambridge, MA: MIT Press.

doi: 10.7551/mitpress/9780262013765.001.0001

Rijk, Rudolf P. G. de. 2008. Standard Basque. Cambridge, MA: MIT Press. Rivas, Albert Mario. 1977. A Theory of Clitics. Ph.D. dissertation, MIT. Roelofsen, Floris. 2008. Anaphora Resolved. Ph.D. dissertation, ILLC. Rollo, William. 1925. The Basque Dialect of Marquina. Amsterdam: H. J. Paris. Rooryck, Johan. 2006. “Binding into pronouns.” Lingua 116: 1561–1579.

doi: 10.1016/j.lingua.2005.08.005

Ross, John Robert. 1972. “Doubl-ing.” Linguistic Inquiry 3: 61–86. Rotaetxe, Karmele. 1978. Estudio estructural del euskera de Ondarroa. Durango: Leopoldo Zugaza. Safir, Ken. 2004. The Syntax of Anaphora. Oxford: Oxford University Press.

doi: 10.1093/acprof:oso/9780195166132.001.0001

Safir, Ken. 2013. “Syntax, binding, and patterns of anaphora.” In The Cambridge Handbook of Generative Syntax, Marcel den Dikken (ed.), 515–576. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511804571.020 Schlenker, Philippe. 2005. “Nonredundancy: Towards a semantic reinterpretation of the Binding Theory.” Natural Language Semantics 13: 1–92.  doi: 10.1007/s11050-004-2440-1 Silverstein, Michael. 1976. “Hierarchy of features and ergativity.” In Grammatical Categories in Australian Languages, Robert M.W. Dixon (ed.), 112–171. Canberra: Australian Institute of Aboriginal Studies. Sobin, Nicholas. 1997. “Agreement, default rules, and grammatical viruses.” Linguistic Inquiry 2: 318–343. Stump, Gregory. 2001. Inflectional Morphology: A Theory of Paradigm Structure. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511486333 Suñer, Margarita. 1988. “The role of agreement in clitic-doubled constructions.” Natural Language and Linguistic Theory 6: 391–434. Tranel, Bernard. 1981. Concreteness in Generative Phonology: Evidence from French. Berkeley: University of California Press. Trask, Robert L. 1981. “Basque verbal morphology.” Iker 1: 285–304. Yang, Charles, Kyle Gorman, Preys, Jennifer and Browczyk, Margaret. 2012. “Productivity and paradigmatic gaps.” Handout from NELS 43, CUNY, October 19–21, 2012. On-line: http:// www.ling.upenn.edu/~ycharles. Yrizar, Pedro de. 1981. Contribución a la dialectología de la lengua vasca. Donostia: Gipuzkoako Kutxa.

 Milan Rezac Yrizar, Pedro de. 1991. Morfología del verbo auxiliar gipuzcoano, vol. 2. Bilbao: Euskaltzaindia. Yrizar, Pedro de. 1992a. Morfología del verbo auxiliar vizcaino, vol. 1. Bilbao: Euskaltzaindia. Yrizar, Pedro de. 1992b. Morfología del verbo auxiliar alto navarro septentrional. Bilbao: Euskaltzaindia. Zribi-Hertz, Anne. 1982. “La construction se-moyen du français et son statut dans le triangle: moyen-passif-réfléchi.” Lingvisticae Investigationes 6: 345–401.  doi: 10.1075/li.6.2.05zri Zribi-Hertz, Anne. 2008. “Le médiopassif à SN préverbal en français : pour une approche multifactorielle.” Congrès mondial de linguistique française 2008: 2645–62. Zuazo, Koldo. 1998. “Nafarroako euskal hizkerak.” In Nafarroako hizkerak, Iñaki Camino (ed.), 1–22. Bilbo: UEU.

part 3

Determiners

Nominals in Basque and their existential interpretation Urtzi Etxeberria CNRS – IKER, UMR5478

Assuming that Souletin, the Basque dialect from Zuberoa, is a previous stage compared to Standard Basque (cf. Michelena 1964, Camino 2015; cf. also Manterola 2012, 2015), this paper tries to explain how Basque historically moves from a situation where bare nouns are allowed (Souletin dialect) to a situation where bare nouns are not allowed in argument position (Standard Basque). The reason we move from system A to system B is argued to be due to a semantic weakening and loss of the null D: (i) in Souletin BNs are full DPs with an empty head occupied by a phonetically null D (with indefinite reference and unspecified for number); (ii) this null D loses its semantic features and since it is null it cannot be reanalyzed and gets lost; (iii) in Standard Basque, due to the fact that null D is no longer available, the definite article [-a] -a semantically flexible element, cf. Etxeberria 2005 et seq- takes its place, as is phonologically a weak element, i.e. a suffix, and so the closest phonological alternative to the null D. Keywords:  definite article; bare nouns; existential interpretation; number neutrality

1.  Introduction1 Basque has usually been described as a language where bare nouns (BNs) c­ annot appear in argument position (cf. inter alia, Laka 1993; Artiagoitia 1997, 1998, 2002,

.  The research conducing to this paper has benefited from the Basque Government project IT769-13; from the ANR project ISQI (ANR-2011-JSH2-004-1); from the CNRS project FR2559 from Fèderation Typologie et Universaux Linguistiques; from the University of the Basque Country project UFI11/14; from the MINECO projects FFI2011-23356, FF201126906, and FFI2012-38064-C02-01, FFI2014-52015-P, FFI2014-51878-P; from the MICIN project FFI2011-29218; and the Aquitaine-Euskadi project “La phrase dans la langue basque et les langues voisines”. The research leading to these results has also received funding from the

doi 10.1075/lfab.13.07etx © 2016 John Benjamins Publishing Company

 Urtzi Etxeberria

2006, 2012; Etxeberria 2005, 2010, 2012a; cf. also Section 2).2 ­However, this property appears to show dialectal variation, as in Souletin, the most ­eastern dialect of Basque, spoken in the area of Zuberoa (a dialect which is considered to be an older version of Basque; cf. Michelena 1964; Camino 2015),3 BNs are accepted in direct object position (cf. also Lafon 1954, 1970; Michelena 1987; Coyos 1999; CasenaveHarigile 2006; Etxebarne 2006; Etxegorri 2013; Manterola 2006, 2008). In this paper we first make a thorough description of the differences between the use of Basque nominal expressions (in argument position) in both Standard Basque and the Souletin. Once this is done, the main aim of this paper will be to try to explain how Basque historically moves from a situation where BNs are allowed to a situation where BNs are not allowed in argument position.4 The Basque definite article derived from the distal demonstrative, as expressed in the example in (1), as has also been shown to be the case in many other languages (cf. Azkue 1905; Michelena 1979; Irigoien 1987; Azkarate and Altuna 2001; E ­ txeberria 2005; Manterola 2012; 2015; cf. i.a. Lapesa 1961; Epstein 1994; de M ­ ulder and Carlier 2006 for Spanish or French; cf. also Bechert 1993, ­Himmelmann 2001).

(1) NP *(h)a(r) → NP-a

Nowadays, it is a bound morpheme that takes the phonetic forms [-a] (when ­singular) and [-ak] (when plural).5

European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement n0 613465. I’m very grateful to Beatriz Fernández and Jon Ortiz de Urbina for inviting me to write this paper, and specially, for their patience. Thanks also to the anonymous reviewers. Usual disclaimers apply. .  In this paper I will not be talking about the predicative uses of Basque [nominal+D] constructions. The reader is referred to Zabala (1993, 2003), Artiagoitia (1997, 2012), Eguren (2006, 2012); cf. also Etxeberria (in prep). .  Due to space considerations, I will not be addressing in depth the idea that Souletin Basque is an older version of Standard Basque; cf. Michelena (1964), Camino (2015), a.o. for extensive discussion on this. The reader is referred to Manterola (2012, 2015), where the historical development of the Basque definite article is presented. .  The division that I make in this paper is between Souletin and the rest of the Basque dialects plus Standard Basque. Thus, when I use the term Standard Basque throughout this paper, I will be making reference to Standard Basque plus the rest of Basque dialects, except for Souletin. .  Some authors argue that the plural form of the Basque definite article [-ak] is not morphologically divided into the definite article [-a] and the plural marker [-k] (cf. Goenaga 1980, 1991, Euskaltzaindia 1993, Artiagoitia 1997, 1998, 2002, 2012, Rodriguez 2003, Trask 2003).



Nominals in Basque and their existential interpretation 

(2) liburu-a/-ak book-D.sg/-D.pl ‘the book/the books’

And these are the phonetic forms that the article takes in all dialects. However, there are some other properties of the use of definite article in argumental nominal expressions that show dialectal variation, as will be shown in Sections 2 and 3. The paper is organized as follows: in Section 2 I present the way in which the definite determiner behaves in Standard Basque, where nominals in argument position cannot appear bare. Section 3 concentrates on showing the behavior of nominals in Souletin, where nominals can appear bare and be arguments but only in direct object position. I argue (in line with Etxeberria 2014) that BNs in Souletin project a full DP with an empty D position occupied by a phonetically null D (cf. i.a. Contreras 1986; Longobardi 1994, 2001; Munn and Schmitt 2005; Cyrino and Espinal 2014), which provides an indefinite interpretation with narrow scope and is unspecified for number. Section 4 argues that BNs in Souletin, and in Basque in general, are unspecified for number or number neutral and their semantic type 〈e,t〉, i.e. a predicate denoting set. In Section  5, I argue that the reason why Standard Basque begins to use the definite article [-a(k)] to express existential interpretation with narrow scope is due to a semantic weakening of the Souletin null D and an eventual loss. This loss forces the overt definite article of Standard Basque – a semantically flexible element; cf. Etxeberria 2005, et seq. – to be used in some positions usually reserved for indefinites.6 Section 6 concludes the paper.

Based on Etxeberria (2005, 2010), I assume that number markers and D are base-­generated in different syntactic position. I will not provide arguments for it; the reader is referred to Etxeberria (2005, 2010); see also Eguren (2006). For ease of exposition, I will refer to [-a] and [-ak] as the singular and the plural D respectively. For a different approach where Number is considered to be a feature, cf. Brouchard (1998, 2002), Dobrovie-Sorin (2012). .  Due to space considerations, I will not be talking about weak definites at all in this paper. Interestingly, weak definites do exist in Basque, as shown by the example in (i). (i) Ane-k egunkari-a irakurri zuen. Ane-erg newspaper-D.sg read aux ‘Ane read the newspaper’ Just note that some properties that weak definites are assumed to have (e.g. habitual or institutionalized activity; stereotypical enrichment) do not apply to the readings the definite article is shown to be obtaining in Basque, cf. Section 2. The reader is referred to Etxeberria (in prep.).

 Urtzi Etxeberria

2.  Standard Basque: Nominals in argument position As we mentioned in the introduction, it’s been assumed (cf. Laka 1993; Artiagoitia 1997; 1998, 2002, 2006, 2012; Etxeberria 2005, 2010, 2012a, among many others) that BNs cannot be used in argument position in Standard Basque and that the use of the definite article is necessary if sentences are going to be grammatical as it is shown in (3) and (4) (the presence of the indefinite article or a weak quantifier also makes the sentence grammatical; cf. Etxeberria 2005, 2008, 2012a). This is actually one of the most characteristic properties of the definite article [-a(k)] in Standard Basque. Subject position: (3) a. Ikasle*(-a) garaiz iritsi zen student-D.sg on-time arrive aux ‘The student arrived on time’ b. Ikasle*(-ak) garaiz iritsi ziren student-D.pl on-time arrive aux ‘The students arrived on time’ Direct Object position: (4) a. Ane-k goxokia*(-a) jan zuen Ane.erg candy-D.sg eat aux ‘Ane ate the candy’ b. Ane-k goxoki*(-ak) jan zituen Ane.erg candy-D.pl eat aux ‘Ane ate (the) candies’

If BNs cannot appear in argument position in Standard Basque, a question that arises is how Standard Basque expresses the kind reading or the existential reading that in some languages like can be obtained by means of BNs. Thus, English, for example, can use bare plurals and mass terms to express the kind reading (Fishes appeared 390 million years ago; Silver has the atomic number 47). In Basque, in order to express kinds, the presence of the definite article is necessary; the result is ungrammatical otherwise (cf. Etxeberria 2005 for more on kinds in Basque). (5) a. Dinosauru*(-ak) aspaldi desagertu ziren dinosaur-D.pl long time ago disappear aux ‘Dinosaurs disappeared a long time ago’ b. Nitrogeno*(-a) ugaria da gure unibertsoan nitrogen-D.sg abundant is our universe.in ‘Nitrogen is abundant in our universe’

But this use of the definite article comes as no surprise as many other European languages also make use of the D to express the kind interpretation, e.g. Romance



Nominals in Basque and their existential interpretation 

languages, or Greek (cf. Kleiber 1990; Krifka et al. 1995; Chierchia 1998b; ­Zamparelli 1998; Fara 2001; Dayal 2004; cf. Borik and Espinal to appear for a different proposal on kinds). Another interpretation that English BNs can obtain (also some Romance languages; e.g. Spanish) is the existential interpretation (Paul ate candies; Paul drank beer). And here comes what is really interesting about the Basque definite article: when Basque definite DPs (plurals and masses) fill the direct object slot of stagelevel predicates, the definite DP can, but need not make reference to a specific set and can obtain the so-called existential interpretation. In other words, the object DPs in (6) are ambiguous between an existential and a definite reading (no kind reading possible). (6) a. Ane-k goxoki-ak jan zituen Ane-erg candy-D.pl.abs eat aux ‘Ane ate (the) candies’ b. Ane-k garagardo-a edan zuen Ane-erg beer-D.sg.abs drink aux ‘Ane drank (the) wine’

In other words, if we were to offer English translations, (6a) and (6b) would be ambiguous between a referential and an existential interpretation: (6a) ‘Ane ate the candies’ or ‘Ane ate candies’; (6b) ‘Ane drank the beer’ or ‘Ane drank beer’ (cf. §5; cf. also Artiagoitia 1998, 2002, 2006, 2012; Eguren 2012; Etxeberria 2005, 2010, for alternative (synchronic) analyses). One could think that the reason why Standard Basque uses the definite article [-a(k)] to get the existential interpretation is because Basque does not have indefinite articles, but this conclusion cannot be right, because Basque does possess singular and plural indefinites: bat ‘a, one’ and batzuk ‘some.pl’ (cf. Etxeberria 2005, 2008, 2012a for more on this). It is important to note that the existential interpretation of the object DPs in the examples in (6) has obligatory narrow scope – in opposition to what happens with the indefinites which can get both narrow and wide scope; cf. Etxeberria (2012a) –, as is the case with BNs in the object position in English (cf. i.a. Carlson 1977). Take the examples in (7): (7) a. #Nere lagunak bi arratoi hil zituen ordubetez    my friend.erg two rat kill aux hour-for ‘My friend killed two rats for an hour.’ b. Nere lagunak arratoi-ak hil zituen ordubetez my friend.erg rat-D.pl kill aux hour-for ‘My friend killed rats for an hour.’

The sentence in (7a) can only be interpreted with the indefinite bi arratoi ‘two rats’ having wide scope over the atelic adverbial ordubetez ‘for an hour’, i.e.

 Urtzi Etxeberria

[bi arratoi > adv. for] and asserts that the same two rats were killed (an accomplishment predicate) again and again; a rather strange state of affairs. The sentence in (7b), on the other hand, is completely grammatical. The reading we get is one where my friend killed different rats and the existentially interpreted definite DP must necessarily take narrow scope below the adverbial ordubetez ‘for an hour’, i.e. [adv. for > arratoiak]. The DP object of the sentence in (7b) can also get a wide scope reading, but in this case we would only get a definite reading, cf. (6); and with the definite interpretation of arratoiak, the sentence in (7b) would be as strange a sentence as (7a). 3.  Nominals in Souletin7 3.1  The definite article The definite article of Souletin is a ‘well-behaved’ definite article; it forces a ­kind-level interpretation when combined with kind-level predicates (just like in Standard Basque), and it always forces a referential interpretation in episodic ­contexts with stage-level predicates. Subject Position: (9) a. Ikasle-a garaiz iritsi zen student-D.sg on-time arrive aux ‘The student arrived on time’ b. Ikasle-ak garaiz iritsi ziren student-D.pl on-time arrive aux ‘The students arrived on time’ Direct Object Position: (10) a. Anek goxoki-a jan zuen Ane.erg candy-D.sg.abs eat aux.sg ‘Ane ate the candy’ b. Anek goxoki-ak jan zituen Ane.erg candy-D.pl.abs eat aux.pl ‘Ane ate the candies’

From the sentences above, (10b) is the crucial example as this is the example that differs from Standard Basque. In (6a), from Standard Basque, the definite DP can

.  Thanks to Battittu Coyos, Oihana Larrandaburu, Marylin Recalt, and especially Maider Bedaxagar, for help with the Souletin data.



Nominals in Basque and their existential interpretation 

get both the referential definite reading and the existential interpretation with narrow scope, in Souletin on the other hand this is not possible and goxokiak in (10b) can get but the referential meaning, as shown by the English translation in bold.8 3.2  Existential interpretation In order to get the existential interpretation (with narrow scope) Souletin makes use of BNs (only in direct object position; and these BNs can be mass or count),9 as shown by the examples in (11) (cf. Txillardegi 1977; Coyos 1999; CasenaveHarigile 2006; Etxebarne 2006).10 Recall that BNs cannot be used in argumental position in Standard Basque; cf. examples (4), (6). Direct Object position: (11) a. Bortüan ikusi dit behi mountain.D-in see aux cow ‘I saw cows in the mountain’ b. Dembora da (…) içan deçadan diru time is is-have aux money ‘It’s time for me to have money’ 

(Bourciez 1895)

c. Zer agitü da? Sagar ebatsi dü what happen aux apple steal aux ‘What happened? She/he stole apples’ d. Manexek hur edan dizü. Peiok ogi jan dizü Manex.erg water drink aux Peio.erg bread eat aux ‘Manex drank water. Peio ate bread’  (Norantz)11 e. Gizon batek jan dizü gezi man one.erg eat aux cherry ‘A man ate cherries’ 

(Norantz)

These BNs get an existential interpretation. To be interpreted existentially here means that the BNs in (11) are not referring to a specific set of whatever the NP

.  The indefinite article (singular and plural) is available also in Souletin and it allows both the wide scope and narrow scope interpretations; cf. example (16) below. .  Mass Ns and count Ns are lexically distinguished in Basque (cf. Etxeberria 2005; cf. also Etxeberria in prep). The reader is referred to Etxeberria (2012a) for extensive discussion on mass/count Qs. .  That BNs cannot be used in subject position is shown in subsection 3.3. .  Thanks to Beñat Oyharçabal, Irantzu Epelde and Jasone Salaberria for sharing with me the data that they were collecting for the project Norantz, now available online: http://norantz. org/web/en/bilaketa.

 Urtzi Etxeberria

.

denotes; rather, they seem to be non-specific, weak indefinites, like bare plurals of other languages in this position (e.g. English, Spanish). This paper will be arguing that BNs in Souletin are unspecified for number in that they can refer to an atom or to a sum; cf. §4. The fact that BNs in Souletin appear in direct object position and get an existential interpretation suggests that an analysis along the lines of Longobardi (1994, 2001) might be on the right track. And I think this is correct; in fact, the assumptions that I will be making in order to account for Souletin BNs are the following: (12) (i) A null element exists if it alternates – is part of a paradigm – with one or more phonologically realized morphemes and if each element of the paradigm contributes a distinct semantic value (ii) DP layer must be projected with a null D (Contreras 1986; Longobardi 1994, 2001)12 (iii) The null D has a default existential interpretation (Longobardi 1994, 2001) (iv) Syntactically, null structure is expected to be subject to licensing conditions (e.g. to appear in object position)

These four assumptions entail the syntactic structure of Souletin BNs as in (13) – leaving aside directionality –, i.e. a full DP with an empty D head occupied by a phonetically null D. (13)

DP NP haür ‘child’

D Ø

The next subsection provides more evidence for the structure in (13) by presenting a more complete picture of the behavior of BNs in Souletin. 3.3  Souletin BNs are syntactically DPs i. Not in subject position One of the predictions of the proposal that Souletin BNs are full DPs with an empty D head is that these BNs will only be able to figure in object positions and

.  Cyrino & Espinal (2014) show that in Brazilian Portuguese BNs in preverbal position cannot get an existential interpretation; cf. also Section 3.3. Thanks to an anonymous reviewer for pointing this out to me.



Nominals in Basque and their existential interpretation 

not in subject positions. The prediction is borne out, as BNs in Souletin are not allowed in subject position (neither with ergative nor with absolutive case).13,14 Subject position: (14) a. Ergative: (i) *Ikasle-k hori egin dü    student.erg that.abs do aux (ii) Ikasle-ak hori egin dü student-D.sg.erg that.abs do aux ‘The student did that’ (iii) Ikasle-ek hori egin düe student-D.pl.erg that.abs do aux ‘The students did that’ b. Absolutive: (i) *Ikasle jin da    student.abs come aux (ii) Ikasle-a jin da student-D.sg.abs come aux ‘The student came’ (iii) Ikasle-ak jin dia student-D.pl.abs come aux ‘The students came’

ii. No kind interpretation One other property of Souletin BNs is that they cannot be combined with kindlevel predicates, and the presence of the definite article is necessary in order to make reference to the species as a whole. This property is expected: if null D can only get the existential interpretation, no kind reading will be available for ­Souletin BNs.

.  Most of the speakers that I have interviewed do not accept BNs in subject position of unaccusatives, although some do. In this paper, I will be assuming that BNs cannot be used in subject position. It is important to note also that the analysis that will be proposed in §5 could be maintained even if BNs were accepted in subject position of unaccusatives. .  Note that there is no restriction on having non-specific noun phrases in subject position in Basque. The indefinite eli bat ‘some’ which can appear in subject position (ikasle eli bat jin dia ‘some students came’) can be interpreted both specifically or non-specifically, showing that the restriction on having BNs in subject position has nothing to do with their non-specific nature.

 Urtzi Etxeberria

(15) a. Lehu*(-ak) desagertzera dia lion-D.pl.abs disappear-all aux ‘Lions are about to disappear’ b. Nitrojeno*(-a) paketa da gure lürraldean nitrogen-D.sg.abs abundant is our country-in ‘Nitrogen is abundant in our country’

iii. Narrow scope BNs in Souletin cannot refer specifically to the set denoted by the NP, cf. (11), suggesting that they take obligatory narrow scope, something expected under the null D proposal. Take (16), for example, where we have the indefinite plural eli bat ‘some’ in direct object position. This sentence is ambiguous between a reading where the direct object is assumed to have wide scope over the verb want and where it is possible to make reference to the boys denoted by the object NP and a reading where the direct object is assumed to have narrow scope below the verb want. A consequence of the latter reading is that the sentence in (16) cannot be followed by naming the names of the boys that belong to the set of boys. (16) some > want a. Anek pottiko eli bat nahi dizü ezagutu. Jon, Peru, eta Mikel Ane.erg boy some want aux meet Jon Peru and Mikel ‘Ane wants to meet some boys. Jon, Peru and Mikel’ want > some b. Anek pottiko eli bat nahi dizü ezagutu. #Jon, Peru, Ane.erg boy some want aux meet   Jon Peru eta Mikel and Mikel

‘Ane wants to meet some boys. Jon, Peru and Mikel’

In the sentence in (17), on the other hand, the BN in object position cannot get but the narrow scope reading and it is not possible to refer back to the members of the set denoted by the NP pottiko ‘boy’ by naming them. (17) Anek pottiko nahi dizü ezagutu. #Jon, Peru, eta Mikel Ane.erg boy want aux meet    Jon Peru and Mikel ‘Ane wants to meet boys. Jon, Peru and Mikel’

So these BNs are really non-specific, narrow scope indefinites, equivalent to incorporated nominals in languages that would allow incorporation, e.g. Greenlandic Eskimo (van Geenhoven 1998). In a language like Basque, which does not allow noun incorporation (at least in the constructions we are considering here), upon loss of null D, as we will argue to be the case later, the only strategy to salvage the structure is using the next available element, which is the phonologically weak D [-a]; cf. §6.



Nominals in Basque and their existential interpretation 

iv. Null D is part of a paradigm As evidence for the existence of null D in Souletin, note that this dialect possesses other means to express indefinite/existential readings: zumait ‘some’ or eli bat ‘some’ in (18). (18) Adixkide zumait /eli bat khümitatü dizügü friend some /some invite aux ‘We have invited some friends’

However, there is a difference between the existential interpretation that these indefinite quantifiers force and the one obtained by BNs in object position (cf. (16), (17)): BNs must necessarily take narrow scope, cf. (17). The weak quantifiers in (18), on the other hand, can get wide scope. It is important to recall also that in Souletin the definite article is needed to express both the definite interpretation and the kind interpretation (cf. ­examples (9), (10), (15)). Thus, if this is the case, the null D0 appears to be part of a paradigm (cf. (12i)): i. D [-a(k)]: definite (referential), and kind reading; ii. bat ‘one’, zumait ‘some’, eli bat ‘some’: indefinite readings with wide/narrow scope; iii. null D0: existential reading with narrow scope. v. Souletin and episodic predicates BNs can combine with any kind of episodic predicate and this is evidence for the existence of the null D head in Souletin. According to authors such as i.a. Espinal and McNally (2011), BNs in general are assumed to not be able to combine with predicates of the type break – a real episodic predicate – which do not accept as internal arguments elements of type 〈e,t〉 or incorporated type elements. Thus, full DPs are blocked in incorporation constructions. In fact, when there is an incorporation process, there arise special semantic effects in that the incorporated predicate (V+NP) designates some typical, characterizing, or generic activity. When incorporating BNs are combined with real episodic predicates the sentence is ungrammatical – or at least pragmatically odd. If the nominal expression contains a null D, as we are arguing to be the case in Souletin, an incorporation process will not be possible; and it follows from here that Souletin BNs would show no restriction to combine with real episodic predicates such as break. The prediction is borne out. (19) Gaur goizeko festan, Peiok godalet hautsi dizü (Souletin) today morning.gen party.in Peio.erg glass break aux ‘In the party this morning, Peio broke glasses’

 Urtzi Etxeberria

Assuming that what we proposed above about Souletin BNs is correct, i.e. that syntactically BNs in Souletin are full DPs with an empty D head occupied by a phonetically null D, there are still a couple of questions that have not addressed yet: (i) why can’t Standard Basque make use of BNs in argument position? (ii) are the Souletin and the Standard Basque systems related? And if they are, how are they related? But before we address these questions, one thing that we need to understand is what the denotation of BNs in Souletin, and in Basque in general, is, and this is exactly what we do in the next section. 4.  Souletin (and Standard Basque) BNs’ denotation Observing the Souletin data in the previous section, one could think that the existential reading of the BNs (or bare plurals, taking into account the nominals in object position we are considering in this paper) is necessarily related to plurality. And this does make sense if we consider that: (i) Spanish or English BNs (which can be interpreted existentially) always appear with the plural number marker [-s] (not mass terms); (ii) in Standard Basque, in order to obtain the existential reading we make use of the plural form of the D [-ak], cf. example (6) (with mass terms we would use [-a]). However, this conclusion is not correct. What is important in the existential interpretation of the BNs (in object position) in (11) (repeated one of the example as (20) for convenience) is that they make non-specific reference to what the noun denotes in the real world. (20) Bortüan ikusi dit behi mountain.D-in see aux cow ‘I saw cows in the mountain’

Thus, what we are going to argue in this paper is that BNs in Souletin, and in Basque in general – as will be shown below –, are unspecified for number or number neutral (cf. Jespersen 1924; Chierchia 1998a; Corbett 2000; Dayal 2004; ­Rullmann and You 2006; Wilhelm 2008, etc.), that is, a BN in Basque can be used to make reference to a singularity or to a plurality, (not to a kind; cf. (15), or in other words, be compatible with atomic and non-atomic entailments).15 Their

.  The property of Basque BNs being number neutral resembles East Asian languages like Mandarin, Cantonese, Thai, etc. where BNs can make reference to singularities or to pluralities. However, despite appearances, there are clear-cut differences between the behavior of Basque BNs and the behavior of the BNs of East Asian languages. The reader is referred to Etxeberria (2014).



Nominals in Basque and their existential interpretation 

semantic type will be 〈e,t〉, i.e. a predicate denoting a set, as they cannot appear bare in subject position and are only allowed in direct object position with absolutive case (in Souletin dialect). (21) Denotation of a Basque BN: In a context where the cows are a, b, and c. [[behi]] = {a, b, c, ab, ac, bc, abc}

In what follows we provide evidence for this idea. 4.1  Numerals/Weak Quantifiers + BNs In Basque (both in Souletin and Standard Basque), numerals combine directly with BNs. Thus, in (22a), the phrase ikasle bat ‘one student’ is semantically singular, while in the example in (22b) the phrase hamar ikasle ‘ten students’ is semantically plural. Yet, the noun ikasle ‘student’ remains completely uninflected for number in both cases. (22) Souletin and Standard Basque a. ikasle bat student one ‘one student’ b. hamar ikasle ten student ‘ten students’

Weak quantifiers such as asko ‘many’ or gutxi ‘few’ also combine directly with the BN and make reference to a plurality. And here again, the BN is completely uninflected for number.16 (23) Souletin and Standard Basque a. ikasle asko student many ‘many students’ b. ikasle gutxi student few ‘few students’

.  The weak quantifiers asko ‘many/much’ and gutxi ‘few/a little’ can co-occur both with count and mass terms in Basque. However, this is not important for the point that I’m trying to make here. The reader is referred to Etxeberria (2008, 2012b, in prep), Etxeberria & Etxepare (2012) for more information on this.

