Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Edited by
LEIDEN | BOSTON
Library of Congress Cataloging-in-Publication Data
Names: Berro Urrizelki, Ane, editor. | Fernāandez, Beatriz, editor. | Ortiz de Urbina,
Jon, editor.
Title: Basque and Romance : aligning grammars / edited by Ane Berro Urrizelki,
Beatriz Fernandez, Jon Ortiz de Urbina.
Description: Leiden ; Boston : Brill, [2019] | Series: Grammars and language
sketches of the world's languages | Includes bibliographical references and
index.
Identifiers: LCCN 2019008329 (print) | LCCN 2019009310 (ebook) |
ISBN 9789004395398 (ebook) | ISBN 9789004395381 (hardback : alk. paper)
Subjects: LCSH: Basque language–Grammar, Comparative–Spanish. | Spanish
language–Grammar, Comparative–Basque. | Basque language–Grammar,
Comparative–Romance. | Romance languages–Grammar,
Comparative–Basque.
Classification: LCC PH5023 (ebook) | LCC PH5023 .B298 2019 (print) |
DDC 499/.925–dc23
LC record available at https://lccn.loc.gov/2019008329
Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill‑typeface.
ISSN 2352-9342
ISBN 978-90-04-39538-1 (hardback)
ISBN 978-90-04-39539-8 (e-book)
1 Introduction 1
Ane Berro, Beatriz Fernández and Jon Ortiz de Urbina
2 Word Order 14
Jon Ortiz de Urbina
Introduction
Ane Berro, Beatriz Fernández and Jon Ortiz de Urbina
1 Introduction
tury was more similar to that of neighboring Romance. As Hualde and Perez
Saldanya (this volume) also show, grammaticalization patterns for causal sub-
ordinate clauses have followed similar paths. As a last instance, Basque shared
with neighboring languages (even with English) important developments such
as the deployment of second personal plural pronouns and verbal morphemes
as markers of respectful second singular forms. All this indicates that shared
developments and changes towards more Romance-like patterns have been
taking place in Basque for a long period now. In the context of this centuries-
long contact, the fact that we can still speak of Basque and Romance as being
typologically ‘different’ languages may also require some explanation. In the
last 150 years, however, factors such as the strengthening of French and Span-
ish state-level political organization, early Romance-exclusive schooling and
contemporary bilingual schooling, sociological changes in population make-up
and urbanization, the development of mass communication systems, among
many others, have led to the extended bilingualism which may underlie the
extensive changes Basque seems to be currently undergoing and which can
be seen in this book. Obviously, language coexistence in space compounded
with extensive language coexistence in the speakers’ brains triggers linguis-
tic changes which make of Basque and Romance contact a data-rich field for
linguistic studies from many different angles. The articles in this book aim at
striking a balance in presenting as wide an overview of the most important
phenomena as possible while providing at the same time detailed data and, in
many cases, theoretical analyses.
Research on Basque and Romance has made it clear that despite their typo-
logical differences (ergative vs. accusative marking, head final vs. head initial
linearization, among others), Basque and Spanish/French show remarkable
similarities. Thus, even if Basque can be descriptively characterized as a fairly
typical OV language and both French and Spanish also display fairly standard
VO features, information packaging mechanisms, however they are construed,
produce in Basque clause-level orders which are (increasingly) similar to those
found in Spanish and French (Ortiz de Urbina, this volume). Similarities are
more evident in colloquial registers and in some dialectal contexts (schooling
is mostly done in the standard variety). Some important similarities are there-
fore formally related to Romance phenomena (whether dialectal or not). For
instance, Differential Object Marking (Bossong 1991, Lazard 2001, Aissen 2003)
is attested in some South-western varieties in the Spanish-speaking area of the
4 berro, fernández and ortiz de urbina
Basque territory and shows great similarities with Spanish general a-marking
(Torrego 2004), or even with the dialectal leísmo of the Basque Country (Fer-
nández & Rezac 2016, Odria 2017 this volume).
In the same vein, the use of the se clitic in French and Spanish is closely
related with the absolutive marking of the subject and izan ‘be’ auxiliary selec-
tion in Basque intransitive verbs, especially in Romance-loan verbs (Alberdi
2003). When the verb in the original Romance language occurs with the se
clitic, the borrowed predicate in Basque takes an absolutive subject and izan
‘be’ auxiliary, whereas when the verb in the original Romance language does
not take se, then the predicate is generally used with an ergative subject and
*edun ‘have’ auxiliary in Basque. As a matter of fact, when Spanish and French
differ in the use of se in certain verbs, the loan verb in Basque shows dialectal
variation, with a contrast between the Basque varieties in contact with Spanish
and those in contact with French. The analysis of loan verbs and the compari-
son of counterpart intransitive verbs in Basque and Romance is of particular
interest, bearing in mind the different case systems of the languages under
study—Basque being ergative (Levin 1983, Ortiz de Urbina 1989, Oyharçabal
1992 among others) and Spanish/French being accusative—and auxiliary alter-
nation. For instance, Spanish, unlike French, does not display auxiliary alterna-
tion, but, even so, the presence of the se clitic in Spanish goes hand in hand with
the selection of the izan ‘be’ auxiliary in Basque loan verbs (Berro, this volume).
Other similarities go beyond dialectally constrained structures, such as the
nature of datives in general and non-argumental datives in particular. Although
recent literature has focused on the high vs. low nature of applied datives espe-
cially after Pylkkänen 2008 [2002], even higher in the architecture of the sen-
tence seem to be the so-called ethical datives, as analyzed by Jaeggli (1982, 1986)
for Spanish and by Jouitteau & Rezac (2007) for French among others. The
sole existence of ethical datives in Basque is not uncontroversial, as affected
(experiencer) datives seem to be of an argumental type, and therefore an in-
depth analysis of affectedness in Basque and Spanish/Basque is important.