 Urtzi Etxeberria

4.2  Predicative uses of BNs In both Souletin and Standard Basque it is possible to use BNs as stage-level predicates, denoting a temporary property; and the same BN, artzain ‘shepherd’ in (24), can be used to predicate of a singularity or of a plurality. (24) Souletin and Standard Basque: a. Miren artzain joan zen Ameriketara Miren shepherd go aux.sg America-to ‘Miren went to America (as) shepherd’ b. Jon eta Miren artzain joan ziren Ameriketara Jon and Miren shepherd go aux.pl America-to ‘Jon and Miren went to America (as) shepherd(s)’

In Souletin, BNs are also used to express individual-level predicates (this is not allowed in Standard Basque; cf. Footnote 2), and again, a BN, haür ‘child’ in this case, can serve as predicate to both singular subjects such as proper names (25a) and plural subjects such as conjoined NPs (25b). Note that this predicative use of BNs is not restricted to capacity nominals (cf. de Swart, Winter and Zwarts 2007) since haür ‘child’ is not a capacity nominal. (25) Souletin: a. Miren haür düzü Miren child is ‘Miren is a child’ b. Miren eta Peru haür tützü Miren and Peru child are ‘Miren and Peru are children’

Before we move on to the next section, recall that at the end of §3.3 we raised two questions and said that it would be interesting to show how to account for the inability to use BNs in Standard Basque and to know whether Souletin and Standard Basque systems are related, and if they are related, how they are related. Let me note as a partial answer to the second question (as I will not get into this in detail in this paper) that one very important assumption that I’m making is that Souletin is closer to Old Basque than the Standard Basque is concerning the nominal system as has been recently shown by Manterola (2012, 2015); cf. Footnote 3. 5.  From Souletin to Standard Basque Thus, we have seen that in present-day Basque there are (at least) two systems when it comes to argumental nominal expressions: (i) Standard Basque, and



Nominals in Basque and their existential interpretation 

(ii) ­Souletin. Recall that in Souletin BNs can appear in direct object position with existential interpretation (and with narrow scope), whereas in Standard Basque BNs cannot be used in argument position and the definite article is used to express definite readings, kind readings, as well as existential readings (with narrow scope). The obvious question that we would like to answer is why/how did Standard Basque (being a later stage of Basque compared to Souletin) begin to use [-a(k)] to express the existential interpretation with narrow scope? In §4 it’s been argued that BNs in Basque are number neutral and that in Souletin they are grammatical in internal argument position with existential interpretation (cf. §4.1). If this is the case, the D does not appear to be making any semantic contribution in the existential reading since it does not provide any kind of definiteness,17 as can be seen in the English translations provided for the sentences in (27).18 (27) a. Anek goxoki-ak jan zituen Ane.erg candy-D.pl.abs eat aux ‘Ane ate candies’ b. Anek ardo-a edan zuen Ane.erg wine-D.sg.abs drink aux ‘Ane drank wine’

Taking all this into account, I believe that what motivated Standard Basque to begin to use [-a(k)]19 in order to get the existential reading (with narrow scope) is twofold: (i) loss of null D, (ii) number morphology. 5.1  Motivation 1: Loss of null D and emergence of -a(k) The null D of the previous stage of Basque, i.e. Souletin dialect, becomes a very weak form semantically as it is non-referential and unspecified for number. This vagueness, i.e. indefiniteness plus number vagueness, eventually dooms this form to loss of its semantic indefinite feature and given that it is a null form, it cannot be reanalyzed and as a consequence it gets lost. The appearance of the overt D in Standard Basque in the indefinite/existential object position, thus, is a kind of reanalysis of the available form D [-a] as a D with existential interpretation (with narrow scope) in these cases. In other words, what I postulate is that the available form D of Basque is reanalyzed and takes over the function of the indefinite null D (cf. Manterola .  Cf. Etxeberria (2005, 2010) for a different synchronic analysis of the Basque [-a(k)]. .  The objects of the sentences in (27) can also get the definite interpretation, cf. (6). I ignore this reading here. .  [-k] is the plural number marker; cf. also Etxeberria (2005, 2010).

 Urtzi Etxeberria

2008 for a possible grammaticisation process -à la Greenberg- of the Basque D; cf. also §5.3). And in fact, this makes sense considering that the Basque D [-a] is a phonologically weak element, and as such, it appears to be the first immediate ‘proximate’ phonologically to the null D.20 As a consequence, Basque can be said to move from the syntactic situation with a null D in (28a) to a syntactic situation where the [-a] takes the place of this null D, (28b). The need to have the D position filled is syntactic: the loss of null D. Syntactically, the definite article [-a] is always a D, but now it also functions as the overt counterpart of the (otherwise) covert indefinite existential in object position, (28). So our proposal is that in this reading the definite article has a weak function: it applies vacuously, i.e., it will be an element of semantic type 〈et,et〉, Thus, in this case, we have an asymmetry between syntax (need to always have functional structure above the NP in Basque, for an NP to function as an argument), and the semantics, which imposes indefinite meaning. Thus, syntactically, the Basque article [-a] has become an element which is necessary in order to create arguments, i.e. unless [-a] is present, nominal expressions cannot be arguments (unless some other element is added, e.g. quantifier, indefinite, etc.). DP

(28) NP haür child

DP D Ø

NP haür child

D -a

An additional function that the Basque definite article has been argued to possess is that of a domain restrictor – a function that D heads may perform crosslinguistically –, where the definite article supplies the context set variable C when combined with strong quantifiers, as in (29) (cf. Etxeberria 2005, 2012b; Etxeberria and Giannakidou 2010, 2014 for arguments in favor of this idea). In these cases, the D applies non-canonically, as it combines with a quantifier of type 〈et,ett〉, not with an element of type 〈e,t〉, as expected.21

.  The reason why Basque did not begin to use the indefinite article bat ‘a/one’ to get the existential reading with narrow scope is probably due to the fact that the indefinite article does have already its own indefinite interpretation(s): existential reading with wide and with narrow scope, in opposition to what happens with Souletin BNs and the existential reading of the Standard Basque D, which cannot get but the narrow scope reading. .  Cf. Etxeberria (2005, 2012b), Etxeberria & Giannakidou (2014) for arguments that prove that these constructions are not DPs of type 〈e〉, but QPs of type 〈et,t〉.



Nominals in Basque and their existential interpretation 

(29) a. mutil guzti-ak(Etxeberria 2005: (37a)) boy all-D.pl b. *mutil guzti c. *mutil-ak guzti

What we have argued is that in the domain restricting function the D does not create a referential expression, but is simply a type-preserving function, i.e. a nonsaturating function, a function that contributes a property. (30) a. [QP [NP mutil N] guzti Q + ak D] b. mutil guzti-ak = (boy) [all (C)] (31) a. ⟦Q⟧ = λP λQ. ∀x P(x) → Q(x) b. ⟦D⟧ = λZ et,ett λP et λQ et Z (P ∩ C) (Q);   Z the relation denoted by Q c. ⟦guzti-ak⟧ = λP Q. ∀x (P(x) ∩ C(x)) → Q(x)

This additional use of the Basque definite article as a domain restrictor provides extra evidence for the fact that the definite article is a semantically flexible element in Basque, one that encompasses classical definite uses, but also weaker functions such as the null existential and domain restriction. In other words, the definite article in Basque in Basque is not always referential (cf. Etxeberria in prep). 5.2  Motivation 2: Number morphology A second motivation why Standard Basque starts using [-a(k)] instead of the ­Souletin null D comes from the fact that Basque begins to mark number morphology explicitly.22 In Souletin (and in Standard Basque, cf. §4), BNs are number neutral and there is no morphological number on the noun itself (as is the case in Spanish or in English where plurality is marked by means of [-s]; cf. Delfitto and

.  According to some authors, e.g. Irigoien (1987), Manterola (2006, 2012), the reason why Standard Basque begins to mark number explicitly by means of [-a(k)] is the result of language contact, as the languages around have overt plural markers, e.g. Spanish and French mark plural number on nouns by means of [-s]. This could of course be the case, however, what is left unexplained is why Souletin did not already take the same route and began to mark plural by means of [-ak], because Souletin is also in contact with French. One possibility would be to think that other languages that are in contact with Souletin, e.g. Occitan’s variant Gascon, would behave just like Souletin in possessing BNs with no number marker, i.e. no plural marker, and in allowing them only in internal argument position. However, this appears not to be correct, as Occitan and its variant Gascon do have plural morphology [-s]. Thanks to Francesc Roca and Xavier Lamuela for help with Occitan data.

 Urtzi Etxeberria

­Schrotten 1991; Brouchard 1998, 2002; Dobrovie-Sorin 2012 for extensive discussion on Number realization and Number interpretation). (32) Bortüan ikusi dit behi. mountain-D.in see aux cow ‘Lit.: I saw cow in the mountain’

It is important to emphasize that Basque possesses a plural marker: [-k]. But this plural marker cannot be applied to nouns directly, as the plural marker is a suffix, and as such categorically as well as phonologically dependent on the presence of another category, in this case, the definite article [-a]. So, unless the definite article is present, the plural marker cannot appear in Basque (cf. Etxeberria 2005, 2010 for extensive discussion on where number is interpreted in Basque). (33) a. *ikasle-k    student-pl b. ikasle-a-k student-D-pl

Thus, the need to mark number on nouns explicitly by means of the plural marker [-k] forces the definite article [-a] to be also present. 5.3  The Basque definite article is on D, not below In §5.1 it has been argued that the Standard Basque article [-a(k)] is syntactically always a D while semantically it is a very flexible element, a property that allows us to account for the various interpretations that it forces, e.g. the existential interpretation. However, one could take another direction, and argue that the Basque determiner [-a(k)] in its existential interpretation is not a determiner but simply a (­singular or plural) number marker. This is in fact a position taken by Artiagoitia (2002, 2006) who claims that Basque DPs have two possible structures depending on the interpretation that they will be getting. When the DP is interpreted existentially, the article will just be filling number specification of DPs; i.e., [-a/-ak] will appear in NumP head. When the DP is interpreted referentially on the other hand, [-a/-ak] must appear in D position. But this analysis is problematic:23 Take example (27b), repeated here as (34) for convenience, with the mass term ardo ‘wine’ in object position. In this situation, as already shown, the object DP ardoa can obtain two interpretations: definite or existential (cf. §2).

.  The reader is referred to Etxeberria (2005) for extensive discussion against Artiagoitia’s analysis.



Nominals in Basque and their existential interpretation 

(34) Anek ardo-a edan zuen Ane.erg wine-D.sg.abs drink aux √ Definite: ‘Ane drank the wine’ √ Existential: ‘Ane drank wine’

In the existential interpretation, Artiagoitia does not treat [-a] as a D head, rather, the article would be placed in [Head, NumP] position and would function as a number marker, a singular number marker. But the mass terms in sentences such as (34) does not denote a singularity. What I think is that mass terms are not number marked (in line with Etxeberria 2005, 2010; cf. also Delfitto and Schroten 1991; Doetjes 1997; Dayal 2004; Krifka 2004, etc.). One other possibility could be to argue that the Basque D has reached the final position of the grammaticisation process proposed by Greenberg (1978) (cf. also Himmelmann 2001) -expressed in (35). In other words, to argue that the Basque article is just a noun marker and that syntactically it would appear in a lower position, closer to the noun. (35) dem → def.art → spec.art → noun marker

However, this proposal cannot be correct as [-a(k)] does not appear with every noun: it does not appear attached to nouns when these are combined with numerals or with weak quantifiers. (36) a. bi/bost/hamar ikasle two/five/ten student ‘two/five/ten students’ b. ikasle asko/gutxi student many/few ‘many/few students’

6.  Conclusions Assuming that Souletin is a previous stage compared to Standard Basque when it comes to the D system (as shown by Manterola 2012, 2015; cf. Michelena 1964; Camino 2015), it has been argued here that Basque historically derived from a stage where BNs were allowed in internal argument position (i.e. object position) to a stage where BNs in argument position are completely ungrammatical, and the definite article is introduced to express existential interpretation with narrow scope. In support of this analysis, the paper first made a thorough description of the use of BNs, and of the use of the definite article in both Souletin and in Standard Basque. I also argued that in Souletin, BNs are full DPs with an empty head

 Urtzi Etxeberria

occupied by a phonetically null D – with indefinite reference and unspecified for number. The result is an analysis of the Basque definite article as a semantically flexible element, one that can take up classical definite functions, but also weaker functions (indefinite – with narrow scope –, as argued here, and non-saturating domain restriction, as argued in earlier work, Etxeberria (2005, 2012b), Etxeberria and Giannakidou (2010, 2014). This paper also provided evidence that BNs in Basque are number neutral, i.e., a BN in Basque can be used to make reference to a singularity or to a plurality (cf. Jespersen 1924; Chierchia 1998a; Corbett 2000; Dayal 2004; Rullman and You 2006; Wilhelm 2008, etc.). Finally, this paper argues that the reason we move from a system like Souletin to a system like Standard Basque is basically due to a semantic weakening and loss of the null D: (i) in Souletin BNs are full DPs with an empty head occupied by a phonetically null D; (ii) this null D loses its semantic features and since it is null it cannot be reanalyzed and gets lost; (iii) in Standard Basque the definite article [-a] is reanalyzed and takes the place of the null D and begins to acquire the function of the lost null D.

References Artiagoitia, Xabier. 1997. “DP predicates in Basque.” In University of Washington Working Papers on Linguistics 15, A. Taff (ed.), 161–198. Washington: University of Washington. Artiagoitia, Xabier. 1998. “Determinatzaile Sintagmaren Hipotesia Euskal Gramatikan.” Uztaro 27: 33–61. Artiagoitia, Xabier. 2002. “The functional structure of Basque noun phrases.” In Erramu Boneta: Festschrift for Rudolf P. G. de Rijk, Xabier Artiagoitia, Patxi Goenaga and Joseba A. Lakarra (eds.), 73–90. Vitoria-Gasteiz: ASJU. Artiagoitia, Xabier. 2006. Euskarazko izen sintagma: arkitektura eta egitura funtzionala. Ms., UPV/EHU. Artiagoitia, Xabier. 2012. “The DP hypothesis in the grammar of Basque.” In NPs and Nominalizations in Basque: Syntax and Semantics, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 21–78. Amsterdam: John Benjamins.  doi: 10.1075/la.187.04art Azkarate, M. & P. Altuna. 2001. Euskal morfologiaren historia. Donostia: Elkar. Azkue, R.M. 1905. Diccionario Vasco-Español-Francés. Bilbo. Bechert, Johannes. 1993. “Definiteness and article systems.” Eurotyp Working Papers I/4. Bouchard, D. 1998. The distribution and interpretation of adjectives in French: a consequence of bare phrase structure. Probus 10, 139–183. Brouchard, Denis. 2002. Adjectives, number and interfaces – why languages vary. Amsterdam: Elsevier. Camino, Iñaki. 2015. “Ekialdeko euskararen iragana.” In Euskal dialektologia: lehena eta oraina. [Supplements of ASJU 69], Irantzu Epelde (eds.), 87–153. Bilbo: UPV/EHU. Carlson, Greg. 1977. Reference to kinds in English. Ph.D. dissertation, UMass. Casenave-Harigile, Junes. 2006. Xiberotar gramatika llaburra. Maule: Sü Azia.



Nominals in Basque and their existential interpretation 

Chierchia, Gennaro. 1998a. “Plurality of mass nouns and the notion of ‘semantic parameter’.” In Events and Grammar, Susan Rothstein (ed.), 53–103. Dordrecht: Kluwer.

doi: 10.1007/978-94-011-3969-4_4

Chierchia, Gennaro. 1998b. “Reference to kind across languages.” Natural Language Semantics 6: 339–405.  doi: 10.1023/A:1008324218506 Contreras, Heles. 1986. “Spanish Bare NPs and the ECP.” In Generative Studies in Spanish Syntax, Ivonne Bordelois, Heles Contreras and Karen Zagona (eds.), 25–49. Dordrecht: Foris. Corbett, Greville G. 2000. Number. Cambridge: CUP.  doi: 10.1017/CBO9781139164344 Coyos, Battittu. 1999. Le parler basque souletin des Arbailles: Une approche de l’ergativité. Paris: L’Harmattan. Cyrino, Sonia and Espinal, M.Teresa. 2014. “Bare nominals in Brazilian Portuguese: more on the DP/NP analysis.” Natural Language and Linguistic Theory 33: 471–521.

doi: 10.1007/s11049-014-9264-6

Dayal, Veneeta. 2004. “Number Marking and (In)definiteness in Kind Terms.” Linguistics and Philosophy 27: 393–450.  doi: 10.1023/B:LING.0000024420.80324.67 Delfitto, Denis. and Schroten, Jan. 1991. “Bare plurals and the number affix in DP.” Probus 3: 155–185.  doi: 10.1515/prbs.1991.3.2.155 Dobrovie-Sorin, Carmen. 2012. “Number as a feature.” In Functional Heads, L. Brugé et al. (eds.), 304–324. Oxford: Oxford University Press.  doi: 10.1093/acprof:oso/9780199746736.003.0024 Doetjes, J. 1997. Quantifiers and Selection. On the Distribution of Quantifying Expressions in French, Dutch and English. The Hague: Holland Academic Graphics. Eguren, Luis. 2006. “Non-canonical uses of the article in Basque.” In Proceedings of the 32nd Annual Meeting of the Berkeley Linguistics Society, M.J. Hauser et al. (eds.). Berkeley. Eguren, Luis. 2012. “Predication markers in Basque.” In NPs and Nominalizations in Basque: Syntax and Semantics, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 243–266. Amsterdam: John Benjamins.  doi: 10.1075/la.187.11egu Epstein, Richard. 1994. “The development of the definite article in French.” In Perspectives on grammaticalization [Current Issues in Linguistic Theory 109], William Pagliuca (ed.), 63–80. Amsterdam: John Benjamins.  doi: 10.1075/cilt.109.06eps Espinal, M.Teresa and McNally, Louise. 2011. “Bare nominals and incorporating verbs in ­Spanish and Catalan.” Journal of Linguistics 47(1): 87–128.  doi: 10.1017/S0022226710000228 Etxebarne, Jüje. 2006. Xiberotar gramatika. Maule: Sü Azia. Etxeberria, Urtzi. 2005. Quantification and domain restriction in Basque. Ph.D. dissertation, UPV/EHU. Etxeberria, Urtzi. 2008. “On quantification in Basque and on how some languages restrict their quantificational domain overtly.” In Quantification: A crosslinguistic perspective, Lisa ­Matthewson (ed.), 225–276. London: Emerald. Etxeberria, Urtzi. 2010. “Making a definite be interpreted as an indefinite.” In Thematic Volume on Bare Nominals – Journal of Portuguese Linguistics, Roberta Pires and Carmen DobrovieSorin (eds.), 29–52. Lisboa: Univ. of Lisboa. Etxeberria, Urtzi. 2012a. “Quantification in Basque.” In Handbook of Quantifiers in Natural Language [Studies in Linguistics and Philosophy, vol 90], Edward Keenan and Denis Paperno (eds.), 83–164. Berlin: Springer.  doi: 10.1007/978-94-007-2681-9_3 Etxeberria, Urtzi. 2012b. “The way the definite determiner affects quantifiers in Basque (and beyond).” In Noun Phrase and Nominalizations in Basque, Urtzi Etxeberria, Ricardo ­Etxepare and Myriam Uribe-Etxebarria (eds.), 79–110. Amsterdam: John Benjamins.

doi: 10.1075/la.187.05etx

 Urtzi Etxeberria Etxeberria, Urtzi. 2014. “Basque nominals: from a system with bare nouns to a system w ­ ithout.” In Weak Referentiality, Ana Aguilar-Guevara, Bert Le Bruyn, and Joost Zwarts (eds.), 335–364. Amsterdam: John Benjamins. Linguistics Today Series.  doi: 10.1075/la.219.14etx Etxeberria, Urtzi. in prep. Nominal expressions in Basque. CNRS-IKER. Etxeberria, Urtzi and Giannakidou, Anastasia. 2010. “Contextual Restriction and the Definite Determiner.” In Context-Dependence, Perspective and Relativity, Francois Recanati, Isidora Stojanovic and Neftali Villanueva (eds.), 93–126. Berlin: Mouton de Gruyter. Etxeberria, Urtzi and Giannakidou, Anastasia. 2014. “D-heads, domain restriction, and variation: from Greek and Basque to St’át’imcets Salish.” In The Structure of NP/DP in Slavic and beyond, Lilia Schürcks, Urtzi Etxeberria, and Anastasia Giannakidou (eds.). Berlin: Mouton de Gruyter. Etxegorri, Philippe. 2013. Biarnoko Euskaldunak. Zingi-Zango Legazpiko Euskara Elkartea/ Eusko Jaurlaritza. Euskaltzaindia. 1993. Euskal Gramatika Laburra: Perpaus Bakuna, Bilbo: Euskaltzaindia. Fara, Delia Graff. 2001. “Descriptions as Predicates.” Philosophical Studies 102: 1–42.

doi: 10.1023/A:1010379409594

van Geenhoven, Veerle. 1998. Semantic Incorporation and Indefinite Descriptions: Semantic and Syntactic Aspects of Noun Incorporation in West Greenlandic. Stanford, CA: CSLI Publications. Goenaga, Patxi. 1980. Gramatika Bideetan. Donostia: Erein. Goenaga, Patxi. 1991. “Izen Sintagmaren Egituraz.” In Memoriae L. Michelena Magistri Sacrum, Joseba A. Lakarra and Iñigo Ruiz (ed.), 847–865. Donostia: UPV/EHU. Greenberg, Joseph H. 1978. “How does a language acquire gender markers?” In Universals of Human Language, J. H. Greenberg et al. (eds.), 47–82. Stanford: Stanford University Press. Himmelmann, Nikolaus P. 2001. “Articles.” In Language Typology and Language Universals: An International Handbook, Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), 831–841. Berlin: Walter de Gruyter. Irigoien, Alfonso. 1987. “Euskarako izen sintagma mugatzailerik gabekoez.” Euskera XXX: 129–139. Jespersen, Otto. 1924. The philosophy of grammar. London: Allen and Unwin. Kleiber, Georges. 1990. L’article LE generique sur le mode massif. Geneve-Paris: Droz. Krifka, Manfred et al. 1995. “Genericity: an introduction.” In The generic book, Gregory Carlson and Francis Jeffry Pelletier (eds.), 1–124. Chicago: The University of Chicago Press. Krifka, M. 2004- “Bare NPs: Kind-referring, Indefinites, Both, or Neither?”. In Empirical Issues in Formal Syntax and Semantics5; Selected Papers from CSSP 2003, Bonami, O. & P. CabredoHofherr (eds.). Paris. Lafon, René. 1954. “Le nombre dans la déclinaison basque.” Via Domitia 1: 111–121. Lafon, René. 1970. “Structure de la déclinaison basque.” Actes du Xeme Congrès International del Linguistes. 279–301. Laka, Itziar. 1993. “Unergatives that assign ergative, unaccusative that assign accusative.” In Papers on case and agreement I, [MITWPL 18], Jonathan D. Bobaljik and Colin Phillips (eds.), 149–172. Cambridge, MA: MIT. Lapesa, Rafael. 1961. “Del demostrativo al artículo.” Nueva Revista de Filología Hispánica 15: 23–44. Longobardi, Giuseppe. 1994. “Reference and proper names: A theory of N-movement in syntax and logical form.” Linguistic Inquiry 25.4: 609–665. Longobardi, Giuseppe. 2001. “The structure of DPs: Some principles, parameters, and problems.” In The Handbook of Contemporary Linguistic Theory, Mark Baltin and Chris Collins (eds.), 562–603. London: Blackwell.  doi: 10.1002/9780470756416.ch18



Nominals in Basque and their existential interpretation 

Manterola, Julen. 2006. “-a euskal artikulu definituaren gainean zenbait ohar.” In Studies in Basque and Historical Linguistics in Memory of R.L. Trask – R.L. Trasken oroitzapenetan ikerketak euskalaritzaz eta hizkuntzalaritza historikoaz, Joseba A. Lakarra and José Ignacio Hualde (eds.), 651–676. Donostia-Bilbo: Gipuzkoako Foru Aldundia-UPV/EHU. Manterola, Julen. 2008. “-a morfemaren erabilera (eza) ekialdeko euskaretan.” In Koldo Mitxeleran Katedraren II. Biltzarraren Aktak, Joseba A. Lakarra, Joaquín Gorrotxategi and Blanca Urgell (eds.). Bilbo: UPV/EHU. Manterola, Julen. 2012. “Synchronic ubiquity of the Basque article -a: A look from diachrony.” In NPs and Nominalizations in Basque: Syntax and Semantics, Urtzi Etxeberria, Ricardo Etxepare and Myriam Uribe-Etxebarria (eds.), 179–208. Amsterdam: John Benjamins.

doi: 10.1075/la.187.08man

Manterola, Julen. 2015. Euskararen morfologia historikorako: artikuluak eta erakusleak. Ph.D. dissertation, University of the Basque Country (UPV/EHU). Michelena, Luis. 1964. Sobre el pasado de la lengua vasca. Donostia: Auñamendi. Michelena, Luis. 1979. ‘Miscelanea Filológica Vasca IV’. Fontes Linguae Vasconum XI, 33. 377–406. Michelena, Luis. 1987. Palabras y Textos. Bilbao: University of the Basque Country UPV/EHU. De Mulder, Walter and Carlier, Anne (2006). “Du démonstratif à l’article défini: le cas de ce en français moderne.” Langue française 152: 96–1. Munn, Alan and Schmitt, Christina. 2005. “Number and indefinites.” Lingua 115: 821–855.

doi: 10.1016/j.lingua.2004.01.007

Oyharçabal, Bernard. 1994. “Contribution de la comparaison typologique à une analyse des rapports ergativité-(in)transitivité en basque.” In La langue basque parmi les autres (influences et comparaison), Jean Baptiste Orpustan (ed.), 115–148. Saint-Etienne-de-Baigorry: Izpegi. Rodriguez, Sonia. 2003. Euskal artikuluaren sintaxiaz. Ms., UPV/EHU. Rullmann, Hotze and You, A. 2006. “General number and the semantics and pragmatics of indefinite bare nouns in Mandarin Chinese.” In Where Semantics Meets Pragmatics, Klaus von Heusinger and Ken Turner (eds.), 175–196. Amsterdam: Elsevier. de Swart, Henriette, Winter, Yoad, and Zwarts, Joost. 2007. “Bare nominals and reference capacities.” Natural Language and Linguistic Theory, 25(1): 195–222.

doi: 10.1007/s11049-006-9007-4

Trask, Larry. 2003. “The Noun Phrase: Nouns, Determiners, and Modifiers; Pronouns and Names’.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 92–134. Berlin: Mouton de Gruyter. Txillardegi. 1977. “L’emploi de l’indéfini en souletin.” Fontes Linguae Vasconum 25: 29–54. Wilhelm, Andrea. 2008. Bare nouns and number in Dene Suline. Natural Language Semantics 16: 39–68.  doi: 10.1007/s11050-007-9024-9 Zabala, Igone. 1993. Predikazioaren teoriak gramatika sortzailean. Ph.D. dissertation, UPV/ EHU. Zabala, Igone. 2003. “Nominal predication: copulative sentences and secondary predication.” In A grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 324–339. Berlin: Mouton de Gruyter. Zamparelli, Roberto. 1998. “A Theory of Kinds, Partitives and Of/Z Possessives.” In Possessors, Predicates and Movement in the Determiner Phrase, Artemis Alexiadou and Chris Wilder (eds.), 259–304. Amsterdam: John Benjamins.  doi: 10.1075/la.22.10zam

part 4

Word order and left periphery

Strategies of verb and verb phrase focus across Basque dialects Arantzazu Elordieta & Bill Haddican UPV/EHU / CUNY-Queens College/Graduate Center

In this chapter we study four different strategies available in Basque dialects to focalize the verb and the verb phrase: (i) verb doubling, (ii) ba-insertion, (iii) V1 strategy, and (iv) expletive egin insertion. We argue that the four constructions reflect different ways of complying with morphophonological requirements of foci, including the requirement that foci be capable of bearing sentence stress. In particular, we argue that these strategies reflect selective deletion of material (PF repair operations) that would otherwise violate these constraints (Bošković and Nunes 2007). In so doing, the discussion provides a partially unified analysis of these four constructions. Keywords:  stress; focus; scattered deletion; expletive insertion

1.  Introduction1 This chapter analyzes four verb (phrase) focalization constructions, which vary across dialects and also by verb class and focus interpretation. We argue that the four constructions reflect different ways of complying with ­morphophonological requirements of foci, including the requirement that foci be capable of bearing sentence stress. In particular, we argue that these strategies reflect selective .  We are grateful to the following people for help with judgements and/or comments on earlier versions of the analysis: Gotzon Aurrekoetxea, Maia Duguine, Esther Elgoibar, Gorka Elordieta, Urtzi Etxebarria, Ricardo Etxepare, Beatriz Fernández, Iñaki Gaminde, Melanie Jouitteau, Richard S. Kayne, Bernard Oyharçabal, Koldo Zuazo, Irune Zuluaga, Miren Zuluaga, the audiences of PLC 36, Wedisyn’s 3rd Workshop on Syntactic variation 2013, and BLS 40, as well as to two very insightful anonymous reviewers whose thorough helpful comments have much improved the paper and raised interesting questions, only some of which we have been able to address in this draft. Work on this chapter was supported by earlier grant FFI200800240/FILO, and by recently awarded grants FFI2014-51878-P and FFI2013-46907-P from the Spanish Ministry of Economy and Competitiveness, and a grant from the Basque Government GIC12/61 IT769-13. All errors are our own.

doi 10.1075/lfab.13.08elo © 2016 John Benjamins Publishing Company

 Arantzazu Elordieta & Bill Haddican

­ eletion of material that would otherwise violate these constraints (Bošković and d Nunes 2007). In so doing, the discussion provides a partially unified analysis of these four constructions. The discussion is organized as follows. In Section  2, we describe the verb doubling strategy found in a handful of western dialects. This strategy, which is restricted to synthetic verb constructions, expresses polarity and contrastive focus. Section 3 describes the ba- insertion strategy, which applies in synthetic verb constructions in a larger set of dialects and has a set of interpretations similar to the verb doubling strategy. Section 4 presents the V1 strategy, which, across dialects, marks polarity focus and in some dialects is used to mark event focus. Section 5 deals with expletive egin insertion, which applies in analytic constructions with new information focus and contrastive focus on the verb or VP. Section 6 summarizes the chapter and discusses remaining challenges for the analysis presented. 2.  The verb doubling strategy We begin by describing the verb doubling construction. As illustrated in (1)–(5), this construction involves doubling of the verb root in a restricted set of contexts to be described shortly. In such constructions, the word containing the verb marker, verb root and infinitival marker appears to the left of a copy of the verb root along with tense and agreement morphology. The first occurrence of the verb obligatorily bears main intonational prominence.2 (1) Ni-k jakin daki-t egia I-erg know.inf know.3sg-1sg truth ‘I know the truth’ [as opposed to‘think’ or ‘believe it’]  (Etxepare & Ortiz de Urbina 2003: 476, (831)) (2) Juen doie, ala etorr-i dator, ba? Go.inf go.3sg or come-inf come.3sg then ‘Well, is he leaving (right now), or is he coming?’  (Mallabia, Zuazo 1998: 207) (3) Orasiño i-tten bere yakin daki-tzu faltzo horrek pray do-imprf too know.inf know.3sg-to you false that ‘That fake (guy) even knows to pray’ (Urduliz, Gilisasti 1997)

.  We use the following abbreviations throughout the article: abs = absolutive, aux = auxiliary, c = complementizer, dat = dative, det = determiner, erg = ergative, evid = evidential, imprf =imperfective, inf = infinitive, loc = locative, neg = negation, pl = plural, prf = perfective, prog = progressive, pst = past, rel = relativizer, sg = singular.