Actually, Bosse, Bruening & Yamada (2012) provide with a fine-grained typol-
ogy that includes Hebrew, French, Japanese among other languages but lacks
both Basque and Spanish materials. Thus, an incursion into affected datives
and even allocutives (Oyharçabal 1993) will give us the opportunity to review
classical syntactic tests (as in Borer & Grodzinsky 1986) and more novel criteria
proposed by Bosse et alia (2012) and Horn (2008, 2013) in order to see to what
extent affected datives are argumental or not in both Basque and Spanish (Fer-
nández, this volume). The interplay between goal, experiencer, ethical datives
on one hand and causee datives in causative structures on the other is also
described for Basque and Spanish in Ortiz de Urbina (this volume). The appar-
introduction 5
grammar and the analysis of phenomena which may be due to their interaction
can have a wider linguistic interest. Throughout the chapters it contains, topics
such as word order, inflection, non-verbal participles, intransitive predicates,
light verb constructions, ethical datives, differential object marking, causative
constructions and causal clauses will be presented. In what follows, we briefly
outline the main contributions in each chapter.
The labels OV/VO or head-first/head-last, at least as general characterizations
of (some) central elements in (some) phrasal constituents, would seem to place
Spanish and French on one side of this typological divide and Basque on the
other. Jon Ortiz de Urbina’s chapter on ‘Word Order’ shows that while Basque
and Romance often differ in the location of their morphological and syntactic
heads, considerable similarities (sometimes even convergence) may be found
at the clausal level. Thus, SOV and SVO may correspond to grammatically neu-
tral, pragmatically bleached orders in both languages (even if, at least in the
case of Basque, certainly not to the most common ordering). However, there are
numerous information packaging mechanisms which account for the majority
of actual clause-level orders, and some of these, old and new, produce simi-
lar orderings in the two languages. Even though Basque does not have the right
peripheral information focus of Spanish, the latter does have as an option a pre-
verbal focus akin to the general-purpose focalization strategy of Basque, and
clause final corrective focus is also increasingly used in Basque. In the course
of the presentation of phrasal order, the chapter also points at ‘non-harmonic’
orders within lexical phrases, appositions and compounds. Their diachronic
development and the comparison with Romance suggests that, at least in some
cases, their presence is not just a function of well-known OV inconsistency.
With virtually universal bilingualism in either Spanish or French and, in the
western area, an important mass of new speakers that make up a good portion
of Basque speakers in heavily populated areas, it would be surprising if a cer-
tain degree of grammatical confluence did not take place. We seem to observe
increasing use of orders which exploit possibilities, traditional or not, which
are compatible with the Spanish/French grammar of the same users.
The chapter ‘Tense, Aspect and Mood’ coauthored by José Ignacio and Céline
Mounole explores the morphological shape of the Tense, Aspect and Mood
(TAM) system in finite verbal forms. As previously noticed in the literature
(Michelena 1981), Basque analytical verbal forms have Romance counterparts
that suggest similar diachronic development. Some similarities between
Basque and Romance are the distinction between indicative and subjunc-
tive, along with conditional and imperative, as far as the mood is concerned,
or the contrast between perfective and imperfective forms regarding aspect.
Nevertheless, there are also certain differences. For instance, Basque shows
introduction 7
a be/have auxiliary alternation whereas Spanish has only have auxiliary and
French selects mainly have and reserves be for a small class of intransitives.
These and other aspects show that Basque has both a convergent and a diver-
gent development with its Romance neighbors.
Ane Berro’s chapter ‘Non-verbal Participles in Basque and Spanish’ com-
pares and contrasts non-verbal participles in both languages, in light of the
literature on adjectival participles (among others, Wasow 1977, Marantz 2001
2007, Kratzer 1994 2000, Anagnostopoulou 2003, Embick 2004). The author
points out a number of differences, regarding the morphological shape of the
non-verbal participles and their aspectual interpretation. In fact, participles
take extra morphemes in Basque when they are adjectival or postpositional,
whereas in Spanish both verbal and non-verbal ones look similar. As for the
aspectual interpretation, Basque non-verbal participles can be used with both
resultative and experiential interpretation, but Spanish non-verbal participles
only trigger the resultative reading. Despite these differences, Spanish and
Basque participles share a number of properties. On the one hand, they show
a similar interaction between the lexical aspect of the verb embedded under
the participle and the aspectual interpretation of the whole structure. On the
other, non-verbal participles in both Basque and Spanish (at least, the Spanish
spoken in the Basque Country) (cf. Gehrke & Marco 2014), allow spatial and
temporal modification of the event underlying the participle.
The chapter ‘(In)transitive Verbs: Unergatives and Unaccusatives’, also
authored by Ane Berro, analyzes the unaccusative/unergative divide of intran-
sitive predicates (Perlmutter 1978 1989, Burzio 1981, 1986, Belletti 1988 1999,
Sorace 1993, Levin & Rappaport Hovav 1995 and subsequent works) in Basque
and several Romance languages. The author focuses on the verb classes that
are systematically aligned in the unergative way or, alternatively, in the unac-
cusative way, both in Basque and Romance, and also border-line verbs (like
Romance loan verbs and some new stative verbs). As proposed by the author,
intransitive stative verbs are the most variable verbs both cross-linguistically
and in Basque, precisely because the only argument of these predicates can be
introduced externally or internally without making a big semantic difference.
Víctor Acedo-Matellán and Anna Pineda’s chapter ‘Light Verb Constructions
in Basque and Romance’ analyzes the morphosyntactic and semantic prop-
erties of Light Verb Constructions (LVC) in Basque and compares them to
their counterparts in Romance—mainly Spanish and French. They argue that
the Light Verb (LV) is responsible for the event- and argument-properties of
the construction, whereas the Non-Verbal Element (NVE) provides conceptual
information of the event involved. As shown in Basque and also Spanish and
French, the choice of the LV is meaningful, given that only DO and GIVE give
8 berro, fernández and ortiz de urbina
References
Aissen, Judith. 2003. Differential object marking: iconicity vs. economy. Natural Lan-
guage and Linguistic Theory 21: 435–483.
Alberdi, Xabier. 2003. Euskal aditz mailegatuen erregimena: hurbilketa. In Jesus Mari
Makazaga & Bernard Oyharçabal (eds.), P. Lafitteren sortzearen mendemugako bilt-
zarra. Gramatika gaiak. Iker 14 (I). Bilbao: Euskaltzaindia. 37–60.
10 berro, fernández and ortiz de urbina
Alexiadou, Artemis, Elena Anagnostopoulou & Florian Schäfer. 2015. External Argu-
ments in Transitivity Altenations. A Layering Approach. Oxford: Oxford University
Press.