Strategies of verb and verb phrase focus across Basque dialects 

(4) Ibil-i dabil beti kale-a-n walk-inf walk.3sg always street-det-loc ‘She is always WALKING in the street’/ ‘She IS always WALKING in the street’ (5) Kostaten da-na kostaten da-la ekarr-i dakar-gu Cost-inf aux-rel cost aux-c bring-inf bring.3sg-1pl te kittu and that’s it ‘Whatever it costs, we will indeed bring it, and end of the story’  (Urduliz, Gilisasti 1997)

Such sentences are interpreted with emphasis on the verb, which some speakers interpret as contrastive focus on the verb, as reflected in the translations for (1)–(4) (Etxepare and Ortiz de Urbina 2003: 476), but for some speakers it can also express positive polarity focus (affirmative denial), as reflected in the translations in (4)–(5) (Altube 1929; Arejita 1980; Euskaltzaindia 1985; Osa 1990). Doubling sentences are infelicitous in out of the blue (‘What happened?’) contexts and as answers to wh-questions questioning the verb, ‘What is X doing?’. In the relevant dialects, either verb doubling or the ba strategy, discussed in Section 3, is obligatory in synthetic contexts on these interpretations. Sentences with a bare verb on such interpretations, on any stress pattern, are bad: (6) *Ni-k, daki-t egia    I-erg know-3sg.1sg truth ‘As for me, I KNOW the truth’

(as opposed to‘think’ or ‘believe it’)

The verb doubling construction is restricted to a handful of Western dialects including Urduliz-Gatika, Mallabia and Arratia dialects, and for some speakers in dialects further east between Western and Central Basque (Altube 1929; Arejita, 1980, 1984, 1988; Aurrekoetxea 1995; Osa, 1990; Zuazo, 1998, 2003, 2008, 2010). In addition, the verb doubling construction is restricted syntactically in several ways, which we describe below. A first restriction on the doubling construction concerns verb class. A closed class of verbs in Basque is formed synthetically such that in imperfective finite contexts, tense and agreement morphology appears affixed to the verb root as in (1)–(6). With synthetic verbs in other aspectual contexts and open class “analytic” verbs in all finite contexts, tense and agreement morphemes do not appear affixed to the verb root, but rather on a separate auxiliary verb (Azkue 1923; Altube 1929; Arejita 1980; Euskaltzaindia 1977, 1985; Hualde & Ortiz de Urbina 2003). The inventory of verbs that behave synthetically varies to some degree from dialect to dialect. ­Euskaltzaindia (1977) and de Rijk (2008) cite 25 synthetic verbs with relatively regular usage, both

 Arantzazu Elordieta & Bill Haddican

transitives and intransitives. We illustrate the different behavior of synthetic and analytic verbs in these two aspectual environments in (7) and (8) respectively. (7) Synthetic verbs a. Jon dator Jon come.3sg ‘Jon is coming’ b. Jon etorr-i da Jon come-prf aux.3sg ‘Jon has come’

[imperfective]

[perfective]

(8) Analytic verbs a. Jon bazkal-tzen ari da Jon lunch-imprf prog aux.3sg ‘Jon is eating lunch’ b. Jon-ek bazkal-du du Jon-erg lunch-prf aux.3sg ‘Jon has eaten lunch’

[imperfective]

[perfective]

The relevance of these facts for verb focus constructions is that verb doubling is only available in synthetic contexts. As shown in (9), analytic verbs never double. (9) *Bazkal-du bazkal-tzen ari da    lunch-inf lunch-imprf prog aux ‘Jon is eating lunch’

Similarly, synthetic class verbs do not double when they behave analytically, that is, in non-imperfective contexts:3 (10) *Ibil-i ibil-i da    walk-inf walk-prf aux ‘She has WALKED’

A second restriction on verb doubling is that it occurs only in contexts where polarity or the verb itself is contrastively focused, as reflected in the translations in (1)–(6). Verb-doubling is blocked in contexts in which additional VP material is focused along with the verb (Elordieta and Jouitteau 2010; Elordieta, 2009, 2010). (11) *[Kalean ibil-i] dabil     street.loc walk-inf walk.3sg ‘S/he is walking in the street’

.  This sets aside instances where the first verb is topicalized and separated from the lower copy by a pause. This is discussed in Section 3 below.



Strategies of verb and verb phrase focus across Basque dialects 

(12) *[Ingeles eta frantses jakin] daki    English and French know-inf know.3sg ‘S/he knows english and french’

A third restriction on doubling concerns word order. A well-known property of foci in Basque is that they must appear left-adjacent to the tense-bearing verb in synthetic constructions or left-adjacent to the verb cluster in analytic constructions (Altube 1929; Euskaltzaindia 1985; Ortiz de Urbina 1989; Uriagereka 1999; Elordieta 2001; Irurtzun, this volume). We illustrate this constraint for synthetic and analytic verbs in (13) and (14) respectively. (13) Synthetic verbs a. Bost-etan JON dator five-at Jon come.3sg ‘JON is coming at five’ b. *JON bost-etan dator  Jon five-at come.3sg ‘JON is coming at five’ (14) Analytic verbs a. Miren JON-EK ikus-i du Miren Jon-erg see-prf aux ‘JON has seen Mary’ b. *JON-EK Miren ikus-i du  Jon-erg Miren see-prf aux ‘JON has seen Mary’

The only elements that can intervene between the verb cluster/tense bearing verb and the focused element are the negative morpheme ez (15), and a class of evidential and speech act particles including ei ‘allegedly’, as in (16). (15) JON ez dator Jon neg come.3sg ‘JON isn’t coming’ (16) JON ei dator Jon evid come.3sg ‘JON is allegedly coming’

Importantly, in verb doubling contexts, no element can intervene between the copies, as shown in (17). (17) *Jakin ez/ei daki-zu zuk hori    Know.inf neg/evid know.2sg you that ‘You don’t/ allegedly know that’

 Arantzazu Elordieta & Bill Haddican

These facts are plausibly related to the further fact that doubled verbs differ from argument and adjunct foci in Basque in being clause-bound. Another welldescribed property of argument/adjunct foci in Basque is that they can raise out of embedded clauses (Ortiz de Urbina 1989). (18) Nor esan duzu [nor dabil-ela]? Who say aux walk.3sg-c ‘Who did you say is walking?’ (19) JON esan dute [JON dabil-ela] Jon say aux walk.3sg-c ‘They said JON is walking’

The verb doubling strategy is restricted to main clause contexts for some speakers (G. Aurrekoetxea, p.c.). However, other speakers also use it in embedded contexts (Arejita 1984; Osa 1990), as illustrated in (20). Importantly, as noted by Elordieta (2010: 44), for such speakers, a “doubled” verb copy cannot raise out of its firstmerged clause in this way (21). In this respect, it differs from V and VP focus in analytic contexts, to be described in Section 4, which can move cyclically across finite clause boundaries. (20) a. [Ibil-i dabil-tza-n] txartelak    walk-inf walk-3pl-c credit.cards ‘Credit cards that WORK’

(Bank advertisement)

b. Eztakitt [joan sixoian-ø ala neg know.1sg   go.inf go.3sg past3sg-c or etorr-i etorren-ø] come.inf come.3sgpast-c

‘I don’t know whether he was coming or going’

(Osa 1990)

(21) *Etorri esan dute [etorri datorr-ela]    come say aux come.3sg-c ‘They said she is COMING’

Verb doubling constructions akin to the Basque strategy discussed so far have been described in a now considerable body of literature on languages including Nupe (Kandybowicz 2007), Kwa (Aboh 2007), Russian (Abels 2001), European Portuguese (Martins 2007), Haitian (Harbour 2008; Koopman 1984; Manfredi 1993), Korean (Jo 2003), Hebrew (Landau 2006, 2007) and Breton (Jouitteau 2008, 2011). All such constructions involve doubling in the context of some topic or focus interpretation. This literature generally distinguishes two kinds of verb doubling constructions. In some languages, the higher copy appears to raise as an XP, in e.g. VP/predicate fronting in Russian (Abels 2001), Hebrew (Landau 2006, 2007) and Haitian Kreyol (Manfredi 1993; Harbour, 2008). A second scenario



Strategies of verb and verb phrase focus across Basque dialects 

described for European Portuguese by Martins (2007) and Nupe (Kandybowicz 2007) is where the higher copy of the verb seems to have raised from the position of the lower copy by head movement.4 In the case of Basque, the facts that doubling constructions can never focus a VP and that the movement is clause-bound and can skip no intervening heads suggests that Basque belongs to the latter class, that is, that the verb raises as a head. In particular, we assume that the higher copy of the verb raises to the head of the same focus projection targeted by argument/adjunct foci. One piece of evidence to this effect is that argument/adjunct foci can never co-occur as shown in (22), arguably due to a ban on multiple foci. (22) *Nor-k jakin daki egia?    who-erg know-inf know.3sg truth ‘Who KNOWS the truth?’ [as opposed to‘think’ or ‘believe it’]

We assume the functional sequence shown in (23), adapted from Laka (1990) and Ortiz de Urbina (1994, 1999) and assumed in much subsequent literature (­Elordieta 2001; Elordieta and Haddican 2014; Etxepare and Uribe-Etxebarria 2009). Here, Evid and Σ stand for the first-merged positions of evidential particle and polarity morphemes respectively. (23) [Top*[Focus [Σ [ Evid [T …]]]]

Specifically, we propose that the higher copy raises from T to Foc, as in (24), which corresponds to the example in (1). As can be seen in the example, the subject is topicalized:5 (24) [TopP Nik Top [FocusP jakin-Focus [TP dakit …egia ]]]]

From the perspective of this proposal, a question that arises is why the tense-­ bearing verb cannot raise to Foc without need for doubling. Such a derivation, presumably, would lead to the ill-formed example in (6), repeated here.

.  Another possibility is Elordieta & Jouitteau’s (2010) and Jouitteau’s (2011) analysis of doubling in Breton as postsyntactic excorporation. .  Linearization of syntactic objects in Basque is of the complement-head type TP-internally, but shows head-complement order regarding most left peripheral heads, including Focus (cf. Ortiz de Urbina 1989, 1999; Laka 1990; Elordieta 2001). For our purposes, both analyses assuming either a head final order or a universal spec-head-complement order can capture the facts discussed here, the latter analysis requiring additional movement operations. We take no position on this issue here. For expository convenience, phrases are displayed in the headcomplement order in the structures below.

 Arantzazu Elordieta & Bill Haddican

(6) *Ni-k, daki-t egia    I-erg know.1sg-1sg truth ‘As for me, I KNOW the truth’

(as opposed to ‘think’ or ‘believe it’)

A further question is why the higher copy appears with infinitival morphology and not in its finite form. In fact, the reverse situation, in which the higher copy is finite and the doubled lower copy is not, is not possible (25a), nor is it possible doubling with two tensed verbs (25b): (25) a. *Dakit jakin egia    know.3sg-1sg know-inf truth b. *Dakit dakit egia    know.3sg-1sg know.3sg-1sg truth ‘I KNOW the truth’

These issues are plausibly related to the status of the lower copy as a tense-bearing verb. Given that V-doubling may apply in all synthetic (V–T raising) contexts and never in analytic contexts where the verb root may not raise to T, V–T movement appears to be crucially implicated in doubling. Specifically, we propose that the inability of the tense-bearing verb to move to Focus is related to a wellknown ban on tense-bearing verbs in sentence initial position, discounting topics (Altube 1929; Elordieta and Jouitteau 2010; Etxepare and Ortiz de Urbina 2003; ­Euskaltzaindia 1985; Ortiz de Urbina 1989, 1994; Uriagereka 1999). We illustrate this constraint, which we will refer to as “*T1” in (26). In (26a), the tense-bearing verb sits in sentence initial position and the result is bad. (26b,c) show word orders where the verb is shielded from the left edge of the sentence by negation and foci, which are fine, or by a non-finite verb form, as in (26d). (26e) shows that ­topics, which are obligatorily separated from material to their right by an intonational break, do not count as first position elements. (26) a. *Dator Mikel  come.3sg Mikel ‘Mikel is coming’ b. Ez dator Mikel neg come.3sg mikel ‘Mikel is not coming’ c. MIKEL/Nor dator/? Mikel/who come.3sg ‘MIKEL/who is coming/?’ d. Etorri da Mikel come-inf be.3sg Mikel ‘Mikel HAS come’



Strategies of verb and verb phrase focus across Basque dialects 

e. *Mikel, dator/?  Mikel come.3sg ‘As for Mikel, he is coming’

*T1 is violable, however, in some embedded clause types, as in the temporal adjunct clause and the embedded yes/no questions in (27) and (28) respectively (Ortiz de Urbina 1994; Elordieta and Haddican 2014). (27) Jon ikusi-ko dut [datorr-en-ean/%ba-datorr-en-ean] Jon-abs see-fut aux   come.3sg-c-loc/ba-come.3sg-c-loc ‘I will see Jon when he comes’ (28) Ez daki-t [datorr-en/%ba-datorr-en] ala ez neg know-1sg   come.3sg-c/ba-come.3sg-c or not ‘I don’t know if (s)he’s coming or not’

It has been suggested that the above pattern is akin to V2, in light of the fact that V2 effects in Germanic languages other than English do not apply in some kinds of embeddings (Ortiz de Urbina 1989, 1994; Uriagereka 1999). According to Ortiz de Urbina’s (1989) influential approach to I-to-C in Basque, Basque instantiates the standard analysis of V2 in Germanic, which models V2 as a conspiracy of an EPP feature on a C-field head and a verb-raising feature on this same head (Chomsky 2000; Roberts 2004; Julien 2009; Jouitteau 2008, 2010; Holmberg 2015; Leu 2015). On such an approach, the contrast in (26a) vs. (27)/(28) might be taken to reflect the need for EPP-driven XP movement to C in root clauses but not in embedded clauses. Elordieta and Haddican (2014, in preparation) nevertheless describe several sets of word order facts that are problematic for this approach and argue instead that the tensed verb does not raise to the same C-head that attracts XPmovement (see also Uriagereka 1999 for a similar proposal). To mention some, the linearization of the [verb+aspect] complex with respect to focalized constituents raises some problems. We saw in (26d) that the main verb in analytic contexts can be a first position element. Under the V2 approach, this means that it raises as an XP. Nevertheless, when a constituent is focalized the non-finite verb appears between the focus and the tensed auxiliary, resulting in the order XPfoc-V-Aux: (29) Liburua eros-i du Jon-ek book-det buy-prf aux Jon-erg ‘Jon has bought THE BOOK’ (focus on the object)

If we want to maintain that the finite verb in second position is in a spec-head configuration with the focused XP, as on the standard V2 approach, the main verb should not be able to intervene between them. It seems thus that in such contexts the lexical verb instead is head-adjoined to the finite auxiliary. In fact, in Ortiz de

 Arantzazu Elordieta & Bill Haddican

Urbina’s (1989, 1994), and Elordieta’s (2001) analysis, the [verb+aspect] complex head-adjoins to T to satisfy a lexicalization requirement on the morphologically weak auxiliary. The [Verb+Aux] then head-adjoins to a C-related head when there is movement to the left periphery, namely in wh-questions and focalizations. But then, this analysis faces the non-trivial obstacle of explaining why the verb can behave like a first-position XP in some contexts but a head in “C” in others. We adopt from Ortiz de Urbina’s proposal the idea that *T1 is morphophonological in nature. A well-known property of finite verbs in Basque is that they are never stress-prominent, that is, they can never bear main stress (Ortiz de Urbina 1989, 1994; Elordieta 2001). This restriction is certainly due some explanation, which we are not able to provide here. Rather, we simply observe this as a property of finite verbs in Basque, given in (30). (30) Finite verbs in Basque are not main stress-bearing

Also crucial to the analysis is the (more or less standard) assumption that Focus must contain an element bearing main stress:6 (31) In Basque, Focus must contain an element bearing main stress

In the well-formed example in (26c), the foci will occupy FocusP, and in (26b) the negative morpheme ez, which bears main intonational prominence in the sentence, will occupy FocusP, having raised from ΣP. The ill-formed examples in (26a,e), however, cannot be parsed with a stress-bearing element in FocusP, since, in both cases, the only possible element that could occupy this position is the prosodically weak finite verb. For the same reason, in the case of verb doubling, what blocks T-Focus movement in sentences like (25a-b) is the absence of a stressbearing item in FocusP. In such cases, the constraints in (30)–(31) are therefore in conflict with the need for the element interpreted as the focus – the verb – to raise to Focus and bear stress prominence. Our proposal for verb doubling in Basque is that it reflects a PF repair by partial spell out of multiple and/or lower copies in a way that accommodates these conflicting needs (Nunes 2004; Bošković and Nunes 2007). Specifically, we take the verb doubling construction to involve spell out of two copies of the verb root – one in T and one in Focus. On the higher copy, tense and agreement morphology will be deleted in compliance with (30) (see Landau 2006 for a similar analysis of doubling in Hebrew). We assume that the morphology associated with infinitives – the verb marker and the infinitival marker – is inserted at vocabulary

.  See Elordieta and Haddican (2014, in preparation) for a discussion of the difference between root and non-root contexts in terms of *T1.



Strategies of verb and verb phrase focus across Basque dialects 

insertion after PF deletion of the tense/agreement morphology. We illustrate this proposal in (32), which represents the two verb copies in (1). Here, for expositional convenience, we use “φ” as shorthand for person/number agreement features on the verb. (32) a. Spell out structure (following T-Foc movement) [FocusP [[[know[FOC]]-Tφ]-Focus [TP [[[know[FOC]]-Tφ] …]]]]

b. Deletion of the higher copy of T (pursuant to (30)) [FocusP [[[know[FOC]]-Tφ]-Focus [TP [[[know[FOC]]-Tφ] …]]]] c. Vocabulary insertion [FocusP jakin [TP dakit …]]

We consider additional evidence implicating *T1 in this derivation in the following section. 3.  The ba-insertion strategy As described in the previous section, verb doubling constructions are restricted to dialects on the western edge of the Basque dialect area. In this section, we discuss an alternative verb focus strategy in these same dialects and in other dialects where doubling is disallowed, which involves inserting a morpheme ba- to the left of the finite verb as in (33).7 (33) Ba-daki-t. ba-know.3sg-1sg ‘I KNOW (not just think)’/‘I DO know’

In all dialects, it is possible to combine doubling with ba-, as in (34), but in such sentences the verb is necessarily interpreted as a topic rather than as focus. (34) Etorri, ba-dato-z come-inf ba-come-3pl ‘As for coming, they are indeed coming’ .  This morpheme is distinct from the homophonous morpheme corresponding to English complementizer ‘if ’. Some grammarians hypothesize that the ba- morpheme may be historically related to the affirmative particle bai, ‘yes’ (see Altube 1929; Arejita 1980; Laka 1990; Osa 1990), but they do not seem to pattern together in its extended usage in non-standard speech. As opposed to bai, which may occur in informal non-standard speech as a contrastive assertive particle preceding a finite verb, ba- only appears with synthetic verbs: (i) baidot / *badot ikusi zure ama. yes-aux    ba-aux see your mum ‘I have seen your mum.’

 Arantzazu Elordieta & Bill Haddican

Importantly, restrictions on ba-insertion are similar to those described above for verb-doubling, particularly in central and western dialects. First, in central and western dialects, the ba- strategy is restricted to synthetic verbs, like doubling constructions. In some eastern dialects of Basque, however, ba- also occurs with the finite auxiliary of analytic constructions: (35) Ba-dut bazkal-du Ba-aux lunch-prf ‘I did have lunch/I had lunch’ (eastern dialects)

Second, in central and western dialects, ba-insertion constructions are limited to polarity focus interpretations and also allows for contrastive verb focus interpretations. In these dialects, they are not possible with contrastive or new information focus on the VP, or with new information focus on the verb. (36), for example, is not a felicitous answer to the Basque counterpart of What are they doing? In some eastern dialects, however, the ba- construction is indeed possible with new information on the verb phrase. The placement of ba- left-adjacent to the tense-bearing verb suggests the possibility that ba- sits in the same position occupied by the verb double in verb doubling contexts. A difference between the two constructions, from this perspective, however, is that ba-, unlike the verb double, need not be left-adjacent to the tense bearing verb. It can be separated from the verb by omen ‘allegedly’ and its western dialectal variant, ei (cf. (16) above and (36)). (36) Ba omen/ei dato-z gu-regana ba allegedly come-3pl we-to ‘They are coming towards us’/‘They are coming towards us’

Like verb copies in the doubling construction, ba- is clause bound, i.e. cannot extract to a higher clause unlike argument/adjunct foci. (37) *Ba esan dute [ba datorr-ela]    ba say aux comes-c ‘They said she is COMING’

We take the appearance of ba- in verb focus sentences like (36) to be a special case of a more general repair strategy involving the morpheme ba- that Ortiz de Urbina (1994) calls “ba-support”, that is, insertion of an expletive ba- morpheme. Ortiz de Urbina’s (1994) analysis of ba-, which we adopt, is that ba- is inserted at PF to rescue configurations that would otherwise violate the *T1 restriction discussed above. In addition, as Ortiz de Urbina (1994) notes, ba- insertion is a last resort strategy -it cannot apply in contexts where it does not repair *T1. We illustrate this in (38), which shows the distribution of ba- in the contexts introduced in (26). (38a,d) shows that ba- repairs *T1 violations, where the finite verb abuts the



Strategies of verb and verb phrase focus across Basque dialects 

left edge of the clause or is preceded only by a topic. (38b,c) show that ba- is not possible where it is not needed, that is, in the presence of negation or foci, which separate the verb from the left edge of the sentence. (38) a. *(Ba)-dator Mikel    ba-come.3sg Mikel ‘Mikel is coming’ b. Ez (*ba-)dator Mikel neg   ba-come.3sg Mikel ‘Mikel is not coming’ c. MIKEL/Nor (*ba-)dator/? Mikel/who    ba-come.3sg ‘MIKEL/who is coming/?’ d. Mikel, *(ba-)dator/? Mikel   ba-come.3sg ‘As for Mikel, he is coming’

In comparing the ba- strategy to the verb doubling strategy (and two other strategies to be discussed shortly), we will focus on central and western dialects and set aside eastern dialects, where, as we have noted, the ba- strategy has a different set of interpretations and applies in analytic contexts as well. For central and western dialects, where the ba- strategy is similar in distribution to verb doubling, we propose that the ba- construction differs from verb doubling constructions in that the entire higher copy of the verb, in Foc, deletes, rather than just the *T1-offending morphology. Ba-support then applies in the usual way to fix the *T1 violation. We illustrate this proposal in (39) which corresponds to (33). (39) a. Spell out structure (following T-Foc movement)8 [FocusP [[[know[foc]]-Tφ]-Focus [TP [[[know[foc]]-Tφ] …]]]]

b. Deletion of the higher copy of V (pursuant to (30)) [FocusP [[[know[foc]]-Tφ]-Focus [TP [[[know[foc]]-Tφ] …]]]] c. Ba-support and vocabulary insertion [FocusP ba- [TP dakit …]]

On this approach, something more will be required to explain a difference we have noted between the ba-insertion strategy and the verb doubling strategy, namely that, in the doubling construction, the higher copy of the verb must be strictly

.  An anonymous reviewer wonders about competition between the V-doubling and bainsertion strategies; that is whether the fact that both are possible among the same speakers means that they are equally economical and that one therefore doesn’t block the other on economy grounds. We set this important issue aside for space reasons.

 Arantzazu Elordieta & Bill Haddican

adjacent to the verb unlike ba-. Specifically, if we take T-Foc movement to happen in the syntax, and if we take the head movement constraint to be a narrow syntactic phenomenon, then T-Foc movement should be blocked by an intervening evidential particle, assuming it is a head (see Ortiz de Urbina 1989, 1994; Ortiz de Urbina 2003, and most literature following).9 This blocking appears to apply in the verb-doubling construction as illustrated in (17), but not in the ba- construction in (36). We speculate that head movement constraint violations are evaluated at PF after chain reduction has applied, so that deletion of the high copy repairs the HMC violation in ba-support, but not in verb doubling where part of the high copy is pronounced. 4.  The V1 strategy The two strategies discussed so far apply only in synthetic verb constructions in the central and western dialects that we have focused on. A prediction of the foregoing analysis is that, in analytic contexts, where tense and agreement morphology is not realized on the verb root-containing cluster, but rather on an auxiliary, focalized verbs should be able to raise to focus. In this section, we propose that a construction we call the V1 construction is indeed evidence to this effect and aim to reconcile it with the syntax proposed for verb doubling and the ba- strategy just discussed. We illustrate this construction in (40), from Ortiz de Urbina (1994). Here the [verb root+aspect] cluster bears stress prominence and appears clause initially (excepting topics), with arguments following. (40) Eros-i du Jon-ek egunkari-a buy-prf aux Jon-erg newspaper-det ‘Jon HAS bought the newspaper’ (general verum focus reading across dialects) /‘Jon has BOUGHT the newspaper’ (also verb focus in eastern dialects)  (adapted from Ortiz de Urbina, 1994)

Ortiz de Urbina (1994) gives the sentence in (40) as an example of polarity focus, and across dialects this construction can have this interpretation. In addition, in eastern dialects this construction is also available for contrastive and new information focus on the verb.10 .  See Etxepare (2010, this volume) for an analysis of the different uses and nature of the evidential particle omen ‘allegedly’ in Basque. .  This fact may be related to the fact that eastern dialects lack a fourth construction to be discussed in the next section -the dummy egin strategy- which marks contrastive and new information focus on verbs in analytic contexts in central and western dialects.



Strategies of verb and verb phrase focus across Basque dialects 

The fact that the V1 construction never focuses phrasal constituents but always either the verb or polarity suggests that the constituent that raises to Focus is a head rather than an XP, as suggested by Ortiz de Urbina (1994). On the other hand, if this movement is head movement all the way to Foc, then the [verb+aspect] cluster will have needed to raise past the auxiliary, apparently skipping intervening heads. In modeling these facts, we propose to take full advantage of the scattered deletion approach embraced in the previous two sections. In particular, we propose that the [verb+aspect] cluster raises all the way to Focus, by successive head adjunction, skipping no intervening heads including T. Through this movement, the focused verb can be in FocusP at PF as required, apparently, of all foci in Basque. The ban on finite verbs in Focus, however, will necessitate a repair, and we propose that this repair involves deleting the tense and agreement morphology from the higher copy, with the consequence that the auxiliary will spell out in a lower position, in T. The [verb root+aspect] cluster spells out in the higher position in the usual way. (41) illustrates this proposal for the higher ­portion of (40). (41) a. Spell out structure (following V-Asp-T-Foc movement)11 [FocusP [[[buy[foc]]-Asp]-Tφ]-Focus [TP [[[buy[foc]]-Asp]-Tφ] …]]]]

b. Deletion of the higher copy of T (pursuant to (30)) [FocusP [[[buy[foc]]-Asp]-Tφ]-Focus [TP [[[buy[foc]]-Asp]-Tφ] …]]]] c. Vocabulary insertion [FocusP erosi [TP du …]]

To summarize, we have proposed that three strategies for verb focus in Central and Western Basque dialects – verb doubling, ba-support and V1 orders – ­differ in two main ways: (i) the way syntax feeds chain reduction; and (ii) the way that chain reduction/copy deletion accommodates prosodic requirements of foci in Basque. We have argued that chain reduction/copy deletion is sensitive to a requirement that FocusP contain intonationally prominent material.

.  A reviewer notes that an important challenge for this proposal is to explain what prevents ba-support from applying in analytic contexts, as in (i). From the perspective of Ortiz de ­Urbina’s last resort approach to ba-, it appears that ba-insertion is not blocked by a competing verb-doubling derivation, but is blocked by the verb raising derivation in (41). (i) [FOC ba [TP erosi du ] We do not attempt an answer to this issue here. Plausibly related to this issue is what blocks verb doubling in analytic contexts. We thank a reviewer for a helpful discussion of these facts.

 Arantzazu Elordieta & Bill Haddican

5.  The dummy egin strategy A final member of our menagerie of Basque verb focus constructions is one in Central and Western Basque dialects where, with all verbs (synthetic and analytic) in non-imperfective contexts and with the open class of analytic verbs independently of their aspect, a focus interpretation on V co-occurs with a dummy or light verb egin ‘do’. We illustrate this in (42), a felicitous answer to ‘What has Miren done with her friends?’ (42) Miren-ek bazkal-du egi-n du lagun-ekin Miren-erg eat.lunch-inf do-prf aux friend-with ‘Miren has EATEN LUNCH with her friends’

Such sentences with egin necessarily have a verb focus interpretation. Egin cannot appear in out-of-the-blue, wide focus contexts like (43), as an answer to ‘What happened?’ but rather co-occurs with narrow contrastive or new information focus strictly on the verb (rather than VP) in most dialects. (43) a. #Miren-ek bazkal-du egin du lagun-ekin    Miren-erg have.lunch-inf do-prf aux friend-with b. Miren-ek lagun-ekin bazkal-du du Miren-erg friend-with have.lunch-inf aux ‘Miren has eaten lunch with her friends’

Rebuschi (1983) and Haddican (2005, 2007) argue that in sentences such as (42), the verb raises to the same left-peripheral focus position targeted by argument and adjunct foci as in (44). As Haddican (2005, 2007) discusses in detail, the focused V in (42) behaves like other kinds of left-peripheral foci in Basque in terms of word order, scope and intonation. Indeed, as regular foci in (44), the focused verb in (42) bears intonational prominence and immediately precedes the verbal complex bearing aspectual, tense and agreement-features: (44) a. Nork eros-i ditu hainbeste liburu? who-erg buy-prf aux so many book ‘Who has bought so many books?’ b. Miren-ek eros-i ditu hainbeste liburu Miren-erg buy-prf aux so many book ‘MIREN has bought so many books’

Although we have described the egin structure as a verb focus construction, ­Haddican (2005, 2007), based on data mainly from the Oiartzun dialect, argues that it may also serve as VP focus, in contexts of remnant topicalization, and hence, like other left peripheral foci, the non-finite verb in sentences like (42) raises to the



Strategies of verb and verb phrase focus across Basque dialects 

focus projection as an XP rather than as a head. The facts are clearest in dialects like Oiartzun’s that allow for remnant topicalization (Ortiz de Urbina, 2002). In the representation in (45), for example, the non-focalized portion of the sentence Mirenek egin ditu … raises as a remnant topic leaving the focused VP right-peripherally. (45) a. Mirenek egin ditu(,) den-denak jan Miren do-prf aux all-all eat b. [Top Mirenek egin ditu [Focden-denak jan] [Mirenek den-denak    Miren do-prf aux   all-all eat  jan egin ditu]] ‘Miren has EATEN THEM ALL’

It seems, though, that the operation of remnant topicalization must apply in order to obtain a VP-focus interpretation with dummy egin. Without topicalization we only obtain a focus reading on the verb itself, excluding the object from being interpreted as focus.12 (46) Mirenek den-denak jan egin ditu Miren all-all eat do-prf aux ‘Miren has EATEN them all’/‘*Miren has EATEN THEM ALL’

In addition, as we noted earlier, all dialects with the egin focus construction allow for verbal foci, like other kinds of foci, to extract to a higher clause as in (47). This is straightforwardly predicted if the verb raises as an XP to Spec,FocusP, but mysterious if the verb raises as a head, a movement which is more restricted locally: (47) Erosi esan didate [erosi egi-n zenue-la etxe-a] buy say aux do-prf aux-c house-det ‘They have told me that you BOUGHT the house’ (as opposed to, say, rent it)

Importantly, in constructions with egin, the dummy verb bears the aspectual morphemes that normally appear on the main verb and the main form appears in its infinitival citation. In (48a), for example, the imperfective morpheme -ten appears affixed on egin, rather than on the main verb sinetsi, ‘believe’. Likewise, (48b) ­illustrates that egin is necessary to hold the perfective aspect morpheme of the main verb which has undergone long focus movement. .  This may be related to the fact that Basque has a more ‘economic’ strategy to mark ­information focus on VP, namely leaving the focus in-situ, without any movement involved. As argued in Elordieta (2001) and Irurtzun (2007), information focus in Basque applies in-situ as a result of the NSR (Cinque 1993; Reinhart 1995; Reinhart & Neeleman 1998), and by ­occurring left-adjacent to the finite verb. The idea behind these proposals is that by being SOV, either O, [OV] or [SOV], the whole sentence, may be interpreted as information focus. Thus, there would be no need to insert egin in (46) in the absence of any movement.

 Arantzazu Elordieta & Bill Haddican

(48) a. Hori sinets-i egi-ten dugu that believe-inf do-imprf aux ‘We BELIEVE that’ b. Erosi esan didate [erosi *egi-n zenue-la etxe-a] buy say aux    do-prf aux-c house-det

Based on this last fact, Haddican (2005, 2007) and Elordieta (2009, 2010) propose that egin is merged to provide lexical support for the aspectual morphemes only in environments where the main verb has more pressing commitments in the left periphery. Haddican (2007) argues it is an expletive element, whereas Elordieta (2010) treats it as a pseudocopy of the verb root, which bears aspectual features. (49) depicts a representation for dummy egin sentences in dialects where these involve head movement, i.e. with narrow focus on the verb rather than a VP. VP focus constructions in dialects like Oiartzun will involve VP movement to spec, FocusP. (49) a. Spell out structure (following V-to Foc movement) [FocusP [buy[foc]]-Focus [TP [[[buy[foc]]-Asp]-Tφ] …]]]]

b. Deletion of the lower copy of V [FocusP [buy[foc]]-Focus [TP [[[buy[foc]]-Asp]-Tφ] …]]]] c. Vocabulary insertion [FocusP erosi [TP egin du …]]

The question that arises with this strategy is similar to the one considered above in the case of T-Focus movement with synthetic verbs, namely why the aspect-­ bearing verb cannot raise as an XP to FocusP without need for egin in the derivation. We cannot offer a particularly explanatory answer to this question, nor do Haddican (2007) nor Elordieta (2010). Manfredi (1993), however, notes that, cross-linguistically, focalized verb phrases tend to bear nominalizing morphology, and proposes that this is a condition on VP focus constructions. We propose that the appearance of egin in western and central Basque is related to this requirement. That is, egin is inserted to host aspectual morphology, so that the focused verb in FocusP may appear with its infinitival morphology, which we take to contain a nominalizing morpheme (Haddican 2007). 6.  Conclusion This chapter has outlined four strategies for focusing verbs and VPs in Basque. The constructions differ in kinds of focus interpretations they may have, as well as the contexts in which they apply – that is, whether they appear with synthetic or analytic verb forms. We summarize the distribution of these four constructions in Table 1.



Strategies of verb and verb phrase focus across Basque dialects 

Table 1.  Focus properties of Verb and VP focus constructions across dialects (P = Polarity Focus, C = Contrastive Focus, N = New Information Focus % = only in some dialect) Dialects Verb class

Strategy

West

Cent.

East

SYNTHETIC

V-doubling (T in high copy deleted)

P,C

*

*

Ba- (deletion of high copy)

P,C

P,C

P,C,N%

V1 (scattered deletion)

P

P

P,C,N

Egin (V/VP movement to Focus)

C,N

C,N

*

ANALYTIC

We have outlined a partially unified approach to the first three of these constructions – the V doubling, ba-support and V1 constructions – which are similar in their ranges of interpretation, and in being head movement rather than XP movement constructions. We have proposed that these constructions differ in the way that chain reduction accommodates phonological needs of foci in Basque. We propose that the egin construction is of a different nature, as it may involve either V or VP focus, depending on the dialect; thus, in dialects like Oiartzun it involves XP movement of foci that can receive new information interpretations. Many questions remain, the most important among these concerning the source of the differences in interpretation among these constructions, particularly between the egin construction in central/western dialects and the other three, as well as the question regarding why all four strategies are not fully available across dialects. Future work might usefully address these issues, which have not so far received extensive formal analysis.

References Abels, Klaus. 2001. “The predicate cleft construction in Russian.” Formal Approaches to Slavic linguistics (FALS) 9: 1–18. Ann Arbor: Michigan Slavic Publications. Aboh, Enoch O. 2007. “Leftward focus versus rightward focus: the Kwa-Bantu conspiracy.” SOAS Working Papers in Linguistics 15: 81–104. Altube, Seber. 1929. Erderismos. Indautxu, Bilbo, 2nd edition. Arejita, Adolfo. 1980. “Aditzaren galdegai funtzioaz.” Euskera 25, 355–369. Arejita, Adolfo. 1984. Euskal Joskera. Bilbao: Labayru. Arejita, Adolfo. 1988. “Focalizaci6n del verbo en la oraci6n subordinada”, en Estudios de Lengua y Literatura, Universidad de Deusto, 39–54. Aurrekoetxea, Gotzon. 1995. Bizkaieraren egituraketa geolinguistikoa. Leioa: UPV-EHUko ­Argitarapen Zerbitzua/Serie Tesis doctorales UPV/EHU.

 Arantzazu Elordieta & Bill Haddican Azkue, Resurrección M. 1923. Morfología vasca (Gramática Básica Dialectal del Euskera). Tomo I: Los afijos vascos; tomo II: Categorías gramaticales. Zalla-Bilbo: La Gran Enciclopedia Vasca. Bošković, Zeljko and Nunes, Jairo. 2007. “The copy theory of movement: a view from PF.” In The Copy Theory of Movement, Norbert Corver and Jairo Nunes (eds.), 13–74. Amsterdam/ Philadelphia: John Benjamins.  doi: 10.1075/la.107.03bos Chomsky, Noam. 2000. “Minimalist inquiries: the framework.” In Step by step. Essays on minimalist syntax in honor of Howard Lasnik, David M.Roger Martin, Juan Uriagereka (eds.), 83–155. Cambridge, MA: MIT Press. Cinque, Giuglielmo. 1993. “A null theory of phrase and compound stress.” Linguistic Inquiry 24, 239–298. Elordieta, Arantzazu. 2001. Verb Movement and Constituent Permutation in Basque. HIL/Leiden University, LOT Dissertation Series, 47. Elordieta, Arantzazu. 2009. “Aldaera sintaktikoak aditzaren fokalizazioan.” Lapurdum 13: 99–112.  doi: 10.4000/lapurdum.2018 Elordieta, Arantzazu. 2010. “Aditzaren bikoizketa foku-topikoa denean.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai [Supplements of ASJU 52], Beatriz Fernández, Pablo Albizu and Ricardo Etxepare (eds.), 37–54. Bilbo: UPV/EHU. Elordieta, Arantzazu and Jouitteau, Mélanie. 2010. “Verb-doubling in Basque and Breton”, talk delivered at the 4th European Dialect Syntax Meeting, Donostia. Elordieta, Arantzazu and Haddican, William. 2014. “Truncation feeds intervention: Two clause type effects in Basque.” Presented at GLOW 37, Brussels, April 2–4, 2014. Elordieta, Arantzazu and Haddican, William (submitted for publication). Mapping clause type effects in Basque. Ms., UPV/EHU-CUNY-Queens College. Etxepare, Ricardo and Ortiz de Urbina, Jon. 2003. “Focalization.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 460–516. Berlin: Mouton de Gruyter. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2009. “Hitz hurrenkera eta birregituraketa euskaraz.” ASJU 43, 335–356. Etxepare, Ricardo. 2010. “Omen bariazioan.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai [Supplements of ASJU 52], Beatriz Fernández, Pablo Albizu and Ricardo Etxepare (eds.), 85–112, Bilbo: UPV/EHU. Euskaltzaindia, 1977. “Aditz sintetikoa.” Euskera XXII (2): 785–850. Euskaltzaindia, 1985. Euskal Gramatika. Lehen Urratsak-I. Bilbo: Euskaltzaindia. Gilisasti Iñaki, 2003. [1997]. Urduliz aldeko Berba lapikokoa. Bilbo: Uribe Kostako Mankomunitatea. Haddican, Bill. 2005. Aspects of Language Variation and Change in Contemporary Basque. Ph.D. dissertation, New York University. Haddican, Bill. 2007. “On egin: Do-support and VP focus in Central and Western Basque.” Natural Language and Linguistic Theory 25: 735–764.  doi: 10.1007/s11049-007-9027-8 Harbour, Daniel. 2008. “Klivaj predika, or predicate clefts in Haitian.” Lingua 118, 853–871.

doi: 10.1016/j.lingua.2007.11.010

Holmberg, Anders. 2015. “Verb second.” In Syntax – Theory and Analysis. An International Handbook, vol. 1, Tibor Kiss and Artemis Alexiadou (eds.), 342–382. Berlin: Mouton de Gruyter. Hualde, José Ignacio and Jon Ortiz de Urbina (eds.). 2003. A Grammar of Basque. Berlin-New York: Mouton de Gruyter. Irurtzun, Aritz. 2007. The Grammar of Focus at the Interfaces. Ph.D. dissertation, UPV/EHU. Irurtzun, Aritz. This volume. “Strategies for argument and adjunct focalization in Basque”.



Strategies of verb and verb phrase focus across Basque dialects 

Jo, Jung Min. 2003. “Variations in predicate cleft constructions in Korean: epiphenomena at the syntax-PF interface.” Presented at the 10th Harvard International Symposium on Korean Linguistics. Jouitteau, Mélanie. 2008. “The brythonic reconciliation.” Linguistic Variation Yearbook 7: 163–200.  doi: 10.1075/livy.7.06jou Jouitteau, Mélanie. 2010. “A typology of V2 with regard to second position phenomena: An introduction to the V1/V2 volume.” Lingua 120 (2): 197–209.  doi: 10.1016/j.lingua.2008.11.011 Jouitteau, Mélanie. 2011. “Post-syntactic excorporation in realizational morphology: Evidence from Breton.” In Formal Approaches to Celtic Linguistics, Andrew Carnie (ed.), 115–142. Cambridge: Cambridge Scholars Publishing. Julien, Marit. 2009. “Embedded V2 in Norwegian and Swedish.” Working Papers in Scandinavian Syntax 80: 103–161. Kandybowicz, Jason. 2007. “On fusion and multiple copy spell-out: the case of verbal repetition.” In The Copy Theory of Movement, Norbert Corver and Jairo Nunes (eds.), 119–150. Amsterdam/Philadelphia: John Benjamins.  doi: 10.1075/la.107.06kan Koopman, Hilda. 1984. The Syntax of Verbs: from Verb Movement Rules in the Kru Languages to Universal Grammar. Dordrecht: Foris. Laka, Itziar. 1990. Negation in Syntax. On the Nature of Functional Categories and Projections. Ph.D. dissertation, MIT. Landau, Idan. 2006. “Chain reduction in Hebrew V(P)-fronting.” Syntax 9: 32–66.

doi: 10.1111/j.1467-9612.2006.00084.x

Landau, Idan. 2007. “Constraints on partial VP-fronting.” Syntax 10 (2): 127–164.

doi: 10.1111/j.1467-9612.2007.00099.x

Leu, Tomas. 2015. “Generalized X-to-C in Germanic”. Studia Linguistica 69 (3): 272–303. doi: 10.1111/stul.12035 Manfredi, Victor. 1993. “Verb focus in the typology of Kwa/Kru and Haitian.” In Focus and Grammatical Relations in Creole Languages, Francis Byrne and Donald Winford (eds.), 3–51. Amsterdam/Philadelphia: John Benjamins.  doi: 10.1075/cll.12.05man Martins, Ana M. 2007. “Double realization of verbal copies in European Portuguese emphatic affirmation.” In The Copy Theory of Movement, Norbert Corver and Jairo Nunes (eds.), 77–118. Amsterdam/Philadelphia: John Benjamins.  doi: 10.1075/la.107.05mar Nunes, Jairo. 2004. Linearization of Chains and Sideward Movement. Cambridge, MA: MIT Press. Ortiz de Urbina, Jon. 1989. Some Parameters in the Grammar of Basque. Dordrecht: Foris. Ortiz de Urbina, Jon. 1994. “Verb-initial patterns in Basque and Breton.” Lingua 94, 125–153.  doi: 10.1016/0024-3841(94)90023-X Ortiz de Urbina, Jon. 2002. « Focus of correction and remnant movement in Basque.” In Erramu Boneta. Festschrift for Rudolf PG de Rijk, Xabier Artiagoitia, Patxi Goenaga and Joseba A. Lakarra (eds.), 511–524. Bilbo: UPV/EHU. Ortiz de Urbina, Jon. 2003. “Modal particles.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 316–322. Berlin: Mouton de Gruyter. doi: 10.1515/9783110895285 Osa, Eusebio. 1990. Euskararen hitzordena komunikazio zereginaren arauera. Ph.D. dissertation, University of the Basque Country (UPV/EHU).

 Arantzazu Elordieta & Bill Haddican Rebuschi, Georges. 1983. “A note on focalization in Basque.” ASJU 4: 29–42. [Reprinted in Essais de Linguistique Basque. [Supplements of ASJU 35], George Rebuschi (ed.) 1997, 31–41. Donostia: UPV/EHU and Gipuzkoako Foru Aldundia]. Reinhart, Tania. 1995. Interface Strategies. Utrecht: OTS Working Papers in Linguistics, Utrecht University. [Published in 1998, by MIT Press, Cambridge, MA.] Reinhart, Tania and Neeleman, Ad. 1998. “Scrambling and the PF Interface.” In The Projection of Arguments: Lexical and Compositional Factors, Wilhelm Geuder and Miriam Butt (eds.), Stanford: CSLI Publications. de Rijk, Rudolf P. 2008. Standard Basque: A Progressive Grammar. Cambridge: The MIT Press. Roberts, Ian. 2004. “The C-system in Brythonic Celtic languages, V2, and the EPP.” In The Cartography of Syntactic Structures, vol. 2, Luigi Rizzi (ed.), 297–327. New York/Oxford: Oxford University Press. Uriagereka, Juan. 1999. “Minimal restrictions on Basque movements.” Natural Language and Linguistic Theory 17, 403–444.  doi: 10.1023/A:1006146705483 Zuazo, Koldo. 1998. “Euskalkiak, gaur.” Fontes Linguae Vasconum: Studia et documenta 30: 191–234. Zuazo, Koldo. 2003. Euskalkiak: Herriaren Lekukoak. Donostia: Elkar. Zuazo, Koldo. 2008. Euskalkiak: Euskararen Dialektoak. Donostia: Elkar. Zuazo, Koldo. 2010. El euskera y sus dialectos. Donostia: Alberdania.

Strategies for argument and adjunct focalization in Basque Aritz Irurtzun CNRS – IKER, UMR5478

In this chapter, I provide an overview of the different strategies employed for Wh-questions and focalizations across Basque dialects. I argue that a core property of Basque syntax is the fact that both Wh- and focus phrases undergo syntactic (A′-type) displacements, and that they exhibit the main characteristics of syntactic displacements (locality, successive cyclicity, sensitivity to islands, etc.). After analysing the “standard strategy”, which is available across all dialects, I provide an overview of the new in situ Wh-question strategy of the young speakers of Navarro-Labourdin and two different strategies that are employed across different dialects to generate reinforced foci: the highly contrastive rightward focus constructions (specific to Southern dialects, and particularly common in High Navarrese), and the ‘reinforced movement’ strategy of NavarroLabourdin (a Northern variety). I finish with a brief description of some other constructions involving foci: the mirative focus constructions of substandard Basque, and the dialectal distribution of different types of split interrogatives. Keywords:  focus; Wh-movement; left periphery; wh in situ

1.  Introduction: The ‘Standard Strategy’1 Basque has been characterized as a S-IO-DO-V language given that, even if the order of constituents in this language is not a fixed one, this is the pattern appearing in an out-of-the-blue or all-new statement (cf., i.a., Ortiz de Urbina,

.  My deepest thanks to B. Fernández and J. Ortiz de Urbina for their editorial support and to two anonymous reviewers for their insightful comments. This work benefited from the projects IT769-13 (Eusko Jaurlaritza), FFI2013-43823-P, FFI2013-41509-P and FFI201453675-P (MINECO). The research leading to these results has also received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement n0 613465.

doi 10.1075/lfab.13.09iru © 2016 John Benjamins Publishing Company

 Aritz Irurtzun

1989 et seq.; Elordieta, 2001; Etxepare and Ortiz de Urbina, 2003; Irurtzun, 2007; ­Erdozia et  al., 2009). However, information-structure affects the order of constituents in Basque and it is, therefore, a ‘discourse-configurational’ language (in the sense of Kiss 1995). In this chapter, I provide an overview of the main syntactic and semantic properties of different question and focalization strategies in Basque. The discussion will concentrate on Wh-questions and argument and adjunct focalization – what is known as “term-focus”–, see Elordieta and ­Haddican (this volume) for an analysis of verb-focalization and Elordieta and Irurtzun (2010) for a study involving verum focus. To begin with, there is in Basque a common Wh-movement and focalization strategy that we could term the ‘standard strategy’, for it is available across all dialects of this language. This strategy is exemplified in (1b) for a Wh-question and (1c) for a subject-focalization, where we can observe that the basic S-O-V word order of an out-of-the-blue statement (1a) is altered and the verb appears immediately following the Wh- (1b) or focal (1c) phrase: (1) a. Jonek ura edan du Jon water drink aux ‘Jon drank water’ b. Nork edan du ura? who drink aux water ‘Who drank water?’ c. [Jonek]F edan du ura    Jon drink aux water ‘[Jon]F drank water’

Actually, a long-standing observation in Basque linguistics (cf. i.a. Altube 1920; de Rijk 1969; Mitxelena 1981; Eguzkitza 1987; Ortiz de Urbina 1989; Rebuschi 1997; Artiagoitia 2000) is that if this change in word order does not take place, the result is ungrammatical, as illustrated in the examples in (2a,b): (2) a. *Nork ura edan du?    who water drink aux ‘Who drank water?’ b. *[Jonek]F ura edan du      Jon water drink aux ‘[Jon]F drank water’

Authors like Ortiz de Urbina (1989, et seq.) or Irurtzun (2007) have analysed these constructions as involving an A′-displacement of the relevant phrase (a Wh-phrase or a focal phrase) to the specifier of some functional projection at



Strategies for argument and adjunct focalization in Basque 

the left periphery of the clause (traditionally, taken to be Cº or Focº).2 This displacement in turn would be followed by a T-to-C movement, as represented in (3), which, in order to avoid repeating the same tree twice, combines the structure under the subject-question in (1b) and the subject-focalization in (1c) (for simplicity, I am omitting the movement of the subject DP from Spec-vP to Spec-TP, and the movement of V-to-v-to-T):3 (3)

CP

Nork/[Jonek]F

C′

edan du

TP tDP

T′ vP

tDP

t[V+v+T] v′ t[V+v]

VP ura

tV

1.1  Cyclicity Furthermore, we can observe the cyclic nature of the Wh- or focus movement in embedded clauses in that the extraction of (for instance) the subject of an embedded clause is always accompanied by the movement of the verb, both in matrix and embedded clauses. If the movement of the embedded verb does not take place, the result is ungrammatical. Departing from the neutral (4a), the necessity of verbmovement can be observed in the contrast between sentences (4b) and (4c), for Wh-extraction, and between (4d) and (4e) for focalization:4 .  See also Elordieta (2001) and Arregi (2002) for alternative proposals and Artiagoitia (2000) and Etxepare and Ortiz de Urbina (2003) for a general overview of the syntax of these constructions. .  See also Laka and Uriagereka (1987) and Uriagereka (1999) among others. .  A reviewer wonders whether all speakers find these examples ungrammatical. The literature converges in this direction and the speakers I have consulted tend to associate the word

 Aritz Irurtzun

(4) a. Jonek [Mirenek ura edan duela] esan du Jon    Miren water drink aux.c say aux ‘Jon said that Miren drank water’ b. Nork esan du Jonek [t edan duela ura]? who say aux Jon drink aux.c water ‘Who did Jon say that drank water?’ c. *Nork esan du Jonek [t ura edan duela]?    who say aux Jon water drink aux.c ‘Who did Jon say that drank water?’ d. [Mirenek]F esan du Jonek [t edan duela ura]    who say aux Jon drink aux.c water ‘Jon said that [Miren]F drank water’ e. *[Mirenek]F esan du Jonek [t ura edan duela]    Miren say aux Jon water drink aux.c ‘Jon said that [Miren]F drank water’

In both (4b) and (4c) we have a Wh-question over the subject of the embedded clause; the only difference between both clauses is that in (4b) we observe OV inversion in the embedded clause (along with the SV inversion of the matrix clause), but in (4c) we do not, resulting in ungrammaticality. Likewise for the focalization constructions in (4d) and (4e). The necessity of verb-movement in the embedded clause has been taken as an indication that the extracted element moves trough the specifier of the embedded CP in its way to the specifier of the matrix CP, this cyclic movement being the catalyst of the verb-movement in the embedded clause.5 Therefore, the derivation corresponding to (4b)-(4d) would be the one in (5):

order in (4c) not with the interpretation in (4c) but with a reading where extraction does not take place from the embedded clause but from the matrix clause (‘Who said that John drank water?’). Unfortunately, we still lack any formal experimental test on acceptability judgements for this type of data. .  See Ortiz de Urbina (1995) for an analysis.



Strategies for argument and adjunct focalization in Basque 

(5)

CP1

Nork/[Mirenek]F

C′

esan du

TP Jonek

T′ vP

tDP

t[V+v+T] v′ t[V+v]

VP tV

CP2 tDP

C′

edan duela

TP

tDP

T′ vP

tDP

t[V+v+T] v′ t[V+v]

VP ura

tV

Importantly, Basque is well-known for also having an alternative to long distance extraction: clausal pied-piping (cf. Ortiz de Urbina 1989, 1993; Arregi 2003; Irurtzun 2007). Consider the data in (6) for Wh-questions and (7) for focalizations. In (6a,7a) we observe that the whole embedded clause is fronted to the left periphery of the matrix CP, and that there is OV inversion in the embedded clause, and SV inversion in the matrix clause, which results in a perfectly grammatical sentence. Lack of inversion (i.e., lack of T-to-C movement) in the matrix clause (6b,7b) or the embedded clause (6c,7c) results in ungrammaticality, as does the lack of inversion in both matrix and embedded clauses (6d,7d):

 Aritz Irurtzun

(6) a. [Nork edan duela ura] esan du Jonek?    who drink aux.c water say aux Jon ‘Who did Jon say that drank water?’ b. *[Nork edan duela ura] Jonek esan du?    who drink aux.c water Jon say aux ‘Who did Jon say that drank water?’ c. *[Nork ura edan duela] esan du Jonek?    who water drink aux.c say aux Jon ‘Who did Jon say that drank water?’ d. *[Nork ura edan duela] Jonek esan du?    who water drink aux.c Jon say aux ‘Who did Jon say that drank water?’ (7) a. [[Mirenek]F edan duela ura] esan du Jonek    Miren drink aux.c water say aux Jon ‘Jon said that [Miren]F drank water’ b. *[[Mirenek]F edan duela ura] Jonek esan du.   Miren drink aux.c water Jon say aux ‘Jon said that [Miren]F drank water’ c. *[[Mirenek]F ura edan duela] esan du Jonek   Miren water drink aux.c say aux Jon ‘Jon said that [Miren]F drank water’ d. *[[Mirenek]F ura edan duela] Jonek esan du   Miren water drink aux.c Jon say aux ‘Jon said that [Miren]F drank water’

Thus, the analysis of the derivation of clausal pied-piping structures is very similar to that of long distance extractions. As a first step, the Wh-phrase/focus phrase is moved to the specifier of the embedded clause, attracting its verbal complex, and then, instead of extracting it to the specifier of the matrix CP (as in the long distance extraction in (5)), it is the whole embedded CP that is moved to the specifier of the matrix CP, generating T-to-C movement in the matrix clause, as represented in (8), which corresponds to examples (6a) and (7a):



Strategies for argument and adjunct focalization in Basque 

(8)   Step 1:

CP1 C′ C

TP Jonek

T′ vP

tDP

[V+v+T] v′ t[V+v]

VP tV

CP2 Nork/[Mirenek]F

C′

edan duela

TP

tDP

T′ vP

tDP

t[V+v+T] v′ t[V+v]

VP ura

tV

 Aritz Irurtzun

Step 2:

CP1

[Nork/[Mirenek]F edan duela ura]CP2

C′

esan du

TP Jonek

T′ vP

tDP

t[V+v+T] v′ t[V+v]

VP



tCP2

tV

Finally, let me note that the parasitic movement of the verb with respect to the Wh-movement can serve as a disambiguator of scope for extractions that could be potentially ambiguous. This is, for instance, the case of modifiers like noiz (when) that could, potentially, be first-merged in either the matrix or the embedded clause, as in the examples in (9): (9) a. Noiz pentsatzen duzu [gerra bukatuko dela]? when think aux    war finish aux.c ‘When is it that you think that the war will finish?’

(When→think)

b. Noiz pentsatzen duzu [bukatuko dela gerra]? when think aux    finish aux.c war ‘According to you, when will the war finish?’

(When→finish)

A direct consequence of the parasitic nature of verb-movement is that a focus or interrogative displacement of an element in the matrix clause generates no inversion in the embedded clause (9a), but on the contrary, extraction from the embedded clause is accompanied by T-to-C movement in both embedded and matrix clauses (9b). As a consequence, the scopal properties of the interrogative have a direct mapping in the syntactic structure of the clause. As a matter of fact, a plausible answer to question (9a) could be something like “Whenever I see the occupation troops leaving the city” (i.e. that’s when I think that the war will come to an end), and a plausible answer to question (9b) something like “When the last city will be liberated” (i.e. I think that when such an event will be accomplished, the war will come to an end). Now, let us give a look at some of the syntactic restrictions that are inherent to the displacement operations in questions and focalizations.



Strategies for argument and adjunct focalization in Basque 

1.2  Islandhood restrictions We just saw that the displacements underlying questions and focalizations in Basque observe some locality and cyclicity restrictions which are shown with the movement of the verbal complex. However, these are not the only characteristic restrictions of Wh-questions and foci. In fact, given that they involve A′-displacements, Wh-questions and focalizations are also sensitive to various sorts of syntactic islands. For instance, extraction out of a coordination phrase results in plain ungrammaticality. This is illustrated in (10b,c): (10) a. Jonek [salda eta legatza] nahi ditu Jon    stock and hake want aux ‘Jon wants stock and hake’ b. *Zer nahi ditu Jonek [salda eta t ]?    what want aux Jon    stock and Lit. ‘What does Jon want stock and?’ c. *Zer nahi ditu Jonek [ t eta legatza]?    what want aux Jon and hake Lit. ‘What does Jon want and hake?’ (11) a. *[Legatza]F nahi du Jonek [salda eta t ]     hake want aux Jon stock and ‘Jon wants stock and [hake]F’ b. *[Salda]F nahi du Jonek [t eta legatza ]    stock want aux Jon and hake Lit. ‘What does Jon want and hake?’

And the same happens with extraction out of adjuncts (12b,c), left branch phrases (13b,c) or complex NPs (14b,c); Wh- or focus movement out of these islands derives in ungrammaticality: (12) a. Jon [abestia entzun duelako] poztu da Jon    song hear aux.because get.happy aux ‘Jon got happy because he heard the song’ b. *Zer poztu da Jon [t entzun duelako]?  what get.happy aux Jon hear aux.because Lit. ‘What did Jon got happy because he heard?’ c. *[Abestia]F poztu da Jon [t entzun duelako]     song get.happy aux Jon hear aux.because Lit. ‘Jon got happy because he heard [the song]F’ (13) a. Mirenek [Jonen liburua] irakurri du Miren    Jon’s book read aux ‘Miren read Jon’s book’

 Aritz Irurtzun

b. *Noren irakurri du Mirenek [t liburua]?  whose read aux Miren book ‘Whose book did Miren read?’ c. *[Jonen]F irakurri du Mirenek [t liburua]    Jon’s read aux Miren book ‘Miren read [Jon’s]F book’ (14) a. [Jonek liburu bat idatzi duelako zurrumurrua] entzun duzu    Jon book one write aux.c.p rumour hear aux ‘You heard the rumour that Jon wrote a book’ b. *Zer entzun duzu [Jonek t idatzi duelako zurrumurrua]?  what hear aux    Jon write aux.c.p rumour Lit. ‘What did you hear the rumour that Jon wrote?’ c. *[Liburu bat]F entzun dut [Jonek t idatzi     book one hear aux    Jon write duelako zurrumurrua] aux.c.p rumour

‘I heard the rumour that Jon wrote [a book]F’

So, as we said, all these restrictions constitute evidence that both Wh-questions and focalizations in Basque have a very similar syntax. In a nutshell, they both involve: i. A′-movement of the focus/Wh-phrase to Spec-CP.6 ii. T-to-C movement, which renders adjacency between the verbal complex (V+v+T) and the moved phrase. iii. Cyclicity in the focus/Wh-movement, which can be observed in the cyclic movement of the verbal complex of each clause. iv. A ban on extraction out of islands. v. Possible clausal pied-piping. This is the panoramic picture regarding the standard constructions.7 In the next sections, I will present the microparametric variability observed in the W ­ h-question and focalization strategies employed across Basque dialects. S­ ection 2 will briefly present the emergent Wh in situ strategy of Labourdin Basque, ­Section 3 will be devoted to two ‘reinforced focus’ constructions ((i) the “­rightward” ­focalization

.  Or Spec-FocP in “split CP” analyses (see, e.g. Ortiz de Urbina 1999; Irurtzun 2007). .  An anonymous reviewer reminds me that Elordieta (2001) mentions a potential point of variation between focalizations and Wh-constructions with respect to “weak crossover” effects (focalizations would generate a weaker agrammaticality effect than Wh-constructions). The observation is interesting but the phenomenon and judgments are not clear, and unfortunately we still have no in-depth study of this important issue.



Strategies for argument and adjunct focalization in Basque 

strategy of Southern dialects and (ii) the “movement+aux” construction of ­Northern dialects), and Section 4 will briefly present some other focus constructions attested in Basque dialects: a substandard mirative focus construction, and two different split interrogatives, which vary in their geographic distribution. Again, it should be emphasized that even if I will be dealing with some strategies that are restricted to some varieties, all varieties of Basque have the standard strategy that we just saw, and that the alternative strategies that I will comment on in the next sections are employed along with the standard ones (in fact, they are typically employed to convey different semantic nuances). 2.  The new wh in situ strategy of Navarro-Labourdin Basque Recently, Duguine and Irurtzun (2014) have discovered that young speakers of Navarro-Labourdin Basque also have another type of Wh-question construction, which is unavailable to older speakers of this dialect, and to the speakers of other dialects in Southern Basque Country (in the Spanish territory). This strategy is illustrated in examples (15) and (16): (15) Nork gereziak jan ditu? who cherries eat aux ‘Who ate the cherries?’ (16) Jonek zer fite jan du? Jon what quickly eat aux ‘What did Jon eat quickly?’

In these constructions, we do not observe the typical adjacency between the Wh-phrase and the verb, characteristic of standard strategies (in (15) the DO appears between the interrogative subject and the verb, and in (16) the adverb ‘fite’ (quickly) is sandwiched between the interrogative DO and the verb). Duguine and Irurtzun’s (2014) analysis is that underlying these examples there is an in situ Wh-strategy. In fact, in these constructions no interrogative displacement seems to take place (no Wh-movement and, as a consequence, no residual V2 effect), but furthermore, these constructions of Navarro-Labourdin Basque also share a range of properties with French wh in situ. In particular, they display intervention effects with negation: Wh in situ constructions are ungrammatical when the Wh-phrase is c-commanded by negation (17), the only way of asking a question with negation on the matrix clause being the Wh-movement strategy (18) (see Bošković 1998, 2000 or Mathieu 1999, 2004 for French data and analysis): (17) *Jonek ez du zer jaten?    Jon neg aux what eat ‘What doesn’t Jon eat?