Allières, Jacques. 1992. Gascón y euskera: afinidades e interrelaciones lingüísticas. ASJU
26 (3): 801–812
Anagnostopoulou, Elena. 2003. Participles and Voice. In Artemis Alexiadou, Monika
Rathert and Armin von Stechow (eds.), Perfect explorations. Berlin: Mouton de
Gruyter. 1–36.
Anagnostopoulou, Elena & Yota Samioti. 2014. Domains within words and their mean-
ings: A case study. In Artemis Alexiadou, Hagit Borer and Florian Schäfer (eds.), The
Syntax of Roots and the Roots of Syntax. Oxford: Oxford University Press.
Belletti, Adriana. 1988. The case of unaccusatives, Linguistic Inquiry 19: 1–34.
Belletti, Adriana. 1999. “Inversion” as focalization and related questions, Catalan Work-
ing Papers in Linguistics 7: 9–45.
Berro, Ane. 2017. Resultative and non-resultative participles in Basque. Ms, University
of the Basque Country & SFL (UMR 7023, CNRS/Paris 8).
Borer, Hagit & Yosef Grodzinsky. 1986. Syntactic cliticization and lexical cliticization:
The Case of Hebrew Dative Clitics. In Hagit Borer (ed.), Syntax and Semantics 19.
New York: Academic Press. 175–215.
Bosse, Solveig, Benjamin Bruening & Masahiro Yamada. 2012. Affected experiencers.
Natural Language and Linguistic Theory (NLLT), 30(4): 1185–1230.
Bossong, Georg. 1991. Differential Object Marking in Romance and beyond. In Dieter
Wanner & Douglas A. Kibee (eds.), New Analyses in Romance Linguistics. Amster-
dam / Philadelphia: John Benjamins. 143–170.
Bosque, Ignacio. 1999. El sistema adjectival. Complementos y modificadores del adje-
tivo. Adjetivo y participio. In Ignacio Bosque & Violeta Demonte (eds.), Gramática
descriptiva de la lengua española. vol. 1. Madrid: Espasa Calpe. 217–310.
Bosque, Ignacio. 2014. On resultative past participles in Spanish. Catalan Journal of Lin-
guistics 13: 41–77.
Burzio, Luigi. 1981. Intransitive verbs and Italian auxiliaries. Ph.D. diss, MIT. Cambridge,
MA.
Burzio, Luigi. 1986. Italian Syntax. Dordrecht: Reidel.
Comrie, Bernard. 1989. Language Universals and Linguistic Typology. Chicago: Chicago
University Press.
Demonte, Violeta. 1983. Pasivas léxicas y pasivas sintácticas en español. Serta Philolog-
ica F. Lázaro Carreter. Madrid: Cátedra. 141–157.
Dixon, Robert M.W. 2000. A Typology of Causatives: Form, Meaning and Syntax. In
Robert M.W. Dixon & Alexandra Y. Aikhenwald, eds. Changing valency. Case stud-
ies in transitivity, Cambridge: Cambridge University Press, 30–83.
Embick, David. 2004. On the Structure of Resultative Participles in English. Linguistic
Inquiry 35-3: 355–392.
introduction 11
Fernández, Beatriz & Milan Rezac. 2016. Differential Object Marking in Basque vari-
eties. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
grammar of Basque. Amsterdam: John Benjamins. 93–139.
Franco, Jon & Susana Huidobro. 2008. Ethical Datives, Clitic doubling and the Theory
of pro. Joyce Brukes de Garavito & Elena Valenzuela (eds.), Selected Proceedings of
the 10th Hispanic linguistics Symposium. Somerville: Cascadilla Proceedings Project.
215–224.
Gehrke, Berit. 2011. Stative passives and event kinds. In I. Reich, E. Horch and D. Pauly
(eds.), Proceedings of Sinn und Bedeutung 15. Saarbrücken: Universaar—Saarland
University Press. 241–257.
Gehrke, Berit & Cristina Marco. 2014. The role of by-phrases in adjectival passives. Lin-
gua 149: 188–214.
Horn, Laurence R. (2008). I love me some him: The landscape of non-argument datives.
In Olivier Bonami & Patricia Cabredo Hofherr (eds.), Empirical Issues in Syntax
and Semantics 7, CSSP, Paris, 169–192. [On-line publication, http://www.cssp.cnrs.fr/
eiss7].
Horn, Laurence R. (2013). I love me some datives: Expresive meaning, free datives, and
F-implicature. In Daniel Gutzmann & Hans-Martin Gärtner (eds.), Beyond Expres-
sives: Explorations in Use-Conditional Meaning. Leiden: Brill. 153–201.
Hualde, José Ignacio. 2003. Introduction. In José Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A grammar of Basque. 1–14. Berlin: Mouton de Gruyter.
Hualde, José Ignacio. 2016. On Basque dialects. In Beatriz Fernández & Jon Ortiz de
Urbina (eds.), Microparameters in the Grammar of Basque. Amsterdam/Philadel-
phia: John Benjamins.15–35
Jaeggli, Osvaldo, A. 1982. Topics in Romance Syntax. Dordrecht: Foris.
Jaeggli, Osvaldo, A. 1986. Three issues in the theory of clitics: case, doubled NP s and
extraction. In Hagit Borer (ed.), Syntax and Semantics 19. New York: Academic Press.
15–42.
Jouitteau, Mélanie & Milan Rezac (2007), The French ethical dative, 13 syntactic tests.
Bucharest Working Papers in Linguistics, IX (1): 97–108.
Krajewska, Dorota. 2012. Resultatives in Basque: a diachronic study. Lingua Posnanien-
sis LIV(2): 56–67.
Krajewska, Dorota. 2013. The diachrony of resultative constructions in Basque. Master
Thesis, UPV/EHU.
Kratzer, Angelika. 1994. The Event Argument and the Semantics of Voice. Ms., Univer-
sity of Massachusetts at Amherst.
Kratzer, Angelika. 2000. Building Statives. Berkeley Linguistic Society 26. 385–399.
Lafon, R. 1980 [1943]. Le système du verbe basque au XVIe siècle. Zarautz: Elkar.
Landa, Alazne. 1995. Conditions on Null Objects in Basque Spanish and their relation to
Leísmo and Clitic Doubling. University of Southern California.
12 berro, fernández and ortiz de urbina
Perlmutter, David, M. 1978. Impersonal Passives and the Unaccusative Hypothesis. Pro-
ceedings of the 4th Annual Meeting of the Berkeley Linguistics Society. 157–190.