 Aritz Irurtzun

(18) Zer ez du jaten Jonek? what neg aux eat Jon ‘What doesn’t Jon eat?’

Furthermore, these wh in situ constructions also pattern like French wh in situ paradigms regarding Wh-islands: Wh-phrases cannot remain in situ in Wh-islands (19), while they can move out of them (20): (19) *Ba-dakizu [nola nori opari bat eskaini]?   yes-know    how who present a offer ‘Do you know how to give a present to whom?’ (20) Nori ez dakizu [nola eskaini opari bat]? who neg know    how offer present a ‘Who don’t you know how to give a present to?’

Last, another characteristic property of Wh in situ constructions in French is that they can be embedded within strong islands (cf. Obenauer 1994; Shlonsky 2012). The in situ Wh constructions of Navarro-Labourdin Basque also show the same asymmetry: while regular Wh-movement displays island effects (21), Wh in situ constructions are just mildly deviant when embedded within strong islands (22): (21) *Nori piztu dute jendearen kexua [t etxea kentzean]?  who light aux people.of anger house remove.when Lit. ‘Who did they lit people’s anger when they took the house to?’ (22) ??[Nori etxea kentzean] jendearen kexua piztu dute?     who house remove.when people.of anger light aux Lit. ‘They lit people’s anger when they took the house to who?’

The syntactic and semantic similarity of these constructions with respect to the French Wh in situ has led Duguine and Irurtzun (2014) to the hypothesis that a catalyst for its emergence in Labourdin Basque is a transfer from French (which would be made possible by other 3rd factor effects like an innate bias for preferring movementless operations). In the next section I will briefly present two strategies that are employed for ‘reinforced focalizations’ across Basque dialects: the “rightward focalization” of Southern dialects (3.1), and the ‘displacement+auxiliary’ constructions of Navarro-Labourdin (3.2). 3.  Reinforced focus strategies In this section I give an overview of some ‘reinforced’ focalization strategies observed across Basque dialects. Generally these strategies are termed ‘reinforced’



Strategies for argument and adjunct focalization in Basque 

because they have a marked focus semantics associated to them; they are generally more contrastive and more presuppositional than the regular focalization constructions (to the point that they could be taken as semantic equivalents of cleft sentences (see below)). The two strategies that I will focus on are the ‘rightward focalizations’ of Southern dialects, and the ‘movement+auxiliary’ constructions of the Northern dialects (in particular, Navarro-Labourdin). 3.1  The ‘rightward’ focalization strategy of Southern dialects We saw that the standard focalization strategy available to all dialects comprises a leftward-dislocation of the focal element. Now, Southern dialects also have a different construction where the focus appears at the right edge of the clause, preceded by the rest of the sentence where the order of constituents tends to be the same as in out-of-the-blue sentences (although, it can certainly vary). For instance, the example in (23) would be a ‘marked’ variant of (24), with an enhanced degree of exhaustivity and a clear topic-focus (rising-falling) intonation: (23) [Jonek]F hautsi du mahaia    Jon break aux table ‘[Jon]F broke the table’

[Standard Construction]

(24) Mahaia hautsi du [Jonek]F table break aux   Jon ‘It’s [Jon]F that broke the table’

[Reinforced Construction]

These rightward focus constructions, as I said, are common to all Southern dialects (see, e.g., the dialectal study in Hualde, Elordieta and Elordieta 1994). Ortiz de Urbina (2002) argues that in this type of constructions we have the regular focus movement of the [+F] marked XP illustrated in (25a), followed by “remnant movement” of the rest of the clause to a higher TopP (25b): (25) a. FocP [ XP[+F] [Focº CP […tXP…]]]] b. TopP[ CP [Topº FocP [ XP[+F] [Focº tCP ]]]]

One of the virtues of an analysis along these lines is that it immediately explains the sentence-final position of the focus (given that it directly follows from the fact that the rest of the clause is remnant-moved to a position higher than Spec-Foc). Besides, it follows naturally that a topic-comment intonational contour accompanies them, and the highly contrastive nature of their interpretation is also explained as deriving from the Topic position of the clause. Furthermore, this analysis can

 Aritz Irurtzun

also explain some scopal relations that would otherwise be hard to explain. Consider the rightward focus construction of (26), where the focal XP takes scope over negation, linear order notwithstanding: (26) Ez da etorri [horregatik]F not aux come    because.of.that ‘He did not come [because of that]F’ (=[That]F , and not (the) other one, is the reason why he did not come).

On this approach, we expect negation not to take scope over the focal horregatik, for Negº is contained within the remnant-moved phrase and from there it cannot c-command the purpose-clause horregatik. The availability of the remnant movement operation (25b) would be restricted to Southern varieties of Basque, which would account for the restricted distribution of these constructions.8 3.2  The ‘movement+aux’ strategy of Navarro-Labourdin The Navarro-Labourdin dialect does not generally display the rightward focalization strategy, but it has a different reinforced construction to express exhaustive focalization: the movement+aux construction. This construction, which coexists in Navarro-Labourdin with the standard construction, is characterized by involving a ‘canonical’ Wh-movement to Spec-CP but instead of being accompanied by movement of the verbal complex to Cº, it only involves movement of the auxiliary to T. This is illustrated in (27b), a reinforced variant of the standard (27a):9 (27) a. [Mirenek]F galdegin du izokina [Standard focus construction]    Miren ask aux salmon ‘[Miren]F ordered salmon’ b. [Mirenek]F du izokina galdegin. [Reinforced construction]    Miren aux salmon ask ‘It’s [Miren]F that ordered salmon’

This is a construction that is restricted to main clauses, and according to the analysis in Duguine and Irurtzun (2010), the reinforced nature of these constructions .  It should be noted that the speakers of these varieties of Basque (all Spanish-Basque bilinguals) also have a very similar construction in Spanish, namely, the non-echoic sentence-final Wh-questions (cf. Uribe-Etxebarria 2002; Etxepare and Uribe-Etxebarria 2004). .  I translate example (27b) with a cleft in order to express the reinforced nature of these constructions. Actually, Lafitte (1944: 48) when commenting on these constructions says that ‘pour le traduire, le français ce que est obligatoire’ [in order to translate it, French ce que is necessary]. See Duguine and Irurtzun (2010) for discussion.



Strategies for argument and adjunct focalization in Basque 

is clearly observed in their semantic interpretation regarding the parameters of presuppositionality and exhaustivity.10 For instance, even if all Wh-questions take as granted that the eventuality described in the question takes place, the presupposition of reinforced constructions is stronger, and a discourse clash or clear incongruence is generated when the presupposition is refuted. Consider the data in (28) and (29). In (28) we observe that the presupposition of the Wh-question in (28A) is directly refuted in the answer in (28B), but there is no sharp discourse clash. On the contrary, the refutation in (29B) clashes directly with the presupposition expressed in the reinforced Wh-question in (29A), as it would do with a clefted question in English: (28) A: Zer jan duzu? what eat aux ‘What did you eat?’

[Standard Wh-construction]

B: Deus ez nothing not ‘I ate [nothing]F’ (29) A: Zer duzu jan? what aux eat ‘What is it that you ate?’

[Reinforced Wh-construction]

B: #Deus ez   nothing not ‘It’s [nothing]F that I ate’

Likewise, reinforced interrogative and focus constructions always require a strongly exhaustive or complete answer interpretation (cf. Duguine and Irurtzun 2010). Thus, contrary to standard constructions, they cannot appear with additive particles (30), and they cannot be employed in a pair-list answer (31):

.  The restriction to main clauses makes that extraction out of embedded clauses using this strategy renders ungrammaticality, as represented in (i), with the reinforced strategy only in the embedded clause, and (ii), with the embedded strategy in both clauses: (i) *[Mirenek]F erran dut [t duela Jon ikusi]     Miren say aux aux.c Jon see ‘I said that Jon saw [Miren]F’ (ii) *[Mirenek]F dut erran [t duela Jon ikusi]    Miren aux say aux.c Jon see ‘I said that Jon saw [Miren]F’

 Aritz Irurtzun

(30) a. Nor besterik jinen da? who else come.fut aux ‘Who else will come?’

[Standard Wh-construction]

b. *Nor besterik da jinen?    who else aux come.fut ‘Who else is it that will come?’

[Reinforced Wh-construction]

(31) a. Jon jinen da, Peio jinen da… [Standard Jon come.fut aux Peio come.fut aux Wh-construction] ‘Jon will come, Peio will come…’ b. *Jon da jinen, Peio da jinen… [Reinforced    Jon aux come.fut Peio aux come.fut Wh-construction] ‘It’s Jon that will come, it’s Peio that will come…’

The availability of this strategy, as I said, is restricted to Northern dialects, and it may be correlated with a range of other phenomena involving participial periphrases (cf. Etxepare 2014 for a recent microparametric analysis). Having briefly analysed the reinforced constructions, in Section 4 I will overview two other focus constructions; a substandard mirative construction of Southern Basque, and two split interrogative (Wh-question+focus) constructions, one of them available to all speakers across Basque dialects, the other one restricted to Southern dialects. 4.  Other Constructions Along with the standard and reinforced constructions of each dialect, there are also other constructions which, even though not evenly spread across the population deserve a brief mention here, given that they also have characteristic correspondences between their syntactic form and their semantic interpretation. Here I would like to briefly comment on two of these constructions; the mirative focus (Section 4.1), and the split interrogative constructions (Section 4.2.). 4.1  Mirative focus constructions Mirativivity in Basque has not got an in-depth treatment in the literature, and its grammatical status is not very clear. However, Etxepare (1998) discusses some constructions that, even though they are not fully grammaticalised nor evenly spread across the population, can be taken as instances of “mirative focus” constructions. Etxepare (1998) concentrates on the differences between the type of focus constructions like (32) that we have seen in Section 1, and



Strategies for argument and adjunct focalization in Basque 

mirative focus constructions like (33) which would involve some sort of exclamation:11,12 (32) [Jonek]F ekarri du ardoa    Jon bring aux wine ‘[Jon]F brought wine’

[Standard focus construction]

(33) [Jonek]F ardoa ekarri du    Jon wine bring aux ‘[Jon]F brought wine!’

[Mirative focus]

According to the analysis in Etxepare (1998), there is a sharp semantic difference between standard constructions like (32) and mirative constructions like (33) in that standard constructions conversationally implicate the eventuality denoted by the open proposition in the (potential) question they answer, whereas mirative constructions conventionally implicate it. This, according to his analysis, would correspond with the different syntactic position that each focal element occupies: whereas standard foci would undergone A′-movement to Spec-CP (along the lines I presented in Section 1), mirative foci would move to a lower A-position, presumably the projection of Infl (or T). 4.2  Split interrogatives The last construction that I would like to comment on is the “split interrogative” construction, which is a confirmatory construction combining a Wh-question and a focal phrase. The Wh-phrase appears to be in its canonical left-peripheric

.  Actually, the terminology employed in Etxepare (1998) is a bit different; he terms “­Emphatic Focus” what we have analysed as (normal) “focus” (basically, the constructions in Section 1), and he terms “Contrastive Focus” the constructions that we analyse here as “mirative”. At the risk of confusing the reader (sorry!), I think that the terminology employed in this chapter is more accurate and, furthermore, it better matches the crosslinguistic use of those terms in the literature, so I will keep to the distinction between “focus” and “mirative focus” when I talk about what Etxepare (1998) calls “emphatic focus” and “contrastive focus”. The reader should be aware of the difference in terminology. .  A reviewer rightly points out that the examples I use to illustrate mirative focus only involve subjects. I do so because thus we can assess the lack of linear continuity between the focus and the verb. Evidence with other syntactic objects would require a longer argumentation (involving an explanation of the properties of scrambling of nonfocal material and speech act modifiers) and it would take us too far from our expository purposes. The reader is referred to Etxepare (1998) for discussion.

 Aritz Irurtzun

­ osition and the focus (the potential answer) in sentence-final position. This is p illustrated in (34): (34) Zer nahi du Jonek, [ogia]F? what.abs want aux Jon    bread.abs Is it the bread that Jon wants? (Lit. ‘What does Jon want, the bread?’)

Irurtzun (2014) argues that we have to distinguish two types of split questions. On the one hand, we would have matching questions, where the Wh-phrase to the left matches in syntactic type and semantic interpretation with the focus phrase to the right. (35) would be a canonical example of a matching construction where the focal element to the right bears the same case (ergative) and theta-role as the fronted Wh-phrase: (35) Nork egin du, [Jonek]F? [Matching type] who.erg make aux   Jon.erg ‘Was it [Jon]F that made it?’ (Lit. ‘Who made it, Jon?’)

These constructions are common to all dialects of Basque. On the other hand, we have what-type questions; constructions that invariably show the dummy interrogative pronoun zer (what) to the left, which does not match with the focal phrase (36). These are restricted to the Southern dialects: (36) a. Zer egin du, [Jonek]F? [What type] what make aux   Jon.erg ‘Was it [Jon]F that made it?’ (Lit. ‘What made it, Jon?’) b. Zer zatoz, [bihar]F? [What type] what come   tomorrow ‘Is it [tomorrow]F that you are coming?’ (Lit. ‘What are you coming, tomorrow?’)

According to Irurtzun (2014), matching type split interrogatives are derived from a simple bi-clausal construction: a regular Wh-question followed by a leftward focalization which undergoes sluicing (following the analysis of Spanish matching constructions proposed by Arregi 2010). Therefore, the fact that all dialects of Basque have this type of constructions is just something to be expected, given that all dialects have both Wh-movement and sluicing constructions. Regarding what type constructions, the analysis put forth in Irurtzun (2014) suggests that they involve a more complex construction with two CPs and multidominance of the clausal spine, which is dominated by an evidential head. As I said, the availability of this complex structure is restricted to Southern dialects (probably, due to a transfer from Spanish, which also has what-type interrogatives like (37), cf. López-Cortina 2003, 2007; Arregi 2007, 2010):



Strategies for argument and adjunct focalization in Basque 

(37) ¿Qué vienes, [mañana]F?    what come    tomorrow ‘Is it [tomorrow]F that you are coming?’ (Lit. ‘What are you coming, tomorrow?’)

Northern dialects, on the contrary, cannot make what-type split interrogatives. 5.  Conclusions In conclusion, in this chapter I have provided an overview of the different question and focalization strategies available across Basque dialects. We have seen that a core property of Basque syntax is that it treats in a very similar way Wh-questions and their answers (focus constructions). The standard strategy (available to all varieties) involves an A′-movement of the focal item to Spec-CP (which is accompanied by movement of the verb) but there are also a range of other alternative strategies employed across Basque varieties. These alternative strategies are generally employed in order to gain some semantic nuance (stronger presuppositionality, exhaustivity, etc.) with respect to the standard strategy.

References Altube, Seber. 1920. De sintaxis euskérika. Donostia: Martin Mena y Cía. Arregi, Karlos. 2002. Focus on Basque movements. Ph.D. Dissertation: MIT. Arregi, Karlos. 2003. “Clausal pied-piping.” Natural Language Semantics 11: 115–143.

doi: 10.1023/A:1024499226731

Arregi, Karlos. 2007. “Syntax and semantics of split questions.” In Romance Linguistics 2006. Selected Papers from the 36th Linguistic Symposium on Romance Languages (LSRL), José Camacho, Nydia Flores-Ferrán, Liliana Sánchez, Viviane Déprez, and María José Cabrera (eds.), 15–28. Philadelphia: John Benjamins.  doi: 10.1075/cilt.287.03arr Arregi, Karlos. 2010. “Ellipsis in Split Questions.” Natural Language and Linguistic Theory 28: 539–592.  doi: 10.1007/s11049-010-9097-x Artiagoitia, Xabier. 2000. Hatsarreak eta Parametroak Lantzen. Vitoria-Gasteiz: Arabako Foru Aldundia. Bošković, Željko. 1998. “LF movement and the Minimalist Program.” In Proceedings of the 28th North East Linguistic Society, Pius N. Tamanji and Kiyomi Kusumoto (eds.), 43–58. ­University of Toronto: Graduate Linguistic Student Association. Bošković, Željko. 2000. “Sometimes in [Spec, CP], sometimes in situ.” In Step by step, Roger Martin, David Michaels and Juan Uriagereka (eds.), 53–156. Cambridge: MIT Press. Duguine, Maia and Irurtzun, Aritz. 2014. “From obligatory Wh-movement to optional Wh in situ in Labourdin Basque.” Language 90.1.  doi: 10.1353/lan.2014.0006 Duguine, Maia and Irurtzun, Aritz. 2010. “Opérateurs d’exclusivité dans les questions et réponses en basque.” Revue de Sémantique et Pragmatique 27: 105–17.

 Aritz Irurtzun Eguzkitza, Andolin. 1987. Topics on the Syntax of Basque and Romance. Ph.D. Dissertation: UCLA. Elordieta, Arantzazu. 2001. Verb Movement and Constituent Permutation in Basque. Utrecht: LOT. Elordieta, Arantzazu and Haddican, Bill. This volume. “Strategies of Verb and Verb Phrase Focus across Basque Dialects”. Elordieta, Gorka and Irurtzun, Aritz. 2010. “The relationship between meaning and intonation in non-exhaustive answers: evidence from Basque.” The Linguistic Review 27.3: 261–291. Erdozia, Kepa, Laka, Itziar, Mestres-Missé, Anna and Rodriguez-Fornells, Antoni. 2009. “­Syntactic complexity and ambiguity resolution in a free word order language: Behavioral and electrophysiological evidences from Basque.” Brain and Language 109: 1–17.

doi: 10.1016/j.bandl.2008.12.003

Etxepare, Ricardo. 2014. “The microparameter in Basque participial periphrases.” Paper ­presented at CamCoS 3, University of Cambridge, Cambridge. Etxepare, Ricardo. 1998. “A Case for Two Types of Focus in Basque.” In UMOP 21: Proceedings of the Workshop on Focus, Elena Benedicto, Maribel Romero and Satoshi Tomioka (eds.), 65–81. Amherst, GLSA. Etxepare, Ricardo and Ortiz de Urbina, Jon. 2003. “Focalization.” In A Grammar of Basque. José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 465–522. Berlin and New York: Mouton de Gruyter. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2004. “In situ wh-phrases in Spanish: locality and quantification.” Recherches Linguistiques de Vincennes 33: 9–34. Hualde, Jose Ignacio, Elordieta, Gorka and Elordieta, Arantzazu. 1994. The Basque dialect of Lekeitio [Supplements of ASJU 34]. Gipuzkoako Foru Aldundia and UPV/EHU. Irurtzun, Aritz. 2007. The Grammar of Focus at the Interfaces. Ph.D. dissertation, University of the Basque Country UPV/EHU. Irurtzun, Aritz. 2014. “Euskarazko galdera erdibituak: analisi bitar baterantz.” In Eridenen du zerzaz kontenta. Sailkideen omenaldia Henrike Knörr irakasleari (1947–2008). Ricardo Gómez and Mari Jose Ezeizabarrena (eds.), 403–422. Bilbo: UPV/EHU. Kiss, Katalin É. (ed.). 1995. Discourse Configurational Languages. Oxford: Oxford University Press. Lafitte, Piarres. 1944. Grammaire Basque (Navarro-Labourdin littéraire), Revised and corrected edition, Bayonne: Editions des “Amis du Musée Basque” and Ikas. Laka, Itziar and Uriagereka, Juan. 1987. “Barriers for Basque and vice-versa.” In Proceedings of NELS 17, Joyce McDonough and Bernadette Phunkett (eds.), 394–408. Amherst: GLSA. López-Cortina, Jorge. 2003. “The structure of split interrogatives.” In Theory, practice, and acquisition: Selected Papers from the 6th Hispanic Linguistics Symposium and the 5th Conference on the Acquisition of Spanish and Portuguese, Paula Kempchinsky and Carlos-Eduardo Piñeros (eds.), 140–155. Somerville: Cascadilla Press. López-Cortina, Jorge. 2007. The Spanish left periphery: Questions and answers. PhD. Dissertation: Georgetown University. Mathieu, Eric. 1999. “Wh in situ and intervention effects.” UCL Working Papers in Linguistics 11: 441–72. Mathieu, Eric. 2004. “The mapping of form and interpretation: The case of optional ­WH-movement in French.” Lingua 114.(9–10): 1090–1132. doi: 10.1016/j.lingua.2003.07.002



Strategies for argument and adjunct focalization in Basque 

Mitxelena, Koldo. 1981. “Galdegaia eta mintzagaia euskaraz.” In Euskal linguistika eta literatura: Bide berriak, Bilbo: Deustuko Unibertsitatea, 57–81 [Reprinted in Mitxelenaren euskal idazlan guztiak VI, 1988, Joseba Andoni Lakarra and Blanca Urgell (eds.), 137–167 Zarautz: Euskal Editoreen Elkartea]. Obenauer, Hans-Georg. 1994. Aspects de la syntaxe A-barre. Effets d’intervention et mouvements des quantifieurs. Thèse d’Etat, Université Paris 8. Ortiz de Urbina, Jon. 1989. Some Parameters in the Grammar of Basque. Dordrecht: Foris Publications. Ortiz de Urbina, Jon. 1993. “Feature percolation and clausal pied-piping.” In Generative Studies in Basque Linguistics, José Ignacio Hualde and Jon Ortiz de Urbina (eds.), 189–219. Amsterdam: John Benjamins.  doi: 10.1075/cilt.105.08urb Ortiz de Urbina, Jon. 1995. “Residual verb first and verb second in Basque.” In Discourse Configurational Languages, K. É. Kiss (ed.), 99–121. New York: Oxford University Press. Ortiz de Urbina, Jon. 1999. “Force phrases, focus phrases and left heads in Basque.” In Grammatical Analyses in Basque and Romance Linguistics: Papers in Honor of Mario Saltarelli, Jon Andoni Franco, Alazne Landa and Juan Martín (eds.). 179–194. Amsterdam: John Benjamins.  doi: 10.1075/cilt.187.11urb Ortiz de Urbina, Jon. 2002. “Focus of Correction and Remnant Movement in Basque.” In Erramu Boneta: Festscrift for Rudolf P.G. de Rijk, Xabier Artiagoitia, Patxi Goenaga and Joseba Andoni Lakarra (eds.). 510–524. Bilbo: Gipuzkoako Foru Aldundia and UPV/EHU. Rebuschi, Georges. 1997. “A Note on Focalization in Basque.” In Essais de linguistique basque. [Supplements of ASJU 35], Georges Rebuschi (ed.), 31–42. Bilbo: Gipuzkoako Foru ­Aldundia and UPV/EHU. de Rijk, Rudolph P.G. 1969. “Is Basque an SOV Language?” Fontes Linguae Vasconum 1: 319–351. Shlonsky, Ur. 2012. “Notes on Wh In Situ in French.” In Functional Heads: The Cartography of Syntactic Structures, Volume 7, Laura Brugé et al. (eds.), 242–252. Oxford: Oxford ­University Press. Uriagereka, Juan. 1999. “Minimal restrictions on Basque movements.” Natural Language and Linguistic Theory 17: 403–444.  doi: 10.1023/A:1006146705483 Uribe-Etxebarria, Myriam. 2002. “In situ questions and masked movement.” Linguistic Variation Yearbook 2: 217–257.  doi: 10.1075/livy.2.09uri

Microsyntactic variation in the Basque hearsay evidential Ricardo Etxepare & Larraitz Uria CNRS – IKER, UMR5478 / IXA, UPV/EHU

This paper studies the syntactic distribution of the Basque hearsay evidential omen across Basque dialects. We test the hypothesis that the complex dialectal variation regarding the syntactic distribution exhibited by omen follows from a basic syntactic microparameter: in central varieties the features that constitute the hearsay evidential are projected as part of the clausal spine; in Eastern varieties, those features are compiled as a separate syntactic term, merged as the Specifier of an independent Evidential Head. This microparameter accounts in a unified way for most of the differences in the syntactic distribution of the evidential. Together with this basic microparameter, the comparative study of Basque dialects seems to support the idea that certain functional projections (like NegP, FocP or EvP) may appear in more than one position in the cartography of the clause, and more specifically, at the edge of cyclic domains. Keywords:  hearsay evidential; modality; left periphery; syntactic cartography

1.  Introduction1 The present chapter has as its aim to describe and account for the syntactic distribution of the Basque hearsay evidential omen across the Basque dialectal continuum,

.  We would like to thank the audiences of the conference The Nature of Evidentiality (June 2012) held at the University of Leiden, as well as of the Workshop L’Estudi de la Variació Sintàctica: Eines Pràctiques i Aspectes Teòrics, held at the University of the Basque Country in Vitoria-Gasteiz in December 2013, for their comments and suggestions. We are also thankful to the reviewers of this paper, who helped us to greatly improve it. We acknowledge financial support from the MINECO Project FF2011-26906 (VALAL, FFI2014-53675-P) (Spain), the MINECO Project FFI2014-51878-P (Spain), the Basque Government’s HiTT, IT769-13 Project as well as the UPV/EHU (UFI11/14) Project from the University of the Basque Country. The research leading to these results has also received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement nº 613465.

doi 10.1075/lfab.13.10etx © 2016 John Benjamins Publishing Company

 Ricardo Etxepare & Larraitz Uria

particularly in its eastern varieties, where a richer variability in the positioning of the hearsay evidential can be observed. We will advance the idea that hearsay evidentiality in Basque, as represented by the particle omen, is syntactically expressed by means of two different syntactic configurations. As we will show, the hearsay evidential in Basque plausibly lexicalizes two syntactic features. Those two features can be combined into an independent syntactic term, or be separately projected as part of the clausal spine. This basic microparameter triggers a number of surface differences in the syntactic distribution of the evidential in central and eastern dialects. For reasons of space, we will not deal with the possible semantic consequences of this variation (see Etxepare 2010a, for a preliminary approach, and Zubeldia 2010, for a detailed analysis of the use of the evidential in central dialects). The paper is organized as follows: in Section 1 we provide a basic characterization of the meaning of the hearsay evidential omen, as well as a preliminary overview of the distributional differences that we find in the Basque dialectal domain; in Section 2, we present in detail the main syntactic differences underlying the varying distribution of omen in central and eastern dialects; in Section 3, we propose the basic syntactic microparameter that underlies those differences. Section 4 explores some further consequences of the underlying microparameter. Section 5 concludes. The empirical data on which this paper is based comprises independent fieldwork realized in 33 eastern Basque speaking villages among native speakers of the corresponding variety aged in between 50 and 90, and data from a dialectal corpus constituted by literary texts spanning from the 1960s to our days. All the data can be freely accessed in the on-line database Basyque, at http://ixa2.si.ehu.es/atlas2/ index.php?lang=en (see Uria and Etxepare 2011). 2.  Omen: Some basic descriptive elements 2.1  The meaning of the evidential Basque possesses an invariant particle omen meaning “reportedly” or “they say”. Thus, (1a) is an ordinary assertion expressing the proposition that Xabier came, whereas (1b) asserts that Xabier came and that the proposition that Xabier came is supported by hearsay evidence: (1) a. Xabier etorri da Xabier come: ptcp is ‘Xabier came’ b. Xabier etorri omen da Xabier come:ptcp evid is ‘Xabier reportedly came’



Microsyntactic variation in the Basque hearsay evidential 

The meaning contribution of omen in (1b) can be stated as follows (adapted from Zubeldia 2010: 12): let S be a sentence, p the proposition expressed by it and Somen a sentence containing omen: (2) Somen(p)-> someone who is not the speaker has said that p

The informal definition in (2) fits well Aikhenvald’s strict definition of evidentiality (2004:3) as a linguistic category whose basic meaning is “source of evidence”. For her, in order to be classified as an evidential, a grammatical formative must have “source of information” as its fundamental meaning. She counts omen among the evidential particles, following Jacobsen (1986). For Aikhenvald, the linguistic category of hearsay evidentiality does not necessarily have a single and fixed position in the clause. Her definition of the evidential category is functional, not structural. Within the cartographic project advocated by Cinque (1999) on the other hand, hearsay evidentials have a fixed position in the clausal spine, in between the category that expresses an evaluative judgement (cf. adverbs such as fortunately), and epistemic modals. This position is illustrated by means of the following two Korean examples. In the first one (3a), it is shown that the evidential affix precedes the evaluative affix in the morpheme-sequence that constitutes the finite form in Korean. In (3b), it is shown that the epistemic affix precedes the evidential one. Assuming that the sequence of morphemes is a faithful reflection of syntactic hierarchic relations, Cinque concludes that the evidential head is higher in the clausal structure than the epistemic modal head, and that the evaluative projection dominates both of them, as in (3c) (Cinque 1999: 71; see also the discussion on credo and dicono in Giorgi 2010): (3) a. Minca-nun ttena-ss-te-kwun-yo M.-top leave-past-evid-evaluat-polite ‘I noticed that M. had left!’ b. Ku pwun-I cap-hi-si-ess-ess-keyss-sup-ti-kka? the person-nom catch-pass-agr-ant-past-epist-agr-evid-q ‘Did you feel that he had been caught?’ c. [SA…[EM Evaluative Mood [EvidP Evidential Mood [EpisMP Epist Mood…]]]]

The existence of a specifically epistemic evidential strategy in the eastern dialects of Basque allows us to establish the relative position of the reportative evidential particle vis-à-vis epistemic modality. This epistemic evidential strategy is based on the apodosis form of the conditional, and can be compared to the conditional based evidential strategy observed in various Romance languages, among which French, the contact language with the Basque eastern varieties. The following is an illustrative example of French.