Perlmutter, David, M. 1989. ‘Multiattachment and the unaccusative hypothesis: the Per-
fect Auxiliary in Italian.’ Probus 1-1: 63–119.
Pylkkänen, Liina. 2008. Introducing arguments. Cambridge / London: MIT Press.
Sorace, Antonella. 1993. Unaccusativity and auxiliary choice in non-native grammars of
Italian and French: asymmetries and predictable indeterminacy. Journal of French
Language Studies 3: 71–93.
Wasow, Thomas. 1977. Transformations and the lexicon. In Peter Culicover, Thomas
Wasow & Adrian Akmajian (eds.), Formal Syntax. 327–360. New York: Academic
Press.
Zubizarreta, María Luisa. 1985. Romance accusatives. Linguistic Inquiry 16: 247–289.
chapter 2
Word Order
Jon Ortiz de Urbina
1 Morphology
1 The abbreviations used in glosses are as follows: aux ‘auxiliary’, dat ‘dative’ det ‘determiner’,
comp ‘complementizer’, erg ‘ergative’, fut ‘future’, gen ‘genitive’, impf ‘imperfective’, nom
‘nominalizing affix’, part ‘partitive’, pl ‘plural’, q ‘question morpheme’. Glosses in a language
like Basque can be very obstrusive, so for readability purposes they have been kept to a min-
imum, providing in most cases only the grammatical information for which the example is
relevant. Thus, the determiner (and agreement) marker -a may be glossed as det, ‘the’ or
simply left out depending on the relevance to the discussion.
2 Apparent exceptions such as erpuru ‘thumb’ or hiriburu ‘capital’, (eri ‘finger’ and hiri ‘town’
+ buru ‘head’), designating types of fingers or towns may not be so when compared with the
expected asteburu ‘weekend’ (aste ‘week’) or iturburu ‘spring’ (iturr(i) ‘fountain’). Notice that
while the latter are interpreted as the head of the week and perhaps as the head of a foun-
tain, the former are not the head of a finger or of a city. Rather, buru receives an interpretation
actually close to the one represented in capital, originally an adjective (from the Latin word
for head) in the collocation capital city, which the Basque word follows closely. Given that
the categorial distinction between nouns and adjectives was not clear in older stages of the
language, we may assume that buru in the first two examples is working as an adjective, and
that these are N Adj compounds of the types described below.
16 ortiz de urbina
nouns with a third person verbal form as first term: saltamontes ‘grasshopper’
(salta ‘jumps’ + montes ‘mountains’), sacacorchos ‘corkscrew’ (saca ‘takes out’ +
corchos ‘corks’). Basque NN secondary compounds (again just as productive in
Basque as they are scarce in Spanish) also follow systematically an OV pattern,
with the deverbal noun to the right as in ile apaingile ‘hairdresser’ (ile ‘hair’ +
apaingile ‘dresser’, from apain- ‘beautify’ and -gile ‘-er’) or suhiltzaile ‘firefighter’
(su ‘fire’ + hiltzaile ‘killer’, from hil ‘kill’ and -tzaile ‘-er’).3
We also find word orders similar to those of syntactic phrases in N Adj com-
pounds. These follow this order in both Spanish and Basque, as expected since
adjectives usually follow nouns in the syntax of both languages; cf. Bsq. bela-
rrimotz ‘small-eared, foreigner’ (belarri ‘ear’ + motz ‘short’), and Span. paticorto
‘short-legged’ (pata ‘leg’ + corto ‘short’). See 2.2.1.3 below for adjective ordering
within nominal phrases. There are few compounds containing adpositions in
either Basque or Spanish, and the few cases that do exist show the expected
order between the adposition head and its nominal complement. Spanish
sin ‘without’ is the most productive preposition found in compounds: sinsen-
tido ‘nonsense’, sinsabor ‘setback’, sinvergüenza ‘shameless, scoundrel’, sinsorgo
‘silly’;4 its Basque postpositional counterpart is gabe ‘without’, also very pro-
ductive, as in paregabe ‘matchless’, etengabe ‘continuous’, lit. ‘without break’,
lotsagabe ‘shameless’ or beldurgabe ‘fearless’. Other prepositions are only occa-
sionally found in Spanish: bajotecho ‘attic’, lit. ‘under roof’, sobrecama ‘bed-
spread’, lit. ‘over bed’, or sobrepeso ‘overweight’. There are even fewer examples
in Basque, partly because many Spanish prepositions correspond to Basque
‘case’ endings, bound inflectional morphemes. There are, however, indepen-
dent postpositions in Basque, of which gabe ‘without’ is a typical example.
Most postpositions, however, are relational nouns inflected for a locative case.
These combine with a noun in the genitive (or a verb in the participial form)
or may be compounded with it;5 in either case, the expected order is main-
tained: etxe(aren) aurrean ‘in front of the house’, bazkalondo ‘chat after eating’
(cf. bazkaldu ondoren/ondoan ‘after having lunch’). Some relational nouns that
form the basis of postpositions, however, can be found in prenominal posi-
tion, only indirectly as a result of accommodation to Romance patterns. Thus,
azpi ‘below’ is used as the equivalent of the Romance prefix sub- in neologisms
like azpiegitura ‘substructure’ or azpimultzo ‘subset’. Basque has a wide array
of derivational suffixes but extremely few prefixes,6 and this type of seman-
tically transparent Latinate prefix is also transparently rendered by the noun
azpi. This is then a case of accommodation to the prefixation possibilities of
Romance, rather than to any kind of prepositional pattern. Notice that the pre-
fixed element is the basic noun without the locational marker, and that there is
no complement relationship between this noun and the one to its right: azpie-
gitura ‘substructure’ is not interpreted as ‘under the structure’ or ‘underside of
the structure’, but as a structure which is located under some other. The loca-
tional noun acts as an adjunct, not as a head.7 Similar considerations apply to
the use of aurre ‘before/front’ as equivalent to the prefix pre- in neologisms
like aurrizki ‘prefix’, aurreritzi ‘prejudice’, lit. ‘before opinion’. The latter is inter-
preted not as what is held before an opinion, but as an opinion formed previous
to contact with data and/or arguments.8
We have already briefly mentioned derivation while discussing relational
nouns. As indicated above, Basque is extremely limited in terms of derivational
prefixes. This in itself fits Greenberg’s findings on the relationship between
affixation and phrasal ordering, since he observed a tendency for SOV languages
to display suffixation exclusively, while SVO languages usually exhibit both. In
principle, if the affix is the head of a derived word, determining its category, one
6 As pointed out by de Rijk (2008), only three: ber- ‘re-’ and the loaned prefixes arra- ‘re’ and
des- ‘un-’.