 Ricardo Etxepare & Larraitz Uria

(4) Le premier ministre serait souffrant the prime minister aux.conditional ill ‘Reportedly, the prime minister is ill’

Forms such as (4) (see Ducrot 1984; Dendale and Collier 2003; Squartini 2004) combine two perspectives: one reflecting the speaker’s point of view, which contributes the proposition that the prime minister is ill, and another one which has its source in someone else’s saying, and which assigns its evidential meaning to those forms. This construction is paralleled by special modal auxiliaries in other languages (German sollen, Schenner 2008, or the Konjunktiv system in that same language, see Wiemer 2010 for references). Eastern Basque has a strategy similar to the Romance ones, based on the apodosis form of the conditional. (5) is an example borrowed from the eastern periodical Herria. The evidential meaning is contributed by the auxiliary litaike, an irrealis finite form based on the verb edin, the auxiliary root used with subjunctives and modals, and the modal affix -ke, which outside the conditional signals the potential mood (see Oyharçabal 2003): (5) Ez da oraino jakina nork erosiko duen neg is still known who buy:fut aux.comp Donapaleuko herria erosletarik litaike StPalais:gen village buyers:abl aux.conditional ‘It is not known yet who will buy it. The village of St Palais is reportedly among the buyers’

Unlike the hearsay particle, the conditional evidential constructions are incompatible with speaker certainty regarding the truth of the proposition. They are in this sense genuinely modal strategies. A sequence such as (6), which assumes the falsity of the reported proposition in (5), is impossible: (6) #Donapaleuko herria erosletarik litaike,   St Palais:gen village buyers:abl aux.conditional, baina egiazki ez da erosleen artean but in.truth neg is buyers:gen among ‘#The village of St Palais is reportedly among the buyers, but it is not among the buyers’

In this, the conditional evidential strategy behaves as an epistemic modal, and differs from the evidential particle omen, which is compatible with speaker certainty (as observed by Etxepare 2010a; Zubeldia 2010): (7) Donapaleuko herria erosleen artean omen da, St Palais:gen village buyers:gen among evid is



Microsyntactic variation in the Basque hearsay evidential 

baina egiazki ez da but in.truth neg is ‘The village of St Palais is among the buyers, they say, but it isn’t really’

Interestingly, the evidential particle can co-occur with the modal strategy, as shown in (8a,b): (8) Erran dutenaren arabera, ba omen litazke said have:rel;gen according aff evid aux.conditional 200 bat bederen desagertuak 200 one at.least disappeared ‘According to what they say, there are reportedly around 200 people missing’

How should we express this state of affairs in formal terms? On the one hand, both items (the evidential modal and the evidential particle) are independent of each other, witness the fact that they can independently occur. On the other hand, the two elements are adjacent, and they are not separable by anything else. A simple way to accommodate those two facts is to propose a sequence of functional heads à la Cinque (1999), for those varieties: (9) …[EvP Omen Evid0 [EpistModP Mood0…]]

2.2  Variation Microvariation studies can shed some light on the issue of how flexibly evidentials of a particular type are represented. The Basque dialectal spectrum presents a certain degree of variation concerning the syntactic behaviour of the hearsay evidential. For instance, all dialects share the following two sequences, where the evidential immediately precedes the auxiliary: (10) a. Ez omen da etorri   a’. Polarity Omen Aux neg evid is come:ptcp ‘Reportedly, she/he didn’t come’ b. Etorri omen da   b’. Lexical Verb Omen Aux come:ptcp evid is ‘Reportedly, he/she has come’

But a corpus-based search yields other possible orders for eastern dialects (simplifying, the dialects spoken in the French side of the Basque Country): they allow a higher degree of freedom in the syntactic realization of the hearsay evidential (11). (11) a. Protestanteen komentu batean omen protestant:gen convent one:loc evid

 Ricardo Etxepare & Larraitz Uria

eskolarazi zuen provide.with.schooling aux

‘He/she got him/her educated in a protestant convent, they say’

b. Eüskaldüna zen lehenago omen popülü ez ikasia basque was before evid folk neg learned ‘The Basques were an ignorant folk then, they say’ c. Joanes XXIII-garrenak hori bezalako hitzak John 23-ord;erg that like:gen words erran eta manatu txi omen say:ptcp and command: ptcp aux evid ‘Pope John the 23rd said and ordered to say words like those ones, reportedly’

The noted examples all add up to the following set of sequences (with Pol a polarity head external to the Auxiliary Phrase, see Laka 1990), divided in terms of two main dialectal areas: (12) a. Omen Verb Aux…  EASTERN DIALECTS b. Aux… omen… c. Verb Aux… omen

d. Pol Omen Aux  CENTRAL DIALECTS and EASTERN DIALECTS

e. Verb Omen Aux

It is not evident, within a cartographic approach, how to deal with the apparent freedom of relative orders available to the hearsay evidential in eastern dialects. This work is an attempt to account for this basic aspect of cross-dialectal syntactic variation and explore the ways in which this variation can be accommodated in a rigid syntactic structure of the cartographic sort. 2.3  Claims We will make the following specific claims in the light of the variation found in the Basque dialectal area: (i) omen is part of the clausal spine in central varieties, in such a way that its complex morphological structure corresponds to the projection of two adjacent heads o- and -men immediately preceding the temporal and modal projections of the finite auxiliary; (ii) in eastern varieties, the relevant features are not part of the clausal spine, but constitute a separate syntactic term, merged as the Specifier of an independent Evidential Head. Many of the differences between central and eastern varieties will be shown to follow from this basic syntactic difference in feature compilation. Since omen is an independent term in eastern varieties, eastern varieties allow a richer array of evidential configurations: we will



Microsyntactic variation in the Basque hearsay evidential 

show that this array follows straightforwardly from the local computations arising from Agree (defined as in Pesetsky and Torrego 2007). In the following section, we review the basic differences in the distribution of the evidential particle in central and eastern varieties. 3.  Structural status of the hearsay evidential 3.1  Omen and the inflected verb/auxiliary In central varieties, omen always shows up together with the inflected auxiliary (cf. 1b, repeated below): (13) Etorri omen da come:ptcp evid is ‘Reportedly, he/she came’

This is not the case in eastern varieties, as shown below: (14) a. Aldiz Euskal Herrian ere gazte batek on.the contrary Basque Country:loc too young.person one:erg omen gauza bera egin-ik, amanda eta presondegia evid thing same done-part, fine and prison ukan zituen had aux ‘On the contrary, a Basque young person having reportedly done the same thing, he suffered imprisonment and had to pay damages’ b. Denbora batez, balearen mihia, hura izan-ki omen time one:ins whale:gen tongue that be-nmlz evid haragi hoberena, meat best Baionako apezpikuari oparitzen zen ardurenik… Bayonne:gen bishop:dat offer:ipfv aux often ‘At the time, the whale’s tongue being reportedly the best piece of meat, it was often offered to the bishop of Bayonne’

That does not mean that central dialects may not appear without an overt auxiliary. Here are some cases (also possible in Eastern dialects): (15) a. Bere bertsoak rimadun prosa omen, ez giar … his verses rhyme.having prose evid, neg force ‘His verses (are) mere prose with rhyme; (with) no force…’ b. Donibanez bildu lorak tisanan artu ezkero, St John:ins collect:ptcp flowers infusion:loc take after

 Ricardo Etxepare & Larraitz Uria

ezinobeak omen zenbait eritasun sendatzeko great evid some illness recover:purp ‘If you take flowers collected during St John’s time, they (are) reportedly great to recover from some illnesses’

There is however an important difference between central and eastern varieties: the former cases can be trivially accounted for in terms of the elision of the finite auxiliary. The eastern ones, which have the evidential particle embedded in a nonfinite dependent (either perfective or gerundive), cannot. 3.2  Omen and the clitic status of the auxiliary The finite forms in Basque, meaning by this auxiliaries, but also directly inflected lexical verbs (so called synthetic verbs, see De Rijk 2008) such as the verb etorri ‘come’ below, cannot occur in the first position of the sentence. In this regard, they behave as clitic auxiliaries (Ortiz de Urbina 1994, 1995; Etxepare and Uribe‑­ Etxebarria 2009). Since they cannot start a sentence, they require the support of a designated set of stress-carrying elements, like overt polarity heads, foci or the lexical verb in the case of the auxiliary: (16) a. *Dator  comes ‘He/she comes’ b. Badator aff-comes ‘He/she comes’ c. Ez dator neg comes ‘He/she does not come’ d. Norbait dator someone comes ‘Someone comes’ e. Etorri da come:ptcp is ‘He/she came’

The hearsay evidential omen cannot save the auxiliary from its initial position in central dialects. Compare in this regard (17a,b) and (18a,b), the latter drawn from eastern dialectal corpora: (17) a. *Omen da paper    evid is paper ‘Reportedly, there is paper’



Microsyntactic variation in the Basque hearsay evidential 

b. *Omen zen gizon bat    evid was man one ‘Reportedly, there was a man’ (18) a. *(Omen) da paper, eta bertzerik     evid is paper and other.stuff ‘Reportedly, there is paper and other stuff ’ b. *(Omen) zen gizon bat,…      evid aux man one ‘Reportedly there was a man…’

(Etchepare 1905)

(Etxamendi 2011 [1987], IV:151)

At least for some speakers in the eastern dialectal area, omen can rescue the auxiliary from first position in eastern dialects (18a, b), but it is unable to do so in central dialects. In central dialects, one of the designated first-position elements must necessarily precede the evidential. We illustrate this fact by the analogous central examples (19a, b), which possess an overt positive polarity head ba-: (19) a. Ba-omen da paper aff-evid is paper ‘Reportedly, there is paper’ b. Ba-omen zen gizon bat aff-evid was man one ‘Reportedly, there was a man’

3.3  Parenthetical omen All eastern varieties share the parenthetical use of omen. Parenthetical omen constitutes an independent intonation phrase, and can be inserted among the major constituents of the sentence, as well as in the right and left edges: (20) a. Erdi aroan, omen, Nafarroa osoa hona middle Age:loc, reportedly, Navarre whole here etortzen zen come:ipfv aux

‘During the middle ages, all of Navarre used to come here’

b. Organo barrokoa ikustekoa da, omen organ baroque see:nmlz;pur is, evid ‘The baroque organ is, reportedly, worth seeing’

Parenthetical omen is absent from central varieties. 3.4  Omen with quotative clauses Eastern varieties also show the evidential preceding root CP structures. Consider thus (21), from the fieldwork corpus.

 Ricardo Etxepare & Larraitz Uria

(21) Bere senarra ospitalean da, omen gaizki d-ela her husband hospital:loc is, evid ill is-comp ‘Her husband is in the hospital. They say that he is ill’

(21) is one of the possible answers to the translation task demanded to the speakers on the French equivalent of the English translation, with an overt verb of saying. The evidential particle omen in this case precedes what looks like a whole CP, headed by the declarative complementizer–ela ‘that’. Root -ela clauses in Basque have a quotative interpretation (see Aikhenvald 2004; Güldemann 2008; Etxepare 2010a). The presence of -ela in a root clause indicates that the clause is an instance of reported speech. Consider in this regard the literary corpus example gathered here, from Sallaberry (1978: 61): (22) Goiz batez mahastiko landan ginen lanean morning one:ins vineyard:gen field:loc we.were working Hil-zeinuak hasten dira ortzia bezala burrunban death tolls start:ipfv are thunder like sounding Eta ari, eta ari and on and on Landa batetik bertzerat oihuka hasi ziren gizonak: field one:abl other:all yelling start:ptcp were men “Norbait hil de-a? -Ez dugu aditu  someone dead is-q neg we.have hear:ptcp Gerla ditake” war can.be Eta tenore baten buruan: and moment one:gen head:loc “Gerla d-ela  war is-comp Irisarriko jendarmak hor dir-ela, Irisarri:gen policemen there are-comp berriaren jakinarazteko etorriak” news:gen communicate-nmlz-pur come:ptcp;pl ‘One morning we are working in the vineyards. Suddenly, the bells start tolling a death toll. They go on and on. The men start yelling at each other from one field to the other. “Is anyone dead? -We did not hear about it”. And after a moment: (lit.) “That it is war. That the policemen from the village of Irisarri have come to tell us the news” ’.

In (22), it is the presence of the complementizer that marks the reported status of the associated proposition. Unlike the reportative evidential however, the reported



Microsyntactic variation in the Basque hearsay evidential 

proposition in -ela clauses always has an identifiable source. Consider in this regard the contrast between (23) and (24): (23) Gizon bat tabernan sartu zen. Ardoa nahi zue-la man one pub:loc come.in:ptcp was wine want aux-comp ‘A man entered the pub. He said he wanted wine’ (24) Gizon bat tabernan sartu zen. Ardoa nahi omen zuen man one pub:loc come.in:ptcp was wine want evid aux ‘A man entered the pub. Reportedly, he wanted wine’

As shown in (23), the -ela-clause cannot refer to a general saying or rumour. It can only attribute the reported proposition to a previously established discourse referent (a man in (23), the people in the fields in (22)). The reportative or hearsay evidential on the other hand, does not specify the source, but refers either to a rumor or a non-identified third person, as shown in (24). Both the hearsay particle and the reportative -ela can be combined. Take the made-up -ela clause in (25a), which attributes the reported saying to the policeman of Irisarri. In the context of something like (25), omen preceding the CP contributes the reading that the information source of the quotative subject (the policeman) to whom the saying is attributed is an unidentified someone: (25) a. Irisarriko jendarma jin da. Gerla d-ela Irisarri:gen policeman come:ptcp is war is-comp ‘The policeman from Irisarri has come. He says that there is war’ b. Irisarriko jendarma jin da. Omen gerla d-ela Irisarri:gen policeman come:ptcp is evid war is-comp ‘The policeman from Irisarri has come. He says that reportedly there is war’

The presence of omen reflects that the underlying agent of the saying uses reported evidence to make his/her assertion. In other words, the agent of the saying and therefore the source of the reported speech is the policeman of Irisarri, and this source manipulates information of the reportative sort to make his claim. The reportative evidential is thus within the scope of the quotative clause. The scope relations are similar to what we find in the English sentence (26): (26) He said that reportedly, there was war

The relative scope of the evidential and the quotative complementizer -ela ­corresponds to what we would expect: the evidential is merged in a functional projection which is within the scope of the CP in Basque (27): (27) Esan didate [etorri omen d-ela] say:ptcp aux    come:ptcp evid is-comp ‘They told me that he reportedly came’

 Ricardo Etxepare & Larraitz Uria

Syntactically though, the evidential particle is at the left edge of the clause, and not in the position it occupies in ordinary embedded clauses. The evidential particle is not a parenthetical here, as it does not constitute an independent intonational phrase, and can occur immediately following an overt verb of saying, as in (28), from fieldwork data: (28) Senarra ospitalean du. Diote omen biziki gaizki d-ela husband hospital:loc has they.say evid very ill is-comp ‘His husband is in the hospital. They say he is reportedly very ill’

The presence of omen at the beginning of the embedded clause, immediately following the verb of saying suggests it occupies a position in the CP domain. Note that the evidential in this case does not contribute reportative evidence for the saying itself, but for the proposition that the person in the hospital is seriously ill. Following an analysis of Blaine and Dechaine for some varieties of Cree (2007), we suggest that in these cases omen occupies the specifier position in a functional projection of the expanded CP-domain (in the sense of Rizzi 1997), to which it has raised from a position under the scope of the complementizer. We will come back to this later on. (29) [FP Omen F0 [CP [IP…(omen)…] -ela]

Nothing like this is possible in central dialects, which nevertheless do possess a quotative strategy like the eastern one. The difference therefore lies in the morphosyntactic status of the evidential particle. 3.5  Synthesis To summarize, in eastern dialects the evidential particle seems to behave as an independent term in the sense that (i) it can rescue the auxiliary from first position; (ii) does not depend on the presence of a finite auxiliary; (iii) occupies more than one possible position in the clausal domain, and (iv) can occur as a parenthetical. None of those properties characterize the central homonymous e­ vidential, which (i) requires the presence of a finite auxiliary, (ii) does not rescue the finite form from first position; (iii) occupies a rigid position in the clausal structure and (iv) cannot occur as a parenthetical. 4.  Compiling the evidential features The four basic properties reviewed in Section 2 distinguishing central from eastern dialects all revolve around the more or less intimate or dependent relation between finiteness, as it is realized in the auxiliary, and the evidential. Thus, in



Microsyntactic variation in the Basque hearsay evidential 

central dialects omen is necessarily associated to the finite auxiliary, and seems to undergo the same first position restrictions operating in the case of the auxiliary. In eastern dialects, the evidential can occur in non-finite constituents and in quotative constructions preceding a fully projected CP domain, and it seems to be able to rescue the auxiliary from first position. A natural hypothesis to entertain in view of these distinctions is that the central evidential, unlike the eastern one, is somehow fused to the finite auxiliary, as part of the set of features which may maximally constitute it. Note however that this cannot be strictly right, since elision can affect the entire IP, auxiliary included, and leave the evidential in place, in all dialects (cf. the cases in (15a,b)). Some principled distinction must separate eastern and central omen, and the relevant factor cannot be a simple morphological extension of the auxiliary to include the evidential. As a preliminary step in the way of accounting for those differences let us point out that the evidential omen is one of the possible instantiations of a wider category that includes the conjectural evidential ote ‘I wonder’ (Eguzkitza 1991). Ote only occurs in interrogative clauses, where the use of omen is forbidden, and it is thus in complementary distribution with the latter. The addition of ote to a sentence has the effect of directing the question towards the speaker itself, and is thus akin in its meaning contribution to English ‘I wonder’: (30) a. Nor etorri da who come:ptcp is ‘Who came?’ b. Nor etorri ote da? who come:ptcp evid is ‘Who came, I wonder’

Both particles share the same rigid position in central dialects (immediately preceding the auxiliary) and the more flexible one in eastern ones (see Etxepare and Uria in preparation). Early generative approaches to Basque grammar have included those two particles under a same modal category (as Eguzkitza 1991; De Rijk 2008). The two evidential particles also share part of their phonological shape, the initial vowel o, which is followed by the rest of the evidential particle. One plausible approach to the bisyllabic evidential particles in Basque is that they are morphologically complex, and can be split into a shared morpheme o- and what I will call the evidential residue, -men in the case of the reportative evidential and -te in the case of the conjectural one. We can only speculate about the hypothetical status of o-. One remote possibility is that it corresponds originally to the proximate demonstrative -o (see Gomez and Sainz 1995), which occurs both as the agreement marker for third person indirect objects in the finite auxiliary, or as a suffixed proximate determiner in plural locatives, such as (31).

 Ricardo Etxepare & Larraitz Uria

(31) etxe-o-tan house-det-loc.sg ‘in the houses next to the speaker’

No clear etymological origin can be found either for -men or -te, although the latter is identical to the old future marker -te, still operative as such in the eastern Souletin dialect (see De Rijk 2008; and Oyharçabal 2003). Under this view, the evidential particles are morphosyntactically complex, and include at least two different syntactic features, one that we will call a deictic feature, expressing speaker dependency, and another one, encoding the paradigmatic evidential component, that we will call the evidential residue. In view of the differences between central and eastern dialects, we can describe the behaviour of the more flexible eastern dialects as one in which the sequence of morphosyntactic features that constitute the bimorphemic word omen behaves as an independent syntactic term. In central dialects, the way those features are arranged does not allow for such a free distribution. One straightforward syntactic formulation of this basic descriptive generalization is the following. Whereas the central evidential omen corresponds to a sequence of two independent functional heads o- and -men, which are recursively merged to the highest feature belonging in the finite auxiliary, the evidential omen in eastern varieties corresponds to a different syntactic arrangement: one in which the two features constitute an independent term, in the sense of Chomsky (1995). This term is merged to a silent evidential head, which in turn selects the IP. Let me represent the two options as follows: (32) a. …[DeicticP o- [Evidential Residue -men [AuxP da]]

(central)

[AuxP da]

(eastern)

b. [Evidential Phrase [Deictic Phrase o- [ -men]]

Ev0

(32) exploits a possibility in the context of the evidential particle that arises within the general question of what a word is. Assuming that morphologically complex words cannot constitute simple heads, it is legitimate, when a string of morphemes has the appearance of a word, to ask as Julien (2002:14) does, “which possibilities are there for the internal organization of that string”. In Julien’s view, which we share, words are perceived entities and not primitives of the system. If a given string is perceived as a word, it only means that the morphemes in question regularly appear adjacent to each other and in a certain order (2002: 36), and the reason they come out that way is because they are syntactically determined to appear that way. The apparent word status of omen and ote in the evidential constructions follows from the regular adjacency of the features involved, and the latter arguably follows as a matter of selection. Strict adjacency on the other hand will follow if the evidential residue does not attract elements to its outer edge. Let us conclude that the evidential residue does not constitute a Probe for Agree. If it is not a Probe, it



Microsyntactic variation in the Basque hearsay evidential 

will host no element in its Specifier, and the adjacency between the deictic feature and the evidential residue will never be broken. A minimal difference between eastern and central dialects lies in the selectional properties of the evidential residue, which directly selects the IP in central dialects, but selects nothing in eastern varieties, being part of a complex phrase which merges to an independent evidential head. From a structure like (32a) it follows immediately that parenthetical omen will not be possible in central dialects, nor a quotative omen, if the latter represents a case in which the evidential residue selects something other than the IP. (32b) takes the (eastern) evidential to be the external argument of a relational head which takes as its internal argument a propositional entity (the IP), and as another argument the evidential complex. No formal issue arises from the fact that the deictic feature and the evidential residue are merged as an independent constituent in one case, and as part of the clausal structure in the other. Since in (32b) the deictic feature and the evidential residue form a constituent, it is possible for them to occupy positions in the vicinity of the evidential head, or to be adjoined as a parenthetical along the clausal spine. We take omen in quotative clauses to be an instance of topicalization. The topic position which hosts omen precedes the overt complementizer in an expanded CP (see Rizzi 1997). The conclusion is inescapable under an antisymmetric view of the order IP-C (see Ormazabal, UribeEtxebarria and Uriagereka 2008, for arguments in favour of an antisymmetric view of IP-C orders in Basque): (33) a. [TopP omen Top0 [CP [IP …(omen)…]i C0ti]] b. Omen etorri d-ela evid come:ptcp is-comp ‘They say (unspecified source) that he/she has come’

In general, the structure in (32b) allows the evidential to migrate to other positions in the clause structure, modulo the relevant triggers for that. One such case is illustrated in Section 2. The evidential in some of the eastern varieties seems to rescue the clitic auxiliary from first position. We can take this to instantiate the movement of the evidential to the specifier of the Polarity Phrase, which licenses auxiliaries: (34) [Polarity Phrase Omen Pol0 [Evidential Phrase (omen) Evid0[ IP]]]

Note that polarity heads precede evidentials and auxiliaries in Basque: (35) [PolarityPhrase Ez/ba- [EvidentialPhraseomen [IP da [AspPetorri]]]]

Since omen corresponds to a sequence of heads in central dialects such a movement to the specifier of the Polarity Phrase is not possible, and the set of affixal heads in the beginning of the clause yields a PF-deviant object (see Etxepare forthcoming).

 Ricardo Etxepare & Larraitz Uria

We will end this section with further indirect evidence in favor of this particular formulation of the differences between eastern and central dialects. This indirect evidence is provided by the fact that only eastern dialects have developed derived adverbs, postpositional phrases and noun phrases from omen, keeping the evidential sense intact. The Basque General Dictionary gathers several lexical entries derived from omen, all of them belonging in the eastern area. They are typically of the form omen+postposition, or reduplicated forms from a basic omen whose meaning seems to be ‘rumor’: (36) a. Bertze omen-amen batzuk kurritu zuten other evid-redup some run had ‘Some other rumors kept running’ b. Omen-ka mintzo da evid-adv speak is ‘He/she speaks by hearsay’ c. Xuberoko jendea omen-ez leguna Soule:gen people evid-ins nice ‘People from Soule, reportedly nice’

Nothing of the sort in (36a-c) is attested in central dialects. Assuming that the derivation of complex lexical items like (36) requires the existence of a basic syntactic term, which is only available in eastern dialects, the lexical data come in support of our hypothesis. 5.  TP-internal and TP-external hearsay evidentials The structure in (32b) allows a richer set of options in the clausal distribution of the evidential particle. Unlike in (32a), the content of the evidential head is not specified in the lexicon, as it can host both the conjectural and the reportative evidential. The evidential features are embedded in the Specifier of the evidential head, and from there they determine its evidential content. This immediately suggests an agreement relation between an unvalued Evidential feature, residing in the evidential head, and the lexical evidential feature included in its Specifier: (37) [[EvidP Omen] Ev[uE] …

We can hypothesize that the evidential feature is valued under agreement with the evidential phrase. Technically, we do not want [uE], the feature content of the evidential head, to be an instance of an uninterpretable feature, but rather of an unvalued one whose value gets specified by its local relation to an evidential phrase, be it omen or ote. For this, we need a theory of Agree which distinguishes



Microsyntactic variation in the Basque hearsay evidential 

valuation from interpretability, unlike Chomsky’s Agree mechanism (2000, 2001). Following a recent proposal by Pesetsky and Torrego (2007), we take unvalued features to come in two guises: as interpretable ones or as non-interpretable ones. For an unvalued feature to arise, a paradigmatic contrast must be established between two values of a same attribute. This paradigmatic contrast is manifest in the dual realization of the evidential category as either a hearsay or a conjectural evidential. In (37), it is the evidential term omen, as opposed to ote, that values the unvalued evidential feature of the Evidential head. Since Agree is actually a local operation that does not force the Goal to be in the specifier of the Probe, but rather is taken to scan the c-command domain of the probe, in our case the evidential head, we can expect to find occurrences of omen that do not immediately precede the evidential head, but just stand in a local relation in the complement domain of the evidential probe. This prediction is borne out: a corpus based search on eastern varieties provides alternative positions for omen; they have the evidential following the finite auxiliary (38a,b), and sometimes, also the lexical verb (38c). Nothing of this sort is available to central speakers. (38) a. Eta denak loriatuak zauden omen leku goxo hartaz and all enchanted were evid place nice that:ins ‘And everyone was enchanted with that nice place’ b. Eüskaldüna zen lehenago omen popülü ez ikasia Basque was earlier evid folk neg educated ‘The Basque people were reportedly an uneducated folk earlier’ c. Kolegioko lagun bat bezala ezagutu ninduen omen college:gen friend one as know:ptcp aux evid ‘He reportedly recognized me as a college friend’ d. Eguneko berria da, auzapezak herriko dirua day:gen news is, major:erg village:gen money ebatsi luke omen steal:ptcp aux.irr evid ‘It’s the news of the day: the major has reportedly stolen the money of the village’ (Informant from Iholdi, Basse-Navarre)

Omen, in the literary corpus cases, is written without a comma, unlike typical parenthetical cases, and in the case of data coming from fieldwork, as (38d), we can confirm that it is produced without any intonational break (in the case in point, with a clear weakening of the first vowel of the evidential). In our database, examples such as (38) are possible in 16 out of 33 inquiry points, according to our acceptance tests. There is thus a substantial number of speakers for whom those orders are impossible. The eastern dialectal domain therefore splits in at least two grammars as far as the distribution of the evidential particle is concerned.

 Ricardo Etxepare & Larraitz Uria

The two basic varieties that we have discussed stand in an inclusion relation: let us call the most restrictive one, the one which does not allow hearsay evidentials in TP-internal positions, Eastern 1. We will call the variety which allows the hearsay evidential both in the TP-external position and in the TP-internal one, Eastern 2. Eastern 2 includes the positions in which the evidential can occur in Eastern 1, and adds the ones which correspond to the TP-internal position. Let us see how we can represent both varieties in a coherent way. 5.1  The restrictive variety: Eastern 1 The position of the hearsay evidential in Eastern 1 is plausibly analysed in terms of a criterial condition specifically involving an independent evidential head and a phrasal evidential particle. Thus, in Eastern 1, the evidential (outside parenthetical instances) only occurs in between the Polarity head and the auxiliary (39a) or in between the lexical verb and the auxiliary (39b): (39) a. [Polarity Phrase Neg Pol0 [Evidential Phrase omen Evid0 [IP Aux [AspP VP]]]] b. [Polarity Phrase [AspP VP] Pol0 [Evidential Phrase omen Evid0 [IP Aux …]]]]

From this position, the evidential can be targeted by higher positions, from either focus or Polarity for instance (40a), or from a Topic position higher than CP, in the quotative cases (40b): (40) a. [FocusP (omen) Foc0 [PolP (omen) Pol0 [EvidP (omen) Evid0 [IP Aux [AspP…] b. [TopicPhrase (omen) Top0 [CP [EvP (omen) Ev0 [IP etorri da] -ela…]]

Eastern 1 also allows parenthetical evidentials. Parenthetical omen, characterized by intonational autonomy, has a very free distribution: it can appear in any of the major constituency breaks in the clausal structure. On the other hand, it cannot occur in the canonical position for the evidential, between the Polarity head and the auxiliary, where intonational independence is impossible: (41) (Omen), etorri (,*omen,) da, (omen)    evid come:ptcp   evid is    evid ‘Reportedly, he/she came’

Furthermore, parentheticals can, however marginally, be combined with another hearsay evidential in its canonical position, yielding cases of more than one evidential particle. Our database attests to this possibility (42): (42) Bere osaba presondegian omen da his/her uncle prison:loc evid is ‘His/her uncle is in prison’ Omen, ez omen du deus aitortu evid neg evid has anything confess:ptcp ‘Reportedly, he/she didn’t confess anything’



Microsyntactic variation in the Basque hearsay evidential 

The special behavior of parenthetical evidentials in comparison to the criterial ones suggest that they relate to the clausal structure in a different way. We will claim that parentheticals relate to the rest of the structure via adjunction. For the sake of concreteness, we adopt the view according to which adjunction is merge without projection of a label (see Chametzky 2000; Hornstein 2009). We take the option of adjunction to be related to the lack of a projected evidential head. This hypothesis is intended as general for all eastern varieties. 5.2  The less restrictive variety: Eastern 2 Eastern 2 allows the possibility of placing the evidential in some position internal to the TP. This possibility is illustrated by the fact that the evidential may follow the auxiliary or even the lexical verb, as shown in (38a-d), as well as by the fact that for a subset of eastern speakers, the evidential can be part of non-finite structures headed by Aspect projections, which presumably, do not possess Tense. This is illustrated in (14a-b). For Eastern 2, which seems to allow a criterial position both TP-internally and TP-externally, we propose an analysis whereby the evidential features, like negation in other languages (see Zanuttini 1997; Rowlett 1998; Haddican 2004, for Basque), may be split in two heads: one in the vicinity of the VP, and another one in the outer edge of the TP: (43) [PolarP1 Pol0 [Evid P1 (omen) Evid0 [TP Aux…[Evid P2 (omen) Evid0 …]]]]

It is probably not a coincidence that the dialects which allow TP-internal evidentials also allow TP-internal negation, as in (44): (44) JON da ez erori Jon is neg fallen ‘It is Jon who did not fall’

In Eastern 2, the evidential can be merged in the lower Evidential projection, and relate to the higher head by Agree, with movement into the higher Evidential Phrase induced by an extra EPP feature. With those basic syntactic foundations, let us provide explicit derivations for the TP-internal word orders available in Eastern 2, repeated from (11). We add a schematic representation of the relative order of evidential, verb and auxiliary in each case: (45) a. Protestanteen komentu batean omen protestant:gen convent one:loc evid eskolarazi zuen provide.with.schooling aux ‘He/she got him/her educated in a protestant convent, they say’ [omen Verb Aux]

 Ricardo Etxepare & Larraitz Uria

b. Langonen zen omen bizi Hipokrataren alaba bakarra Langon:loc aux evid live Hypocrates:gen daughter only ‘It was in Langon that Hipocrates’ only daughter reportedly lived’ [Aux omen V…] c.