7 This is also the use found in other words containing azpi- in the traditional vocabulary, such
as azpijoko ‘foul play’, lit. ‘under play’, or azpizapo ‘traitor’, lit. ‘under toad’.
8 Some morpheme ordering relationships within the number system were already claimed to
be linked to other phrasal relations in Lehman (1975). In particular, he claimed that in addi-
tive numbers (for instance, thirteen, three plus ten, as opposed to thirty, three times ten) the
smaller number added to the bigger one (the decimal in this type of system) behaves like
the verb in the O/V relation. Regardless of the overall validity of this correspondence, it does
capture the relation between the Basque and Spanish systems. Few numbers are additive in
these languages, 11 through 19 in Basque and 11 through 15 in Spanish, the remaining higher
numbers being either factorial (berrogei ‘40’, lit. 2 20 in Basque and cuarenta in Spanish, with
initial 4) or coordinations (berrogeitamar ‘50’, lit. 2 20 and 10 in Basque or dieciséis ‘16’, lit. 10
and 6 in Spanish). Basque additive numbers take the form ten + unit (hamairu ‘13’, lit. 10 3)
and, more opaquely, unit + ten in Spanish (doce ‘12’, 2?).
18 ortiz de urbina
would expect that head-first languages would favor prefixation, whereas head-
final languages like Basque would favor suffixation; however, this is clearly not
the case.9
2 Syntax
After examining morpheme ordering phenomena, let’s turn now to word order
phenomena in a more strict sense, devoting special attention to interlinguis-
tically or intralinguistically unexpected patterns, i.e., unexpected when com-
pared to patterns typical of OV languages and/or incongruent with other order-
ing internal to Basque. We begin with appositions in 2.1, devoting 2.2 to word
order in major lexical phrases and 2.3 to ordering at the clausal and sentence
level.
2.1 Appositions
The relative order between type and token in restrictive appositions correlates
with other VO/OV characteristics, so that the type expression corresponds to the
V position and the token phrase to the O position. In fact, this was the content
of Greenberg’s Universal 23, where he showed a correlation between proper
noun/common noun appositions and the order of genitive/noun pairs.10 Since
the genitive/noun order patterns with the V/O order, this type of apposition
would then show overall mirror image effects when comparing Basque and
Spanish. The differences between the two languages are clear in this respect:
11 Cinque (2013) also notes counterexamples in both OV and VO languages, but the only
Romance language mentioned in this respect, Italian, conforms to the common/proper
order in all instances he checks. English displays more deviations (Church Street, Coney
Island, Potomac River), perhaps connected to the head last organization of NN com-
pounds.
20 ortiz de urbina
Ama Teresa Kalkutakoa ‘Mother Teresa of Calcutta’, or even Ama Lurra ‘Mother
Earth’, but less so in the much more productive cases involving relations like
izeko ‘aunt’, osaba ‘uncle’ or amona ‘grandmother’. Example (8) is a celebra-
tory radio message for the Bizkaia Irratia radio station indicating the persons
wishing their 6-year-old relative a happy birthday, while (9) displays two family
relation names, each in a different order:
(8) Bere aita Gaizkak, ama Isabelek eta neba Aritzek. Baita
her father G. mother I. and brother A. also
osaba Mikelek, aitona Jose Martin eta amona
uncle M. grandfather J.M. and grandmother
Mari Karmenek.
M.K.
‘Her father Gaizka, mother Isabel and brother Aritz. Also uncle Mikel,
grandfather Jose Martin and grandmother Mari Karmen.’
(9) Bere osaba Pedrok Guillerma amatxiren ganik ukan zuen etxe
his uncle Pedro Guillerma grandm.gen from have aux house
gotorrari.
strong.dat
‘To the strong house that his uncle Pedro had (received) from grand-
mother Guillerma.’ (Oncle, Ardoy SFran 84)
The Orotariko Euskal Hiztegia [Basque General Dictionary], from which the
latter is taken, lists examples with preposed and postposed osaba, but the ten-
dency seems to be towards preposing.
2.2.1 NP s
Not unlike the situation described in 2.1, Basque behaves in general like an OV
language, with some ‘anomalies’, some but not all of which can be attributed
to contact phenomena with Spanish and Romance in general. We will describe
determiners and quantifiers in 2.2.1.1, adposition phrase modification of nouns
in 2.2.1.2, adjective modifiers in 2.2.1.3 and relative clauses in 2.2.1.4.
word order 21
12 The sound [h] is not pronounced in this dialect, so there is no difference here between
the initial and final determiners.
13 The Biscayan pattern in the plural seems closer to the Spanish article+demonstrative pat-
tern: hónek mutilok ‘those guys’, with the proximal article -ok instead of a doubled demon-
strative. Since the proximal plural form -ok derives historically from the proximal plural
demonstrative *hauk (rather than the distal *hak, which is the source of the plural article),
the mismatch is actually smaller than seems to be the case at first sight. In any event, final
-ok is still the determiner, unlike the final demonstrative in Spanish.
22 ortiz de urbina
all ‘that tall guy with glasses’. In contrast, the final demonstrative in the Basque
examples is exclusively final, since it is a regular determiner after all.
Quantifiers make up a semantic class sometimes behaving like determiners,
and therefore incompatible with them. Thus, muchos in mis muchos amigos
‘my many friends’ is compatible with a determiner and, therefore, probably not
a determiner itself, while bastantes ‘quite a few, several’ cannot cooccur with
the genitive determiner, as shown in *mis bastantes amigos ‘my several friends’.