Kolegioko lagun bat bezala ezagutu ninduen omen college-gen friend one as know aux evid ‘He reportedly recognized me as a college friend’ [Verb Aux… omen]

(45a) is accounted for by the movement of a lower Polarity Phrase including the VP to the specifier of the higher polarity phrase, an instance of predicate fronting well established for Basque (see Haddican 2004, 2008). The movement carries along omen in the specifier of the low evidential projection: (46) …[PolP1 [PolP2 Ø [EvidP omen Evid0 [VP eskolarazi]]i Pol0 [AUXP zuen ti ]]]

In the second case (41b), omen is in-situ, in the specifier of the low evidential phrase, following the auxiliary. A focal phrase rescues the auxiliary from first position: (47) [FocP Langonen Foc0 [TP zen…[EvidP omen Evid0 [VP …bizi ]]]]

The last case also has omen in-situ, and it is the VP which raises to the higher Focus Phrase, leaving the evidential behind. In this case, predicate fronting targets the verbal phrase: (48) [FocP [VP DP V] Foc0 [PolP Pol0 [TP ninduen…[EvidP omen Evid0 …]]]]

6.  Summary Let us summarize the different syntactic representations of the hearsay evidential in the Basque dialectal continuum. In central varieties, we reached the conclusion that the hearsay evidential lexicalizes a sequence of evidential features which merge independently within the clausal spine: (49) …[Deictic Phrase o [Evidential Residue -men [AuxP … ]]

(Central)

In eastern dialects, we have distinguished two large linguistic varieties, named Eastern 1 and Eastern 2. In Eastern 1, the evidential particle occupies a criterial position which belongs in the left periphery of the sentence, just above epistemic modals and Tense, à la Cinque: (50) …[EvidP omen]] Ev[uEv] [ModalP/TP … ]

(Eastern 1 and Eastern 2)



Microsyntactic variation in the Basque hearsay evidential 

In the less restricted eastern variety we have called Eastern 2, the evidential particle can merge in two different positions, which we took to be criterial, one in the left periphery of the clause, and another one in the periphery of the verbal phrase. This configuration is not unlike what one finds in split negation, including negation in Basque. The relevant configurations can be represented as in (51). Crucially, the evidential particle can Agree with both the inner evidential head and the higher one, with Agree understood as Feature Sharing, as in Pesetsky and Torrego (2007): (51) …[ (omen Ev[uEv] [AuxP AUX…[ omen Ev[uEv] [vP …]]]]  (Eastern2)   \_____/ \_______________/\____/  (Agree)

A last structural option is represented by parenthetical omen, which merges to the structure by adjunction. We will assume parenthetical adjunction to occur at the major constituency joints of the clause: CP and VP. We assume that in this case no Evidential Head is projected. As we expect, this option is only available to eastern varieties, in which the deictic feature and the evidential residue form a syntactic term, independently of the clausal configuration. For a cartographic approach to the functional structure such as Cinque’s, the Basque dialectal data provide conflicting results. On the one hand, it presents evidence that varying surface syntactic distribution does not constitute in and of itself evidence against a rigid arrangement of functional heads, as Cinque wants. On the other hand, the distribution of the hearsay evidential points to the existence of more than one syntactic position for evidentials of this type, at least in a subset of the eastern varieties. The kind of evidential spreading that we observed in Eastern 2 is reminiscent of the kind of split categorical representation that other functional items, such as negation, seem to show (see Cinque 2006, for the special status of negation in the cartography of the clause). Haddican (2004) has given concrete arguments in favour of this configuration in Basque, which is also confirmed by dialectal data (cf. (44) in this paper). In pushing this parallelism further, we note also that focused items in Basque, if Etxepare and Uribe-Etxebarria (2008) are right (see also Belletti 2005), may occupy more than one position in the Basque functional sequence, either a position in the left periphery of the clause or one at the vicinity of the vP. It may be that a subclass of the functional projections, perhaps on account of their light selectional restrictions, has syncategorematic properties, being able to occur at the edge of different cyclic domains. The validity of such a conclusion will largely depend on whether the different positions occupied by, say, reportative evidentials, contribute exactly the same meaning in each case. This seems to be largely true, but a systematic contrastive analysis of the dialectal differences regarding the reportative evidential has not yet been seriously attempted (see Etxepare 2010b for a first approach).

 Ricardo Etxepare & Larraitz Uria

References Aikhenvald. Alexandra. 2004. Evidentiality. Oxford University Press. Belletti, Andrea. 2005. “Extended doubling and the VP periphery.” Probus 17–1: 1–35.

doi: 10.1515/prbs.2005.17.1.1

Blaine, Eleanor M. and Dechaine, Rose-Marie. 2007. “Evidential types: evidence from Cree ­dialects.” International Journal of American Linguistics 73–3: 257–291.  doi: 10.1086/521728 Chametzky, Robert A. 2000. Phrase Structure: From GB to Minimalism. Oxford: Blackwell. Chomsky, Noam. 1995. The Minimalist Program. Cambridge: MIT Press. Cinque, Guillermo F. 1999. Adverbs and Functional Heads. A Crosslinguistic Perspective [Oxford Studies in Comparative Syntax]. Oxford: Oxford University Press. Cinque, Guillermo F. 2006. Restructuring and Functional Heads. Oxford: Oxford University Press. De Rijk, Rudolf. 2008. Basque Grammar. Cambridge, MA: MIT Press. Dendale, Patrick and Collier, Danielle. 2003. “Point de vue et évidentialité.” Cahiers de ­Praxématique 41: 105–130. Ducrot, Oswald. 1984. Le dire et le dit. Paris: Les Editions de Minuit. Eguzkitza, Andolin. 1991. “Partikula modalez ohar pare bat.” ASJU 25 (3): 961–963. Etxepare, Ricardo. 2010a. “From hearsay evidentiality to samesaying relations.” Lingua 120: 604–627.  doi: 10.1016/j.lingua.2008.07.009 Etxepare, Ricardo. 2010b. “Omen bariazioan.” In Euskara eta euskarak: aldakortasun sintaktikoa aztergai [Supplements ASJU 52], Fernandez, Beatriz, Pablo Albizu and Ricardo Etxepare (eds), 85–112. Bilbo: UPV/EHU. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2008. “On negation and focus in Spanish and Basque.” In Gramatika Jaietan. Patxi G ­ oenagaren omenez. Bilbao: University of the Basque Country. Supplements of the International Journal of Basque Language and Linguistics, Xabier Artiagoitia et Joseba Lakarra (eds), 51. 287–310. ISBN: 9788498600858. Etxepare, Ricardo and Uribe-Etxebarria, Myriam. 2009. “Hitz hurrenkera eta birregituraketa euskaraz.” In Beñat Oihartzabali gorazarre. Festschrift for Bernard Oyharçabal [Supplements of ASJU 43], Ricardo Etxepare, Ricardo Gómez and Joseba A. Lakarra (eds), 335–356. Bilbo: UPV/EHU. Gómez, Ricardo and Koldo Sainz. 1995. “On the origin of the finite forms of the Basque verb.” In Towards a History of the Basque Language, José Ignacio Hualde, Joseba A. Lakarra and Robert L. Trask (eds), 235–273. Amsterdam: John Benjamins.  doi: 10.1075/cilt.131.10gom Giorgi, Alessandra. 2010. About the Speaker. Oxford: Oxford University Press. Güldemann, Tom. 2008. Quotative Indexes in African Languages. A Synchronic and Diachronic Survey. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110211450 Haddican, Bill. 2004. “Sentence polarity and word order in Basque.” The Linguistic Review 21 (2): 87–127.  doi: 10.1515/tlir.2004.004 Haddican, Bill. 2008. “Euskal Perpausaren Espez-Buru-Osagarri hurrenkeraren aldeko a­ rgudio batzuk.” In Antisimetriaren hipotesia vs. Buru Parametroa. Euskararen hitz hurrenkera ezbaian, Iñigo Arteatx, Xabier Artiagoitia and Arantzazu Elordieta (eds), 69–96. Bilbo: UPV/EHU. Hornstein, Norbert. 2009. A Theory of Syntax. Minimal Operations and Universal Grammar. Cambridge: Cambridge University Press.



Microsyntactic variation in the Basque hearsay evidential 

Jacobsen, William H. 1986. “The heterogeneity of evidentials in Makah.” In Evidentiality: The Linguistic Coding of Epistemology, Wallace L. Chafe and Johanna Nichols (eds), 3–28. ­Norwood, NJ: Ablex. Julien, Marit. 2002. Syntactic Heads and Word Formation. Oxford: Oxford University Press. Laka, Itziar. 1990. Negation in Syntax. Ph.D. dissertation, MIT. Ormazabal, Javier, Uribe-Etxebarria, Myriam and Uriagereka, Juan. 2008. “Hitz hurrenkera eta NZ-mugida: azalpen parametriko baten bila.” In Antisimetriaren hipotesia vs buru parametroa: euskararen oinarrizko hitz hurrenkera ezbaian. Iñigo Arteatx, Xabier Artiagoitia and Arantzazu Elordieta (eds), 25–50. Bilbo: UPV/EHU. Oyharçabal, Bernard. 2003. “Tense, aspect and mood.” In A Grammar of Basque, José Ignacio Hualde and Jon Ortiz de Urbina (eds), 249–282. Berlin: Mouton. Ortiz de Urbina, Jon. 1994. “Verb-initial patterns in Basque and Breton.” Lingua 94: 125–153.  doi: 10.1016/0024-3841(94)90023-X Ortiz de Urbina, Jon. 1995. “Residual verb second and verb first in Basque.” In Discourse Configurational Languages, K. Kiss (eds), 99–121. Oxford: Oxford University Press. Pesetsky, David and Esther Torrego. 2007. “The syntax of valuation and the interpretability of features.” In Simin Karimi, Vida Samiian and Wendy K. Wilkins (eds) Phrasal and clausal architecture: Syntactic Derivation and Interpretation. In honor of Joseph Emonds. ­Amsterdam: John Benjamins.  doi: 10.1075/la.101.14pes Rizzi, Luigi. 1997. “The Fine Structure of the Left Periphery”, In Elements of Grammar, L. ­Haegeman (ed.), Kluwer, Dordrecht. Rowlett, Paul. 1998. Sentential Negation in French. Oxford: Oxford University Press. Schenner, Mathias. 2008. “Semantics of evidentials: German reportative modals.” Proceedings of ConSOLE XVI, 179–198. Squartini, Mario. 2004. “Disentangling evidentiality and epistemic modality in Romance.” ­Lingua 114–7: 403–444. Uria, Larraitz and Etxepare, Ricardo. 2011. “Basyque: aplicación para el estudio de la variación sintáctica.” Linguamática 3–1. Wiemer, Björn. 2010. “Hearsay in European languages. Toward an integrative account of lexical and grammatical marking.” In Linguistic Realization of Evidentiality in European languages, Gabriele Diewald and Elena Smirnova (eds), 49–59. Berlin: Mouton de Gruyter. Zanuttini, Raffaella. 1997. Negation and Clausal Structure. New York: Oxford University Press. Zubeldia, Larraitz. 2010. “Omen partikularen analisi semantiko eta pragmatikoa.” Ph.D. ­dissertation, UPV/EHU.

Consulted corpora Basyque (Basic Syntactic Atlas of the Basque Language). Online at http://ixa2.si.ehu.es/atlas2/ index.php?lang=en. Ereduzko Prosa Gaur [Contemporary Reference Prose]. Online at: http://www.ehu.eus/euskaraorria/euskara/ereduzkoa/laguntza.html. Etchepare, Jean. 1905. Jean Etchepare mirikuaren idazlanak, III. Kazetaritza (A), 1903–1915. Etxamendi, Eñaut. 2011. Idazlanen bilduma, 4. Baiona: Maiatz.

 Ricardo Etxepare & Larraitz Uria Orotariko Euskal Hiztegia [General Basque Dictionary]. 1987–2005. Luis Michelena and Ibon Sarasola. Bilbo: Euskaltzaindia, Desclée de Brower, Mensajero. Online at: www.euskaltzaindia.eus. Sallaberry, Etienne. 1978. Ene sinestea. Zarautz: Itxaropena. XX. mendeko Euskararen Corpus Estatistikoa [Statistical Corpus of 20th century Basque]. Online at http://xxmendea.euskaltzaindia.eus/Corpus/.

Index of Basque varieties mentioned in the text Many Basque place names possess two forms, the Romanized/Romance one, written with a Spanish or French orthography (often the only official designation until the end of the 20th century and even up to the present in some areas), and another one closer to the Basque pronunciation, current or historical. The latter one has become official in the Basque Autonomous Community and in parts of Navarre, and is written following Standard Basque orthographic conventions. In the following index, the form used by the author(s) who cite the varitey is registered; thus, the first entry in this index is given as Aezkoa, with Basque orthography, rather than Aezcoa, with Spanish orthography. Where both forms are used by different authors, page numbers are given under the most often quoted form, and the least quoted name is included without page numbers, referring the reader to the most often quoted version of the name. Occasionally, parentheses will be used to collapse the two forms into a single entry. The names of broader dialects, including traditionally recognized varieties, have been entered following similar criteria: the authors’ choices have been respected, and page numbers are given after the most often quoted form. The labels northern, western, eastern and southern correspond to macrodialectal areas and are given in the body of the book in capital letters when they correspond to dialects identified with that name within specific dialectological analyses. For this index, both uses of the term have been fused into a single entry. See Hualde (this volume) for an overview of Basque dialects. A Aezkoa  108, 135 Ahetze  135, 165 Aincioa  166 Ajánguiz  163–164 Alavese  27–28 Albóniga  175, 177–178, 182 Amaiur  135, 165 Antzuola  109, 135, 165, 175–185 Araitz  93, 104, 107, 108, 122, 123, 125, 130 Araitz-Betelu  93, 104, 107, 122, 123, 130 Arrasate  109, 135, 165 Arratia  105, 165, 223 Azpilkueta  107

B Baigorri  54 Bakio  165 Barrika  165 Basaburua Nagusia  109 Basaburua Ttikia  108 Basauri  13, 101, 106, 109, 117, 134, 135, 156, 165 Baztan  108, 135 Bergara  109, 136, 165 Bermeo  11, 109, 136, 165, 167, 169, 171–175, 178 Berriatua  159, 160 Bertizarana  108 Beskoitze  136 Betelu  123, 125 Biscayan, see below

Bizkaian  16, 17, 19, 21, 22, 23, 24, 25, 26, 27, 28, 33, 132, 142 Bortzerriak  108, 135 Butron  176 C Central  19, 23, 27, 109, 115, 165, 178, 223, 232, 233, 238, 266, 270, 271, 272, 276, 277, 278, 279, 280, 284 Ciordia  166 D Deba  136 Dima  93, 100, 105, 107, 108, 110, 116, 117, 118, 122, 125, 126

 Microparameters in the Grammar of Basque E Eastern  19, 20, 23, 25, 27, 58, 73, 108, 115, 150, 232, 234, 266, 267, 268, 269, 270, 272, 276, 277, 278, 280, 281 Eibar  109, 136, 165 Elgoibar  104, 105, 109, 110, 111, 112, 136 Ermua  109, 136, 165 Ermua/Eitza  109 Errenteria  108 Erro(ibar)  105, 108, 136, 166 Erronkari, see Roncalese Esteribar  105, 108 F Foru(a)  109, 136, 165, 169 G Gatika  176 Gernika  163 Gipuzkoan  16–25, 29–30, 33, 161, 164 Goierri  109 H Hernani  54 High Navarrese  16, 18, 20–21, 24–30, 166, 243 Hondarribia  106–109, 136 I Igorre  109, 136 Imo(t)z  109, 130, 136 Irun  35, 107–109, 136 Iruri (Zuberoa)  154 Itsasondo  116–117 L Labourdin, see below Lapourdian, see below Lapurdian  9, 16, 17, 20, 25, 28–29, 33, 142, 155, 243, 252, 254, 261–262

Larrau  109 Larresoro  165 Lasarte-Oria  109, 136 Laudio  136 Lazkao  136, 138 Leioa  109, 136, 165 Leitza  54 Lekeitio  35, 93, 99–100, 103, 104–112, 122, 136, 165, 262 Loizu  166 Low Navarrese  16–18, 20, 25, 29, 33 M Malerreka  108 Mallabi(a)  136, 165, 222–223 Markina  121, 136, 165 Marquina, see above Maruri  165, 176 Maya, see Amaiur Milafranga  121 Mondragón  165 N Navárniz  163 Navarrese  17–18, 105, 108, 142 Navarro-Labourdin [Lapurdian]  10, 243, 253–256 Northern  19, 73–74, 81–82, 253, 258, 261 O Oiartzun  109, 136, 165, 236, 238–239 Olazagutia  166 Olazar  175 Ondarroa  136, 144, 147, 157, 159, 165, 170 Oñati  121, 136, 165 Ordizia  109, 137, 165 Orio  109, 137, 165 Otxandio  109, 138, 165

P Pasaia  106, 109, 138, 165 Pasaia Donibane  180 Plentzia  176 R Roncalese  16, 29, 30, 136 S Sopela  109, 138, 165 Souletin  9, 16–20, 24–29, 33, 49, 195–217, 278 Southern  81, 253, 255, 260 Standard Basque  9, 16, 18–19, 23–30, 33, 63, 70, 86, 99–100, 102, 110, 116, 141, 195–199, 200, 201, 206, 207–214, 242 Sunbila  108 T Tolosa  104, 106, 109, 120, 136, 157 U Ugao  138 Ultzama  105, 108, 138, 165 Urduliz  165, 222–223, 240 W Western  19, 22–29, 70, 73, 77, 80, 87, 109, 115, 140, 150, 155–158, 165, 166, 170, 178, 185, 223, 231–233, 238 Z Zaldibia  138, 158, 165 Zamudio  138, 165, 170, 178 Zeberio  138, 156, 164–167, 178, 181–182 Zegama  109, 138, 165 Zuberoan, see Souletin

Name Index A Abels, Klaus  226, 239 Abney, Steven  129 Aboh, Enoch  226, 239 Agirre, Julen  93 Agirretxe, Joxe Luix  138, 165, 180, 185 Aguilar-Guevara, Ana  216 Aikhenvald. Alexandra  267, 274, 286 Aissen, Judith  8, 11, 75, 89, 95, 103, 129 Alberdi, Xabier  22, 34, 42, 62 Albizu, Pablo  5, 11, 12, 45, 55, 62, 64, 82–83, 89–91, 93, 99, 112–113, 120–121, 123, 125, 127, 129–132, 134, 139, 140, 144, 150–156, 168, 185, 187, 191, 240, 286 Albright, Adam  161, 185 Alcázar, Asier  45, 62 Aldai, Gontzal  40, 62 Aldamizetxebarria, Ensunza  19, 34 Aldridge, Edith  7, 11, 61–62 Alexiadou, Artemis  62, 64, 213, 240 Allières, Jacques  32, 34 Altube, Severo  6, 11, 223, 225, 228, 231, 239, 244, 261 Altuna, Patxi  196 Amadas, Laia  53, 64 Anagnostopoulou, Elena  121, 130 Anand, Pranav  149, 185 Angos, Alberto  136 Apalauza, Amaia  130, 136 Aramaio, Itziar  159, 160, 185 Aranberri, Fernando  136 Arejita, Adolfo  223, 226, 231, 239 Ariztimuño, Borja  140, 186 Arraztio, Kontxi  93, 104, 107, 122, 125 Arregi, Karlos  25, 34, 78, 89, 110, 119, 126, 127, 130, 136, 144–147, 149, 154, 157–158,

160, 166, 168, 170–178, 181, 183, 185, 245, 247, 260–261 Arregui, Anna  185 Arretxe  101–102, 118, 135, 156, 165, 185 Arteatx, Iñigo  6, 11, 113, 130, 152, 286, 287 Artiagoitia, Xabier  11, 13, 20, 34, 115, 126, 129, 130, 134, 144, 147–148, 155, 185, 186, 195–199, 212–214, 241, 244–245, 261, 263, 286–287 Artola, Fernando  130 Aurrekoetxea, Gotzon  102, 130, 221, 223, 226, 239 Austin, Jane  97, 101–102, 104, 130 Ayerbe, Mikel  39 Azkarate, Miren  196 Azkue, Resurrección María de  16, 34, 155–156, 165, 186, 196, 223, 240 Azurmendi, Joxe Migel  138 Azurza, María Eugenia  136 B Baauw, Sergio  147, 186 Baerman, Matthew  161, 186 Baker, Mark  3–4, 11, 155, 186 Barker, Chris  146, 186 Beavers, John  130 Bechert, Johannes  196, 214 Béjar, Susana  120, 130, 146, 186 Belletti, Andrea  285–286 Bereziartu, Juan Inazio  136 Berro, Ane  7, 11, 22, 34, 39–40, 44–45, 49, 51, 53, 60–63, 93 Bhatt, Rajesh  97–98, 104, 111–112, 127, 130 Bilbao, Leire  144 Blaine, Eleanor M  276, 286 Blake, Barry  67, 69, 73, 89 Blanche-Benveniste, Claire  143, 148, 186 Bleam, Tonia  98, 130 Blume, Kerstin  8, 11, 67, 69, 71–74, 84, 89

Boas, Franz  177, 186 Bobaljik, Jonathan  5, 11–12, 40, 63–64, 68, 89–90, 176, 186, 216 Boeckx, Cedric  4, 10–11 Boeder, Winfried  91 Bonaparte, Louis Lucien  3, 7, 16, 18- 19, 25, 33–34, 100, 102, 130 Bonet, Eulàlia  5, 11, 119, 123, 128, 154, 176, 183, 186 Bordelois, Ivonne  215 Borer, Hagit  4, 6, 11 Borik, Olga  199 Bošković, Zeljko  221–222, 230, 240, 253, 261 Bossong, Georg  75, 89, 95, 103, 130 Branigan, Phillip  176, 186 Bresnan, Joan  161, 186 Brewer, Mary A  174, 186 Broadbent, Judith M  161, 186 Brouchard, Denis  196, 212, 214 Browczyk, Margaret  191 Brown, Dunstan  161, 186 Burguete, Xabier  138 Büring, Danel  143, 186 Burzio, Luigi  113, 130 Butt, Miriam  131 Bybee, Joan  107, 131, 174, 186 C Cabodevilla, Josu  135, 165, 186 Camino, Iñaki  20, 34, 135, 195–196, 213–214 Cardinaletti, Anna  147, 186 Carlier, Anne  196, 217 Carlson, Greg  199, 214, 216 Carnie, Andrew  112, 133, 241 Caro Baroja, Julio  31, 34 Casenave-Harigile, Junes  196, 181, 214 Chametzky, Robert A.  283, 286 Chierchia, Gennaro  199, 206, 214–215 Chomsky, Noam  3–5, 11, 112, 119, 127–128, 131, 140, 184, 186, 229, 240, 278, 281, 286

 Microparameters in the Grammar of Basque Cinque, Guglielmo  12, 170, 186, 237, 240, 267, 269, 284–286 Collier, Danielle  286 Collins, Chris  5, 11, 216 Comrie, Bernard  76, 87, 89–90 Contreras, Heles  197, 202, 215 Corbett, Greville  161, 186, 206, 214–215 Couquaux, Daniel  113, 131 Coyos, Battittu  196, 200–201, 215 Croft, William  11 Cyrino, Sonia  197, 202, 215 Cysouw, Michael  144–145, 147–148, 186 D Dayal  199, 206, 213–215 de Mulder, Walter  196, 217 de Rijk, Rudolf P. G.  4, 11, 48, 50, 58–59, 63, 87, 92, 113–115, 129, 152, 161, 191, 214, 223, 241–242, 244, 263, 272, 277–278, 286 de Swart, Henriette  208, 217 Dechaine, Rose-Marie  276, 286 Delfitto, Denis  147, 186, 211, 213, 215 Demirdache, Hamida  53, 63 den Dikken, Marcel  146, 148, 150, 186 Dendale, Patrick  268, 286 Deustuko Hizkuntzalaritza Mintegia  87, 89 Diesing, Molly  96, 98, 112, 127, 131, 186 Dixon, Robert M.W  9, 186, 148 Dobrovie-Sorin, Carmen  95, 196, 212, 215 Doetjes, Jenny  213, 215 Dowty, D. Robert  45, 63 Ducrot, Oswald  268, 286 Duguine, Maia  52, 63, 91, 111, 131, 144, 221, 253–254, 256–257, 261, 286 Duhau, Henri  136 E Egaña, Aitzane  136, 158, 166–168, 187 Eguren, Luis  115, 131, 196, 199, 215 Eguzkitza, Andolin  62, 244, 262, 277, 286

Elexpuru, Juan Martin  136 Elgoibar, Esther  221 Elordieta, Arantzazu  9, 11, 23, 26, 34- 35, 93, 99–100, 103, 105, 112, 120, 131, 136, 144, 152, 188, 221, 224–230, 237–238, 240, 244–245, 252, 255, 262, 286–287 Elordieta, Gorka  26, 35, 99, 103, 105, 136, 144, 188, 221, 244, 255, 262 Elortza, Jerardo  135 Embick, David  128, 131, 161, 174, 176, 184, 187 Emonds, Joseph E  150, 187 Enç, Mürvet  95, 131 Epelde, Irantzu  35, 165, 187 Epstein, Richard  196, 215 Erdozia, Jose Luis  93, 138, 160, 244, 262 Espinal, M. Teresa  197, 199, 202, 205, 215 Etxabe, Karmele  137–138, 158, 165, 187 Etxebarne, Jüje  196, 201, 215 Etxebarria, Juan Manuel  138, 156, 164- 165, 167, 181, 187 Etxebarria, Toribio  136 Etxeberria, Urtzi  93, 117, 160, 196, 207, 211 Etxegorri, Philippe  49, 63, 196, 216 Etxepare, Ricardo  5, 7, 10–11, 13, 20–21, 23, 34–35, 39–40, 49, 56, 59, 62–63, 88–90, 92–93, 99, 103, 112–115, 117, 119–120, 125, 127, 130–132, 134, 144–145, 150–152, 154–155, 163, 187, 191, 207, 214–215, 217, 221–223, 227–228, 234, 240, 244–245, 256, 258–259, 262, 265–266, 268, 272, 274, 277, 279, 285–287 Euskaltzaindia  23, 43, 61, 89, 128, 130, 133–134, 194, 214, 221, 223, 226, 238, 284, 286 Ezeizabarrena, Mari Jose  101, 132, 262

Fernández, Beatriz  3, 5, 7–8, 11–13, 25- 26, 34–35, 39–40, 50, 62–64, 67, 70, 72, 76, 78, 84, 88–112, 117–122, 129- 134, 139, 144, 150–152, 154, 156, 158, 165–166, 173, 181, 187, 191, 195, 221, 240, 243 Fernández Landaluce, Javier  144–145, 147–148, 186 Fernández-Ordóñez, Inés  72–74, 80, 90, 93, 96, 102 Fischer, Olga  127, 132 Fodor, Janet Dean  140, 188 Fraile Ugalde, Ainhoa  136 Fraile Ugalde, Idoia  136 Franco, Jon  145, 188, 263

F Fábregas, Antonio  46–47, 55, 63 Fara, Delia Graff  199, 216 Felser, Claudia  153, 187

H Haas, Josef  188 Haddican, Bill  9, 23, 34, 56, 63, 127, 221, 227, 229–230, 236, 238, 240, 244, 262, 283–286

G Galarraga, Aritz  144 Gallego, Ángel  150, 188 Gaminde, Iñaki  117, 132, 136, 138, 165, 167, 169, 170–172, 188–189, 221 Garai, Andoni  93 Garai, Jesus M.  135 Garai, Oier  93 Garmendia, Larraitz  137–138, 158, 165, 187 Garro, Eneritz  135 Giannakidou, Anastasia  210, 214, 216 Gilisasti, Iñaki  222–223, 240 Giorgi, Alessandra  267, 286 Goenaga, Patxi  129, 133, 196, 214, 216, 241, 263 Gómez, Ricardo  31, 34–35, 90, 124, 132, 134, 262, 277, 286 Gorman, Kyle  191 Gorrochategui, Joaquín  35, 62 Gould, Stephen Jay  10 Grashchenkov, Pavel  62, 64 Green, Georgia  140, 188 Greenberg, Joseph H  6, 12, 210, 213, 216 Güldemann, Tom  274, 286 Gutiérrez-Rexach, Javier  107, 132

Haider, Hubert  78, 90 Haig, Geoffrey  103, 124, 132 Hale, Ken  4, 5, 11–12, 40, 55, 61, 63, 72, 90, 131 Hamann, Cornelia  147, 188 Harbour, David Adger  224, 238 Harley, Heidi  184, 188 Haspelmath, Martin  155, 188 Heap, David  188 Heath, Jeffrey  20, 34, 174, 176–177, 188 Heger, M. Klaus  145, 188 Heim, Irene  146, 149, 188 Hentschel, Gerd  91 Higginbothan, James  47, 63 Himmelmann, Nikolaus  78, 92, 132, 196, 213, 216 Holmberg, Anders  3, 7, 12–13, 40, 127–128, 133, 229, 240 Holmer, Arthur  7, 12, 40, 63 Hornstein, Norbert  283, 286 Hualde, José Ignacio  5, 7, 11–12, 15, 25–30, 34–35, 56, 58, 63–64, 91, 99, 103, 105, 115, 130–136, 140–142, 163, 165, 174, 181, 188, 217, 223, 240–241, 255, 262–263, 286–287 Hudson, Richard  161, 188 Hurtado, Irene  104, 136 Ibarra, Orreaga  93, 105, 136, 138, 165, 188 I Iglesias, Aitor  93, 105, 136 Inoue, Atsu  140, 188 Iribertegi, Paula  163 Irigoien, Alfonso  196, 211, 216 Irurtzun, Aritz  9, 144, 160, 225, 237, 240, 243–244, 247, 252–254, 256–257, 260–262, 286 Iturain, Iñaki  137 J Jacobsen, William H.  267, 287 Jauregi, Oroitz  28, 35 Jelinek, Eloise  96, 98, 112, 127, 131, 133 Jespersen, Otto  206, 214, 216 Jo, Jung Min  226, 241 Johns, Alana  62, 64, 186

Name Index  Jouitteau, Mélanie  139, 143, 148, 191, 221, 224, 226–229, 240, 241 Julien, Marit  229, 241, 278, 287 K Kandybowicz, Jason  226, 241 Kayne, Richard  3, 6–7, 10, 12, 39, 50–51, 54, 64, 85, 93, 114, 143, 150, 188, 221 Keenan, Edward L.  87, 90, 215 Keine, Stefan  128, 133 Keyser, Samuel  11, 40, 55, 63, 72, 90, 131 Kiss, Katalin É  4, 12, 244, 262–263 Kiss, Tibor  240 Kleiber, Georges  199, 216 Koopman, Hilda  226, 241 Koopman, Willem  132 Kornfilt, Jaklin  95, 133 Krifka, Manfred  199, 213, 216 L Lafitte, Piarres  35, 156, 162, 188, 256, 262 Lafon, René  196, 216 Laka, Itziar  5, 7, 12, 33, 35, 40–41, 50–53, 57, 62, 64, 68, 90, 99, 120, 127, 130, 133, 136, 140, 145, 154, 173, 181, 188, 195, 198, 216, 227, 231, 241, 245, 262, 270, 287 Laka, Enara  167, 169, 171, 189 Lakaba, Ana  136 Lakar, Maite  135 Lakarra, Joseba  13, 31, 34–35, 62, 64, 90, 129, 132, 134–135, 214, 216–217, 241, 263, 286 Lambrecht, Knud  145, 189 Lambton, Ann K.S.  95, 133 Lamuela, Xavier  211 Landa, Alazne  96, 133, 263 Landa, Josu  68, 88, 90, 120, 132, 156, 274 Landau, Ivan  230, 241, 226 Lapesa, Rafael  196, 216 Larrañaga, Jone  135, 178, 182, 184, 189 Larson, Richard  5 Lazard, Gilbert  75, 90, 95, 133 Le Bruyn, Bert  216