There is nothing semantically odd about the latter, since the same meaning can
be rendered in a fully grammatical way as bastantes amigos míos, with an adjec-
tival genitive míos, crucially not a determiner. Strong Basque quantifiers like
guzti ‘all’, den ‘all’, bakoitz ‘each’ or gehien ‘most’ must in fact cooccur with the
definite determiner, while weak quantifiers like asko ‘many’, batzu(e)k ‘some’,
gutxi ‘few’, zenbait ‘some’ and hainbat ‘some’ cannot coocur with them (see
Etxeberria 2005). Thus, we find lagun guzti*(ak) ‘all friends’ with obligatory plu-
ral determiner but lagun gutxi(*ak) ‘few friends’ with a bare quantifier. From
the point of view of word order, we find again the basic mirror image between
Basque and Spanish, so that the set expression to express thanks contains ini-
tial muchas ‘many’ in Spanish but final asko ‘many’ in Basque (muchas gracias
versus eskerrik asko ‘many thanks’). But, at the same time, we also find the occa-
sional disharmonic orders in OV languages, so that even if most quantifiers are
postnominal in Basque, there are also some prenominal ones. In particular,
zenbait ‘some’ and hainbat ‘some’ precede the noun they quantify over: zen-
bait lagun ‘some friends’.14 These prenominal orders may be syntactically more
complex than the postnominal ones, since the former contain operators which
typically induce displacement in clauses and/or noun phrases: zenbait appears
as zeinbeit in the earliest printed book in Basque, containing the wh-word zein-,
and hainbat contains hain ‘so’, which, as discussed in 2.2.2.2 below also appears
in a displaced position with respect to adjectives it combines with. If extra syn-
tactic positions and processes affect these prenominal quantifiers, then, the
overall mirror-image ordering of quantifiers and nouns in Spanish and Basque
would be more thorough at a more abstract level of analysis.15
14 Zenbait may (rarely) follow the noun Nahiko(a) ‘enough, quite a few’ is more complex,
since depending on the dialect it can precede or follow the noun it quantifies over; it can
also occur as an external quantifier, as if floating out of the NP, taking then the determiner:
a phrase like enough guys can be found across dialects as mutil-a nahiko-a, mutil nahiko-a,
nahiko-a mutil, nahiko-a mutil-a, nahiko mutil-a and nahiko mutil. The form is also mor-
phologically complex and may contain the genitive suffix -ko.
15 For an analysis of Basque NP s, see Artiagoitia (2002).
word order 23
The preferred adnominal preposition in Spanish is de, for both argumental and
adjunct modifiers of the noun (as in (10) and (11)), as well as attributive mod-
ifiers (14). Other prepositions are also possible, as in (12) and (13). In all cases,
the PP follows the head noun. As famously put by de Rijk (1993), Basque nouns
are not very ‘hospitable’ to any type of phrasal dependents, which require the
intervention of a genitive marker to be able to attach to the head noun. Argu-
mental, thematic relations are typically marked with the genitive suffix -ren
(as in (10)), while more peripheral adnominals are marked with -ko, added to
an adverbial or a PP as in (12) or directly to the nominal in the case of loca-
tive/temporal relations like (11) or (13) and in attributive relations as in (14).16
Although the status of these ‘genitive’ markers is not uncontroversial (see de
Rijk 1993), positionally they are always placed at the end of the phrase, and
16 De Rijk (1993) claims all cases of apparent bare nominals to which -ko is attached actually
hide PP s, whereas genitive (r)en would attach to NP s.
24 ortiz de urbina
we can equate them with postpositions for word order considerations, forming
PP s just like those headed by de or other prepositions in Spanish. Prepositional
phrases are more ‘harmonic’ with head initial VP s and NP s, as in Spanish, while
postpositional phrases are more typically found with head final phrases, as in
the Basque examples above.
Let us examine now deviations from these patterns. As usual, it is OV lan-
guages like Basque that exhibit some degree of deviation from the expected
pattern: there are no prenominal PP s in Spanish, while postnominal PP s do
exist in Basque, if in a limited way. First of all, dative adnominals constitute
both a gap in the complementation pattern of nominal heads and a possible
anomaly in the expected ordering relationships between head noun and com-
plement. Thus, goal complements of nouns cannot be attached to the head by
using either of the genitive markers discussed above, neither as nominals with
-(r)en or as dative marked nominals with -ko:
c. entzulegoaren eskaintza/eskerrak
audience.gen offer/thanks
‘the offer/thanks of the audience’
b. abisua marinelei
warning sailors.dat
‘a warning to sailors’
Thus, a song like Gracias a la vida ‘Thanks to life’ is usually rendered as Eskerrak
bizitzari, literally ‘Thanks life to’, identical to the order in Spanish and English
in the relevant respect. It is not clear, however, whether the two phrases are
really integrated into a single expression or act with at least relative indepen-
dence.17 Certainly, in the rare cases where the goal can be expressed with a
dative, it occupies the pre-head position expected for an OV language. Com-
pare for instance the previous example eskerrak bizitzari ‘thanks to life’ with
an expression like berari esker ‘thanks to him’, where esker ‘thank’ has lost its
nominal status and is used as a postposition requiring a dative complement. In
any event, the possibility that the examples in (16) involve complements which
are not really fully integrated with the head means the variation in word order,
while still a piece of data that must be explained, is not a counterexample for
the right-headed nature of noun+adnominal pairs.
Real counterexamples would be expressions of the type N+XP-(r)en/-ko,
where the genitive bearing nominal follows the head noun, and these do occur;
however, most are old and found in lexicalized expressions of legal or religious
origin where Romance influence is expected, as in aita pontekoa ‘godfather’
(lit. ‘father of the [baptismal] font’), aita besoetakoa ‘godfather’ (lit. ‘father of
the arms’), aita zerukoa ‘father of heaven’, aingeru guardakoa ‘guardian angel’
(Span. ángel de la guarda), aita gurea ‘our father’ (as in the prayer, Span. Padre
nuestro), etc. Similarly, corresponding to Spanish last names with a toponym
preceded by de, earlier texts or forms fixed in songs often give sequences like
Peru Abendañoko, Span. Pedro de Abendaño, Matxalen Busturiko ‘Magdalen of
Busturia’; these, however, are not productive at present.
While complement and adjunct PP s are prenominal in Basque and post-
nominal in Spanish, their position relative to each other is more flexible. The
most neutral ordering is the expected one, namely adjunct-complement-head in
17 See Berro & Fernández (to appear). They show that these constructions may be found in
titles or as complements of a predicative head, but not in the normal nominal functions
such as subject, etc.
26 ortiz de urbina
18 Euskaltzaindia (1985: 243) points out that preposed adjectives can be found in literature,
giving examples like justu manamenduak ‘fair commands’, venial faltak ‘venial sins’ or
enganoso mundua ‘deceitful world’. The type of adjective and expression, hardly vernacu-
lar, may indicate this is/was a stylistic device for elevated register.
word order 27
ity hinges on these specific suffixes, rather than on a general semantic type,
for instance, indicates the positional properties may stem from properties of
the affixes themselves. Be that as it may, the restrictive versus non-restrictive
interpretations obtained in well-known postnominal and prenominal order-
ings of adjectives in Spanish phrases like nieve blanca/blanca nieve ‘white snow’
or incas valientes/valientes incas ‘bold incas’ does not have any structural cor-
relate in Basque, where both interpretations can be associated with the basic,
postnominal position of the adjective: lagun maiteak ‘dear friends’ receives
an unmarked non-restrictive interpretation (probably for pragmatic reasons),
while lagun minak ‘close friends’ receives a restrictive interpretation. We will
briefly dwell on the three types of adjectives which may be preposed in Basque
in the remainder of this subsection.
Beginning with -(t)ar derived items, de Rijk (2008: 127) explicitly treats them
as nouns: “Nouns derived with this suffix freely combine with other animate
nouns, acting as a first or as a second component of a bipartite compound”,
exemplifying with erromatar soldadua/soldadu erromatarra ‘the Roman sol-
dier’. Correspondingly, he translates mundutar ‘earthly’ as earthling, zerutar
‘celestial, heavenly’ as ‘celestial being’ or ezkertiar as ‘leftist, left-winger’. But
at least some -(t)ar derivatives seem to be used as bona fide adjectives, shar-
ing nonetheless the positional properties of better-behaved items: soñu eztitsu
ta zerutarra ‘a sweet and celestial tune’ (Bilbao, Ipuin Barreka 158), zerutar ta
jainkotiar gauzetaz ‘on celestial and devout things’ (Gazt MusIx 166).19 Even
if most -(t)ar derivatives were actually nouns and we were dealing with com-
pounds, however, we would probably be transferring the problem to morphol-
ogy: erromatar soldadua ‘Roman soldier’ may be equivalent to Erroma soldadua
‘Rome soldier’, a right-headed compound or an apposition, but it is more diffi-
cult to analyze soldadu erromatarra, the alternative order, as headed by a noun
erromatarra: this expression corresponds to Spanish el soldado romano, rather
than something like appositive el romano soldado. It is of course possible that
some of these derived words have become adjectives, but that only nouns can
be found in prenominal position, so the basic line of de Rijk’s analysis may per-
haps still be maintained.
Basque grammars (Euskaltzaindia 1985: 244; Trask 2003: 138) often mention
that loaned adjectives of origin like frantses ‘French’ share with the native
19 These are literary uses, where the derived item has a qualifying function; the preposed
order here may be taking a free ride on the preposing capacity of -(t)ar forms, produc-
ing a pattern similar to preposed qualifying adjectives in Spanish. See below for a short
discussion on qualifying and classifying uses of preposed adjectives.
28 ortiz de urbina
adjectives built on -(t)ar like frantziar the ability to appear preposed, so that
both frantses/frantziar mutilak ‘French youths’ are acceptable, as well as post-
posed mutil frantsesak/frantziarrak. But, leaving aside the preference for post-
posed orders, this is far from a general fact: many adjectives do not sound
felicitous in prenominal position (??aleman/*suizo/*espainol mutilak ‘German/
Swiss/Spanish youths’), and frantses is by far the most common adjective of
this group. Like nafar ‘Navarrese’ (not a loanword), these are found in some
lexicalized expressions like frantses belar ‘alfalfa’ (lit. French grass), nafar hiz-
kuntza ‘Navarrese language’ (i.e., Basque), or nafar-ilar ‘Navarrese pea, gar-
banzo bean’,20 where it is not clear the first element is always an adjective,
rather than a noun.21 Certainly, euskal ‘Basque’ as in euskal dantzak ‘Basque
dances’ is a nominal bound form of euskara ‘Basque language’, and as such can-
not even be found following the noun.22
Turning now to the case of words formed with the suffix -dun, the standard
analysis identifies it as a reduced form of the finite relative verb duen ‘that has
…’, so that a form like euskaldun jendea ‘Basque (speaking) people’ would actu-
ally be something like ‘people who have the Basque language’. The prenominal
order would follow directly from the fact that relative clauses are prenomi-
nal in Basque. Postnominal orders would then have to be explained, probably
as analogy with normal adjectives. As de Rijk (2008:503) points out, however,
there is little evidence for a finite form doubling as a suffix, at least in Basque,
while there is solid evidence for verbs as second elements of exocentric com-
pounds being grammaticalized as derivational suffixes. This would be the case
of -gin ‘-er’ (<egin ‘make, do’), -garri ‘-able’ (<ekarri ‘bring’) and, in this case
-dun ‘-ed’ (<*edun ‘have’). If this is so, then, the ability for words bearing this
suffix to appear prenominally must now be explained, since -garri adjectives,
for instance, do not share it.23
20 The -ar sequence in nafar ‘Navarrese’ belongs to the stem, although its similarity with the
suffix -ar may also be connected to its preposability.
21 Perez Gaztelu et al. (2004: 158) explicitly analyze these as exocentric compounds, although
the categorial identification of the first element is not questioned.
22 That is, something like *dantza euskarak or *kirol euskarak ‘Basque sports’ does not seem
to be possible, although Trask (2003: 139) points out that Eastern varieties admit expres-
sions like liburu euskara bat ‘a Basque [language] book’, specifically referring to the lan-
guage.
23 If the widespread analysis of this suffix as akin to the relative finite form reflects not
just an analysis by linguists but a reanalysis by speakers, we would expect prenominal
forms to be less common in earlier texts. As far as I know, this has not been researched so
far.
word order 29
24 While roughly similar numbers of OV languages are NRel or RelN (with varying propor-
tions depending on the sample used), VO languages with RelN are exceedingly rare: 1 out
of 61 in Dryer’s 1992 sample and 5 out of 421 in Dryer (2013). Mandarin Chinese (a language
with other OV features such as prenominal genitives) is the clearest example.
25 Cinque actually classifies Basque as a rigid OV language; the fact that it is not turns out to
be actually helpful for his analysis, in view of the data discussed in the text.