Legarra, Hiart  39, 165, 167, 169, 171, 188 Legate, Judith  7, 12, 61, 64 Leizarraga  15, 35 Lersundi, Mikel  93, 138, 180, 185 Leu, Tomas  229, 241 Levin, Beth  4, 12, 68, 90 Lipták, Anikó  146, 148, 186 Loida, Loren  137 Longobardi, Giuseppe  197, 202, 216 López-Cortina, Jorge  260, 262 M Mahajan, Anoop  62, 64 Maienborn, Claudia  47, 64 Makazaga, Jesus Mari  132, 136, 93 Maling, Joan  85, 90 Manfredi, Victor  226, 238, 241 Manterola, Julen  31, 35, 160, 195–196, 208–209, 211, 213, 217 Marantz, Alec  161, 187 Marín, Rafael  46–47, 55, 63 Markaida, Bene  138 Markaida, Eukene  138 Markman, Vita  62, 64 Martins, Ana  226, 241 Masica, Colin  133 Massam, Diane  64, 186 Mateu, Jaume  64 Mathieu, Eric  253, 262 McCawley, James D.  140, 149, 189 McFadden, Thomas  8, 12, 67, 69, 71, 78, 82, 85, 87, 90, 110, 133 McNally, Louise  205, 215 Mendikoetxea, Amaya  11, 80, 85, 90, 97, 129, 133, 168, 185, 189 Mestres-Missé, Anna  262 Michelena, Luis  9, 12, 16, 31–32, 35, 62, 68, 75, 90–91, 162, 195–196, 213, 216–217, 244, 263, 288 Miller, Philip H.  145, 189 Mohanan, Tara  95, 97, 103–104, 133 Molina-Azaola, Gainko  78, 89, 110, 126, 130 Morgan, Jerry L.  140, 189

 Microparameters in the Grammar of Basque Morin, Yves-Charles  143, 148–149, 189 Moulton, Keir  114, 119, 315 Mounole, Céline  35, 70, 75–76, 90, 93- 94, 97, 103–106, 133, 189 Mugarza, Pello  136 Müller, Stefan  79, 90 Munn, Alan  197, 217 N N’Diaye, Genevieve  135, 165, 189 Ndayiragije, Juvenal  186 Neeleman, Ad  237, 242 Nevins, Andrew  25, 34, 119, 127, 130, 136, 139, 144–147, 154, 157–158, 160–161, 166, 168, 170–178, 181, 183, 185, 189 Newman, Paul  148, 189 Newmeyer, Frederick J.  3, 10, 12–13 Noyer, Rolf  128, 131, 161, 174, 176, 184, 187–189 Nunes, Jairo  39, 221–222, 230, 240–241 O Obenauer, Hans-Georg  254, 263 Odria, Ane  5, 8, 13, 39, 70, 76, 78, 87–88, 91, 93–94, 97, 101, 104–105, 109–111, 121, 129, 133 Oihartzabal, Beñat (see Oyharçabal, Bernard) Olaetxea, Ortzuri  138, 180, 185 Olondo, Leire  136, 167, 169, 171, 189 Oñederra, Lourdes  29, 35 Ormaetxea, José Luis  135 Ormazabal, Javier  5, 8, 13, 76, 88, 91, 93, 95–96, 98, 103–104, 107, 119, 121, 123, 127–128, 133–134, 145, 189, 279, 287 Ortiz de Urbina, Jon  3–7, 11–13, 20, 23, 34, 63–64, 67, 72, 78, 80, 82, 84, 87- 88, 90–91, 93, 97, 102–103, 111, 120- 121, 125–126, 130, 131–135, 144, 150, 156, 162, 168, 173, 186–189, 195, 217, 222–223, 225–230, 232–237, 240–247, 252, 255, 262–263, 272, 286–287

Osa, Eusebio  223, 231, 241 Oyharçabal, Bernard  5, 11, 21–22, 34- 35, 40, 64, 78, 81, 87–93, 110, 118, 120, 131–133, 141, 144, 148, 150, 154–155, 162–163, 189, 201, 217, 221, 268, 278, 286–287 P Paperno, Denis  215 Perlmutter, David  145, 190 Pesetsky, David  271, 281, 285, 287 Plank, Frans  79, 90–91 Platzack, Christer  128, 133 Plazaola, Esteban  135 Postal, Paul M.  150, 190 Preminger, Omer  68, 91, 127, 134, 145, 154, 173, 190 Preys, Jennifer  191 Pullum, Geoffrey  140, 149, 161, 190 Pylkkänen, Liina  82, 92, 110–111, 134 R Ramchand, Gillian  61, 64, 187 Raz, Shlomo  95, 134 Rebuschi, George  5, 13, 50, 65, 99, 113, 134, 181, 190, 236, 242, 244, 263 Reinhart, Tanya  143, 146, 190, 237, 242 Reuland, Eric  143, 190 Rezac, Milan  5, 8, 13, 25–26, 34, 35, 50–51, 67, 70, 76, 90–94, 97–113, 119–130, 132, 134, 143, 145–146, 148, 150, 152, 155–157, 164, 173, 176, 180–181, 184, 186, 190 Rhodes, Richard  144–145, 176, 191 Richards, Norvin  161, 191 Ritter, Elizabeth  116, 118, 134 Rivas, Albert Mario  154, 191 Rizzi, Luigi  4–5, 13, 242, 276, 279 Roberts, Ian  3, 13, 229, 242 Roca, Francesc  211 Rodriguez, Sonia  196, 217, 262 Rodríguez-Fornells, Antoni  262 Roelofsen, Floris  146, 149, 191

Rollo, William  136, 165, 191 Romero, Juan  5, 8, 13, 76, 88, 91, 93–98, 103–104, 107, 119, 121, 123, 127–128, 133–136, 145, 171, 189 Romero, Asier  136, 165–168, 172 Rooryck, Johan  143, 146, 148, 191 Rosen, Sara T.  116, 118, 134 Ross, John Robert  161, 191 Rotaetxe, Karmele  136, 165, 191 Rothmayr, Antonia  47, 64 Rowlett, Paul.  283, 287 Rullmann, Hotze  206, 217 S Sæbø, Kjell, J.  116, 134 Sáez, Luis  70, 82, 92 Safir, Ken  143, 149, 191 Sag, Ivan A.  145, 189 Sagarzazu, Txomin  106–107, 136 Sainz, Koldo  31, 34, 124, 132, 277, 286 Salaburu, Pello  89, 93, 133, 135 Salanova, Andrés Pablo  62, 65 Salazar, Belene  138 Sallaberry, Etienne  274, 288 Sarasola, Ibón  4, 13, 42, 65, 68, 91, 93, 133, 288 Schenner, Mathias  268, 287 Schlenker, Philippe  143, 147, 149, 191 Schmitt, Christina  197, 217 Schroeder, Christoph  91 Schroten, Jan  213, 215 Schultze-Berndt, Eva  78, 92, 112, 132 Shain, Cory  95, 134 Shlonsky, Ur  125, 135, 254, 263 Sigurðsson, Halldór Ármann  82, 92 Silvera, Xabier  144 Silverstein, Michael  103, 134, 177, 191 Simpson, Jane  79, 92 Sobin, Nicholas  140, 149, 191 Squartini, Mario  268, 287 Starke, Michael  147, 186 Stowell, Tim  113, 135 Stump, Greogy  161, 191 Suñer, Margarita  145, 191 Svenonius, Peter  59, 65

T Thráinsson, Höskuldur  5, 11 Torrego, Esther  8, 13, 62, 65, 70, 92, 95, 96, 98, 104–105, 107, 127, 135, 271, 281, 285, 287 Tranel, Bernard  160, 191 Trask, Robert L.  27, 34–35, 99, 118, 121, 135, 150, 196, 217, 286 Tsoulas, George  56, 63 Txillardegi [José Luis Álvarez Emparantza]  49, 102, 130, 201, 217 U Urgell, Blanca  62, 190, 217, 263 Uria, Larraitz  10, 265, 266, 277, 287 Uriagereka, Juan  6, 13, 131, 225, 228–229, 240, 242, 245, 261–263, 279, 287 Uribe-Etxebarria, Myriam  11, 39, 49, 53, 56, 59, 63, 65, 113–115, 129–131, 152, 214–217,

Name Index  227, 240, 256, 262–263, 272, 279, 285–287 Urkijo Orueta, Natxu  136 V van der Hulst, Harry  34 van der Wurff, Wim  132 van Geenhoven, Veerle  204, 216 van Kemenade, Hans  132 Vicente, Luis  112, 135 W Wiemer, Björn  268, 287 Wilhelm, Andrea  206, 214, 217, 242 Winter, Yoad  208, 217 Y Yang, Charles  161, 191 You, A.  206, 214, 217 Yrizar, Pedro de  25, 35, 102, 107–108, 135, 155, 163, 164, 166, 171, 176, 191

Z Zabala, Igone  78, 89, 92–93, 110, 113, 115, 119, 135, 196, 217 Zagona, Karen  215 Zamparelli, Roberto  199, 217 Zanuttini, Raffaella  283, 287 Zelaieta, Edu  135 Zribi-Hertz, Anne  192 Zuazo, Koldo  7, 13–21, 25–26, 29–32, 35, 100, 104, 108, 115, 135–136, 164, 192, 221–223, 242, 287 Zubeldia, Larraitz  266–268, 287 Zubizarreta, María Luisa  110, 135 Zuluaga, Irune  221 Zuluaga, Miren  221 Zvolenszky, Zsófia  146, 148, 186 Zwarts, Joost  208, 216–217 Zwicky, Arnold  161, 190

Subject index A Abaza  155 absolutive  7, 20, 23, 203 absolutive displacement  157, 158, 159 absolutive shift  158, 159, 160 pronouns in displacement  159 accomplishments  84 accusative  7, 68 activities  47, 84 additive particles  257 adjacency  6, 225, 232, 278 lack of wh-verb adj.  253 adjunct control  113 adjunction  283 adjuncts  251 adposition  40, 41, 50–62 adverbs  46 affectedness  105, 122, 130 affricates  28 agent  43, 52, 80 agentivity  7, 39, 41, 53, 61, 105, 118 Agree  113, 121, 123–128, 146, 271, 278, 280–283 agreement  4, 13, 91, 93, 106, 107, 122–134, 141, 149, 230, 235, 280 +person  122–123 absolutive  94, 119 agreement mismatches  101 agreement ‘monsters’  162 allocutive  105 causee argument  88 clitics and agreement  145, 147, 154 dative  21, 67, 88, 94, 97, 99–100, 109, 117–123, 127, 129, 154, 170 dative agr. optionality  21, 150, 155 DOM and agreement  8, 91, 94, 101–102, 108, 126, 127 doubling  122, 154, 180 ergative  43, 157

ergative deletion  168 ergative in unaccusatives  157 feature  280 from plural to polite  31–32 gaps in unaccusatives  158 impoverishment  167, 172–175, 181, 182–185, 189 increasing complexity  178 loss and default filling  180 markers  8, 146, 277 monotransitive  99, 106 morphemes  142 morphological gaps  161 number  57 object agreement  5, 25, 95 obliteration  167, 170–175, 182, 183 paradigm loss  164 person  119, 120, 124 pluralizer  26, 32–33, 58, 158, 162 position of person affixes  179 prefixing/suffixing in ED  180–184 remodelling  174 remote  154, 163 simplification  171, 178 single dative restriction  154 tripling  154 variation  141 verbal agr. productivity  162 with double datives  88 Algonquian  176 allocutive  141, 165, 171, 181, 183 allomorphy  99, 128, 142, 145, 166, 172, 175, 184 analogy  33, 163 analytic verbs  40, 100, 140, 223–224, 232–236 Ancient Common Basque  31–32 animacy  25, 76–77, 83, 93–98, 102–105, 112, 119, 127, 169 anticausative  169 antipassive  69, 176 antisymmetry  279

applicative  50–51, 82, 94, 109–111, 114, 118, 121, 126–128, 151 argument position  9, 195–198, 206, 209–210, 213 argument structure  11, 22, 61, 63, 89, 90, 125, 131 article, (see also determiner) definite  196, 203 grammaticalization  31 indefinite  198–210 aspect  40–41, 53, 60, 238 Aspect head  56 aspectual classes  45 frequentative  43 habitual  45, 49 imperfective  27, 41, 43, 49, 50, 223, 224, 236, 237 inchoative  44 perfective  44, 237 progressive  22, 52–53, 125, 222, 242 -tzen aspectual suffix  27, 41–45, 48–55, 57, 60 aspiration  28 atelic adverbial  199 auxiliary  41, 99, 101, 140, 149, 184, 268, 279 auxiliary preposing  20 bivalent  25 bivalent intransitive  26 transitive  26 trivalent  25 Axial Parts  56, 59, 65 B bait- complementizer  21 bare nouns  9, 58–59, 195–217 bare plurals  198, 202, 206 Bascoid  32 Basque Academy, (see also Euskaltzaindia)  16, 18, 25, 33 Basque Country Spanish  96, 123 ba-support  231–235, 239 be  51–52, 54 Belhare  145

 Microparameters in the Grammar of Basque bilingualism  16, 18, 24, 26, 30, 145, 256 binding  6, 150 Condition B  8, 139–140, 143–147 Borer-Chomsky conjecture  4 borrowings  31 Breton  226–227, 240–241, 287 C Cantonese  206 capacity nominals  208 cartography  10, 125, 242, 265, 267, 270, 285 case, (see also under ergative, absolutive, dative) ablative  24 allative  40, 58, 77, 121, 122, 158, 165 alternation  78, 79, 84, 86, 87, 103, 109, 111, 124, 158 benefactive  24 Case  119 comitative  24 dative doubling  87, 119–123 genitive  20, 24, 25, 27, 40, 56, 57, 58, 62, 68, 85, 86, 113 inessive  40, 48, 49, 53, 54, 57 inherent  40, 69, 94, 96, 109, 112, 115, 120, 121, 124 licensing  120 locative  25, 53, 54, 72 partitive  24, 59 prolative  24 relational -ko  22 structural  52, 93–96, 109, 112–113, 115, 119–127 Catalan  183 causal clauses  21 causatives  78, 80, 87, 110, 121, 150 causee  87, 88, 120–121, 155–156 c-command  6, 97, 120, 253, 256, 281 Celtic  32 central coincidence  7, 39, 41, 49, 51, 53, 55, 60–61 Chinook  134, 177, 186, 190 Choctaw  177 Chukchi  176 circumlocutions  140, 158, 160–161, 165, 167–168, 171 clausal pied-piping  9, 247–248, 252, 263

clausal spine  10, 260, 265–267, 270, 279, 284 clitic doubling  123, 127, 132–133, 145 clitics  143, 146–147, 150, 164, 279 collective plural  143 comparative  56 complementizers  68, 222, 231, 274–276, 279 complex NPs  251 complexity  4–5, 70, 78, 89, 142, 164, 178, 181–183, 262 compounds  56–58, 75 conditional  171, 179, 267 configurationality  3–6 conflation  72, 82 contact phenomena  4, 10, 22, 24, 28, 31, 76–77, 86, 102, 211, 254, 260, 267 copula  43, 46–50, 64, 113–117 stage level  49 covaluation  143 Cree  276 crossover  6, 252 D dative  5, 24–26, 149 adjectives and datives  151 and ECM  152 dative displacement  5, 13, 101–106, 134, 158 dative flag  99, 101, 118, 128 dative marked objects  7, 8, 67, 71–75 dative marker  26 ethical datives  117 high and low  119–123, 150–155, 165 in unaccusatives  156 nonagreeing  155, 158 of interest  155 of possession  153, 155 default tense-mood prefixes  181 definiteness  76, 103, 106, 197, 199, 200–201, 209, 215 degree  43, 46–47 deletion  176, 231–235, 238 demonstrative, (see also determiner)  23, 31, 196, 277 depictives  78, 81, 110–112, 121, 132 derivation  27 derivational morphology  27

determiner  56, 58, 59, 195–214 null  9, 195–197, 202–206, 209, 211, 214 deverbal nouns  85–86 dialects (see also the Basque Varieties Index) Ancient Common Basque  31 classification  16–17 dialectal boundaries  16–19, 30–31, 36 dialectal diversification  30, 34 dialectal markers  19–30 intelligibility  16 koiné  32 lexical differences  19 literary dialects  33 origins  30–32 Differential Object Marking  5, 8, 11, 13, 26, 34, 67, 70, 73, 80–81, 83, 86–137, 169, 172, 181 Basque DOM vs. Spanish DOM  97 DOM and animacy  103–105 DOM and dative objects  75–78 DOM and definiteness  106–107 DOM and person  106–107 DOM and tense  107, 124 Hindi DOM  97–98 Spanish DOM  95–97 diminutives  29 discourse  244, 257, 275 dissimilation  29 Distributed Morphology  161 DOM. See Differential Object Marking domain restrictor  210–211 dummy verb, egin  9, 23, 234–238 E English  12, 45, 50, 56–59, 75, 79, 91–92, 110–115, 118, 120, 122, 125, 129, 132, 145–149, 152, 161, 163–164, 198–202, 206, 209, 211, 214, 224, 229, 231, 257, 274–277 epenthesis  29 EPP  153, 229, 242, 283

ergative  7, 23, 39, 40–41, 43, 47, 49, 50–52, 54, 61–62, 99 ergative displacement  124, 141–142, 158, 160, 163–164, 171, 178–185 ergative subjects  7, 4–43, 53, 64 nonagreeing  169 ergativity  3–7, 11–12, 34, 39, 53, 55, 61–64, 89–90, 98, 124, 127, 133–134 European Portuguese  226, 241 Euskaltzaindia  25, 45, 63, 91, 130, 132, 135–136, 196, 216, 223, 225, 228, 240, 286, 288 evaluative judgement  267 Even  148 eventualities  47 evidentiality  10, 265–288 conditional strategy  268 conjectural evidential  280 conjectural ote  277 decomposing particles  277 derivation from omen  280 Evid head  227 evidential particle  225, 234, 265, 267 finiteness and ev. particle  271, 272, 276 freedom of realization  269 multiple evidentials  282 parenthetical evidentials  279, 282, 285 position w.r.t. auxiliary  281 quotative interpretation  273, 274, 275, 276, 277, 279, 282 reportative ev. vs. epistemic modality  267 reportative evidential  275, 280, 285 Exceptional Case Marking  8, 94, 96, 109, 112–113, 119, 149, 151, 153 and adjectives  153 exclamation  259 existential readings  9, 195, 197–213 experiencer  50–51, 80, 114, 116, 120, 149–153, 156, 165 expletive  113, 232, 238 extended projection  113

Subject index  external argument  40–41, 51–55, 60–61, 68, 83, 103, 105, 111, 279 extraction  245–248, 250–252, 257 F finiteness  103, 107, 125–126, 230 Fin head  109, 127 focus  6, 9, 23, 221–239, 240, 242, 243, 244, 245, 246, 251, 252, 254, 256, 259, 260, 261, 272, 282 ba strategy  223, 231–234 contrastive  223–224, 232, 234, 236, 255 dummy egin strategy  236–238 emphatic  259 exhaustive/complete interpretation  256–257 Focus phrase  6 information  237 in-situ  237 mirative  253, 258–259 mirative and contrastive  259 movement+auxiliary  255 multiple  227 narrow  238 new information  232 polarity  223–224, 232–234 presuppositions  255 raising  226 reinforced  254, 256 right edge  255–256 sentence final  260 strategies  243–263 in-situ  253–254 mirative  258–259 rightward  255–258 split interrogative  259–261 standard (‘V2’)  243–253 verb focus  221–242 verb-doubling  222–231 VP focus  236, 238 wide  236 French  9–10, 12, 16, 18, 22, 24, 28, 32, 42, 99, 110, 114–115, 119, 131, 134, 143–151, 154–156, 196, 211, 215, 224, 253–256, 262, 267, 269, 274, 287 fricatives  28

functional categories  5–6, 10, 241, 265, 285 G gaps  139–192 interpretive  143–149 morphological  161–166 stopgaps  166–171 syntactic  149–154 garden paths  184 Gascon  32, 211 Georgian  5, 134, 155 German  8, 69, 71, 78–79, 82, 85–87, 90–91, 110, 133, 268, 287 Germanic  85, 112, 128, 229, 241 gerund  113, 125 goal  69, 70, 75, 78, 82, 85, 88–89, 109, 120–121, 127, 156, 158 grammaticalization  31, 105, 140, 161, 170, 210, 213, 215 Greek  199, 216 Greenlandic Eskimo  204 Guarani  95 H Haitian  226, 240–241 Hausa  148 have  51–52, 54–55, 61, 114 heads adjunction  230 Aspect  283 Deictic  278 Evid and EpistMod  269 Evidential  260, 270, 278, 280 evidential omen as complex head  270 head adjunction  235 head movement  128, 227, 234, 235, 238, 239 head parameter  6, 11 I-to-C  229 order of evidential, epistemic and evaluative  267 polarity  227, 270, 272, 279 silent  278 T-to-C  245, 248 T-to-Focus  230, 234, 238 V–T movement  228 Hebrew  95, 226, 230, 241 Hindi  8, 65, 94–98, 103–104, 111–112, 129, 131, 133

 Microparameters in the Grammar of Basque holder  48, 53–54, 61 Hungarian  146, 148 hypercorrection  77, 86 hypothetical  171, 178, 180, 182 I Iberian  164 Icelandic  82, 92, 119 idiom chunks  113 imperative  22, 26, 41, 165, 171, 179 impersonal clauses  8, 67, 78–85, 97, 166–170, 175–177 implicatures  259 incorporation  204–205, 217 indefinite plural  204 indefinite pronouns  28 indefiniteness  197, 203, 209 infinitival complement  43, 125 infinitive  41, 49, 153, 228, 230, 237 initiator  61–62 internal argument  49, 60, 88, 103, 105, 110–112, 209, 211, 213, 279 interpretability  281 inter-speaker variation  19 intervention effects  9, 253, 262 intonation  21, 222, 228, 230, 235–236, 255, 273, 276, 281–282 inversion  246–247 Iranian  103, 124, 132 irrealis -ko  41, 57 islands  9, 140, 149, 251–254 isoglosses  15–19, 24, 28, 31, 81, 164 K kind reading  198–199, 203, 205 Korean  226, 241, 267 Kwa  226, 239, 241 L last resort  154, 232, 235 Latin  27–32 left branch phrases  251 left periphery  3, 6, 9–10, 230, 238, 243, 245, 247, 262, 265, 284–285 leísmo  8, 93–98, 101–102, 111, 123, 133

LF  174, 261 light verbs  68 linearization  146, 174, 183, 229 M macroparameters  3 Mandarin Chinese  206, 217 markedness  173 mass terms  198, 206–207, 213 microparameters  5, 6, 10, 262, 265–266 microvariation  6, 269 minimal domain  128 Minimalist Program  3–4, 11, 261, 286 modals  49–50, 267–268 Mohawk  155 mood  179 indicative  27, 171, 179 potential  22, 26, 27, 161, 178–184, 252, 259–260, 268 subjunctive  22, 26, 165, 171, 179, 268 movement auxiliary vs. verb complex  256 cyclicity  245–246 focus to A and A’ positions  259 N nasalization  29 negation  20, 161, 230, 233, 253, 256, 283, 285 new information  234, 239 nominalization  20, 53, 62, 65, 85–86, 151, 161, 171 noun number  197 null elements  202 numerals  23, 207, 213 Nupe  226 O object absolutive  68 accusative  68 agreement  5 dative/absolutive alternation  72–75, 124–126 animate  76 canonical marking in Basque  98–99

dative  67–92 direct  20, 25, 32, 40, 51, 57, 68, 78, 80, 95, 103, 111, 123, 196–199, 201–202, 204, 207, 209 genitive objects  20, 85–86 human  103 implicit  72–74, 82 indirect  4–5, 8, 21, 25, 67–92, 93–101, 103, 109–112, 119–129, 277 object shift  94, 96, 98, 109, 111–112, 118–119, 127–128 path  74 position  9, 121, 199, 202, 204–206, 209–213 Ojibwa  144–145 Old Basque  27, 208 orthography  28 overlapping reference  145–146 P P head/feature  118, 127–128 palatal  28 palatalization  26, 28, 34 participle  22, 26–27, 40, 49, 57–58, 125 bare perfective  22 future  27 participial periphrases  258 perfective  22–23, 27 predicative  56 resultative  56 passive  8, 52, 67, 78, 85, 96–97, 110, 113 Persian  95, 133 Person Case Constraint  8, 11, 13, 91, 119–123, 134, 140, 147, 154–160, 171 phi-features  140–141, 164, 166, 170–171, 175–176, 180, 183 phonologically weak element  210 pitch and stress  30 plurality  146 pluralizer  99, 141–142, 172, 184 polarity  284 positive polarity ba-  232, 234, 273 possession  114–115 inalienable possession  72, 74

possessor  114, 116, 120, 130, 149, 155–156 postpositions  25 pragmatic coreference  146–147 pragmatics  168, 177 predicate aspectual  71, 74–75 bare analytic  48–53, 55–57 bivalent ergative  23, 72 bivalent transitive  99 bivalent unaccusative  73, 74, 80–81 bivalent unergative  7, 8, 67–92, 94, 102–103, 111–112 change of position  44 change of state  44 communication  71 contact  75 decomposition  61 ditransitive  8, 67–89, 95–103, 112, 117, 121, 130, 133, 177 episodic  205 eventive  53, 55–61 fronting  226 individual-level  115, 117, 152 interaction  69 kind-level  200, 203 motion verbs  156 non-agentive  53 non-eventive  43, 50 perception predicates  43, 47, 80, 113, 152, 153 physical contact  71–72 predicate fronting  284 primary  49 psych-verbs  153, 156 relative motion  71 secondary depictive  70 secondary predicate  8, 67, 70, 78–82, 88, 91, 109–110, 115–116, 132, 217 stage-level  115, 116, 152, 199, 200, 208 stative  7, 39, 41, 44, 45, 46, 47, 48, 50, 55, 61 taking dative objects  69–92 transitive  40 unaccusative  4, 39, 52, 54, 65, 80, 87–88, 99, 149, 156, 203, 216

Subject index  unergative  7, 40, 48, 68–75, 85, 88, 99 presupposition  257 Probe  278, 281 process  61 process adverb  43 Q quantifiers  119, 198, 205, 207, 210, 213, 215 questions dummy wh-word  260 exhaustive/complete interpretation  257 matching questions  260 pair-list answer  257 Span. sentence final wh  256 split confirmatory constructions  259 strategies  244 wh in situ  9, 243, 252, 253, 254, 261, 262 wh-movement  243–244, 250–256, 260–261, 286 wh-questions  9, 223, 230, 243–244, 247, 251–252, 257, 261 reinforced  257 wh-words  6, 9 yes/no particle al  21 yes/no questions  21, 229 yes/no suffix  21 R raising  113–116, 149–150, 232, 234, 237, 284 realizational morphology  173, 241 reanalysis  209 reciprocal  144, 169 referentiality  211 reflexive  22, 44, 84–85, 107, 112, 143–151, 169 register  86 relational nouns  74 relative clauses  21–22, 87, 162–163 accessibility  78 appositive  21 case-matching  87 remnants  6, 236–237, 241, 255–256

repair strategy  122, 140, 147, 151, 154, 156–158, 161, 232, 235 no repair for PCC  160 PF repair  230 reported speech  274 restructuring  126, 135, 150 right-periphery  237 Romance  11, 31–32, 42, 89, 92, 95, 130, 145, 146, 147, 199, 261–263, 267–268, 287 Romanian  95 Russian  226, 239 S scope  9, 119, 129, 197, 199, 201, 204–205, 209, 210, 213, 236, 250, 256, 275–276 wh  250 sluicing  260 small clause  49–53, 113–115, 149, 152–153 source  80 Spanish  4, 8, 16, 18, 22–28, 42, 54, 70, 72–73, 75–77, 80, 82, 85–86, 94–98, 101–111, 114–115, 119–125, 128–130, 133, 145, 147, 154, 168–169, 196, 199, 202, 206, 211, 215, 256, 260, 262 speakers new speakers  18–19, 30 traditional speakers  18 young speakers  18, 30, 42, 54, 158, 169, 172, 183, 253 specificity  93, 95, 97, 103, 106, 112, 131 Specified Subject Condition  150, 153 speech act particle  225 spell out  230–231 Standardization  4 states  7, 39, 42–48, 60–61 verbal states  53–55 stativity  7, 34, 44, 46, 47, 55 stigmatisation  184 stress  24, 221, 223, 230, 234, 272 strong pronouns  143 subject absolutive  42–43, 52, 74 adjunct  111

 Microparameters in the Grammar of Basque ergative  39, 42, 48, 51–53, 61, 68, 114 impersonal  83 position  201–203, 207 unaccusative  57 superlative  56 syncretism  141 synthetic verbs  23, 40–41, 45, 50, 100–101, 117–119, 141, 222–225, 228, 231–238, 272 as clitics  272 paradigm loss  166 T temporal adjuncts  98, 229 temporal clauses  56, 59, 65 tense  103, 171, 174, 178 barred from initial position  228 future  41, 57

hodiernal  22 non-habitual  46 past  26 present perfect  22 prospective  44, 56 tense-mood system  179 Thai  206 theme role  69 Tigre (Xasa)  95 topic  228, 231, 233, 255, 282 topicalization  227, 236–237, 279 TopP  227, 255, 279 transitive predication  50–51, 112–119, 152, 160 transitivity variation  22–23 transmission  18 Turkish  95, 133 U Urdu  8, 94–98, 103–104, 111–112, 129, 131

V v (small v)  7–8, 43, 56, 61, 94, 109, 111, 113, 118, 125–128, 205, 245–252 V1  9, 221–222, 234–235, 239, 241 V2  9, 229, 241–242, 253 valuation  281 verbs, see under predicate doubling  9, 221–239 radical  22 citation form  41 vowels  29 W word order  3–5, 9, 10–11, 135, 153, 225, 229, 236, 243–245, 262, 286 Z Zazaki  124

This book is an endeavor to present and analyze some standard topics in the grammar of Basque from a micro-comparative perspective. From case and agreement to word order and the let periphery, and including an incursion into determiners, the book combines ine-grained theoretical analyses with empirically detailed descriptions. Working from a micro-parametric perspective, the contributions to the volume address in depth some of the exuberant variation attested in the diferent dialects and subdialects of Basque. At the same time, although the contributions focus mainly on Basque data, cross-linguistic evidence is also presented and discussed. Ater all, the goal pursued in this book is to attempt to explain variation in Basque as a particular instantiation of variation in human language at large. The volume presents and analyzes a wide range of empirical phenomena, many typologically marked among European languages, and will therefore be a welcome resource to linguists looking for detailed description and/or theoretical discussion. “This is an essential volume for all researchers working on generative approaches to dialectal variation and the Basque language. A descriptively nuanced and theoretically signiicant contribution to the microparametric endeavor, providing detailed discussions and accounts by some of the most outstanding linguists in the ield.” Itziar Laka, UPV/EHU “We are lucky to have Basque, rich in all its dialectal glory, well and alive – and even luckier to also enjoy an inexhaustible quarry of hard-working linguists studying it all. First and foremost, this is true for studies in general linguistics, which must never forget the counterpoint that languages out of the mainstream ofer us. It is also true, more broadly, for a deeper understanding of the great mystery of language variation, which a language with under a million speakers shows great examples of. Last but not least, studies like the present one remind us of the importance of preserving our cultural heritage: Basques can bless their lucky stars, but their eforts can also serve as a great example of what is possible with systematic work and passion, for so many communities around the world.” Juan Uriagereka, University of Maryland “This impressive collection continues in the rich tradition of Basque linguistics. It contains micro-comparative studies touching on a wide range of topics in Basque syntax, emphasizing the importance of dialect variation. The individuation of the microparameters that underlie this variation can be seen to play a key role in deepening our understanding of general properties of the human language faculty.”

isbn 978 90 272 0830 9

Richard S. Kayne, New York University

John Benjamins Publishing Company