30 ortiz de urbina
and at present are still occasionally found in eastern dialects. The following
examples illustrate the more standard prenominal relative (a) and the post-
nominal substandard one (b):
More frequent in older texts, but following more closely Romance relative
structure and rarely crossing the barrier into the spoken language, is a postnom-
inal non-restrictive relativization strategy with overt relative pronouns (i.e.,
special uses of otherwise interrogative pronouns) as in the following example
from Leizarraga’s 1571 translation of the New Testament:
The relative clause is introduced by the pronoun zein ‘which’, identical to the
interrogative, and is marked with the complementizer bait-. In contrast, (17b)
displays the typical gap strategy of Basque relatives, varying only in the left
position of the antecedent; since case marking is attached to the end of the
constituent, the absolutive plural determiner -ak that is attached to the postce-
dent in a standard Basque relative like (17a) appears in (17b) at the end of the
relative, attached to the complementizer -n. The Romance type of postnomi-
nal relative is, however, less interesting from the point of view of ordering, since
its position may be accounted for by the fact that it is a non-restrictive paren-
thetical element, rather than by the type of phrase internal relations we are
examining here. More interesting is the type of relative clause which appears
to the right of and in an appositive relation with the antecedent, as in (19), from
de Rijk (2006: 488):
26 A very similar sentence is given in de Rijk (2008: 489) as an example of restrictive appos-
itive, although in fact a negated existential verb intervenes here between the antecedent
and the relative:
(i) Harginik ez da edaten ez duenik
stonecutter.part not is drink not aux.comp.part
‘There is no stonecutter who does not drink’
As discussed in the following text, these could also be analyzed as extraposed relatives.
32 ortiz de urbina
27 Compare with English: the??/a man came to see me who I had met the day before.
word order 33
PP complements and adjuncts to adjectives are far less common28 and can be
attached directly to the Adjective head without any other intervening gram-
matical element which may rule its position relative to the Adjective. In con-
structing these complex AdjPs, while the position of adjective complements
and adjuncts with respect to their head is quite flexible, there is a preference in
more neutral contexts for a head final order:
28 As Goenaga (in progress) points out, there are relatively few complement-taking adjec-
tives in Basque, especially, as will be seen below, in attributive contexts.
29 Comparatives and superlatives are realized as inflectional suffixes on the adjective (han-
diago ‘bigger’, handien ‘biggest’), so their ordering is not considered here.
34 ortiz de urbina
tive, but, unlike samar ‘rather’, they also follow the determiner, so that they
seem to be extraposed: compare common handi samarra and marked handia
erabat.
Manner adverbs (biziki, zeharo, izugarri) and PP s (benetan ‘really, truly’)
which modify adjectives receive degree interpretations; they form a relatively
open class, and occupy a preadjectival position; in similar Spanish examples,
on the other hand, degree adverbs precede (just like regular degree words)
while PP s follow, just as regular PP s without degree interpretation: tremen-
damente aburrido ‘terribly boring’, but aburrido en exceso ‘boring in excess’.
As Bosque (1999: 220–223) points out, however, both of them function gram-
matically like degree expressions and are therefore incompatible with regular
degree words: *tremendamente muy aburrido ‘terribly very boring’, *muy abu-
rrido en exceso ‘very boring in excess’. Similarly in Basque *biziki oso ederra
‘extremely very beautiful’, *oso biziki ederra ‘very extremely beautiful’ (with
very as degree for beautiful), *benetan oso ederra ‘truly very beautiful’,30 *oso
benetan ederra ‘very truly beautiful’ (again, very connected with the adjective).
The quantifier asko ‘many’ (hitz asko ‘many words’) is occasionally used as a
degree word with adverbs and adjectives. In that case, it still occurs predomi-
nantly in final position, as seen with the adverbial examples maiz/laster/pozik
asko ‘very often/soon/happy’. The use of asko as a degree word with adjec-
tives, sparsely attested in earlier texts in spite of the fairly common expression
segur asko ‘most likely, lit. very certain’, is currently found in central dialects
(according to Trask 2003: 141 and the OEH entry for asko),31 as in eder askoa
da ‘it is very beautiful’, where askoa bears the agreement marker as a pred-
icative element. We will return to its attributive use in the following sec-
tion.
The extension of asko from its original function as a noun quantifier to a
degree word for gradable adjectives can be related to the ‘across the board’
nature of some Basque quantifiers/degree words in terms of their categorial
selectional properties. As de Rijk (2008: 242) points out, especially in north-
ern dialects words like oso, anitz, biziki, sobera ‘very’ can modify across nouns,
verbs, adjectives and adverbs: oso adiskide ‘very (much a) friend’, oso poztu ‘to
rejoice very much’, oso erraza ‘very easy’, oso errazki ‘very easily’. Some Span-
ish modifiers are also quite flexible in terms of the category of the lexical
element they modify: bastante amigo ‘rather friend’, bastante rápida(mente)
30 Benetan oso ederra is acceptable with a non-degree interpretation of benetan more similar
to ‘in truth’ than to ‘truly’, the interpretation relevant in this paragraph.
31 The use of askoa described here is specifically classified as Gipuzkoan in the OEH.
word order 35
‘rather quick(ly)’, fumar bastante ‘smoke quite a lot’. Muy also extends to nouns,
targeting the prototypical qualities associated with the entity denoted by the
noun (muy amigo, muy gallo) but not to verbs, where mucho is used instead
( fumar mucho).
The Basque quantifier asko ‘many’ has developed in different directions in
terms of the distribution with categories other than nouns with respect to the
equivalent Spanish quantifier mucho/a(s) ‘many’. On one hand, the extension
to adjectives described above (eder askoa ‘very beautiful’) is not paralleled in
Spanish, where mucho cannot occur with adjectives; on the other hand, in
southern dialects, asko ‘many, much’ has virtually displaced oso ‘very’ as a ver-
bal modifier, presumably helped by the fact that its Spanish equivalent muy
‘very’ does not modify verbs, while mucho ‘many, much’ does.
33 The similarity mentioned there with respect to relative clause extraposition with indefi-
nite head nouns in English is intriguing in that there is no apparent extraposition involved
in the Basque examples with simple or complex appositive adjectives.
word order 37
The second example shows that the position is not only to the left of the noun
but, more specifically, at the left periphery of the nominal complex. The last
example shows the similar behavior of the negative polarity degree word batere
‘at all’.34 Moreover, at least oso ‘very’ can be extraposed to the right, as in etxe
handia oso ‘a very big house’ (but not *etxe handia hain/nahiko ‘so big a house/a
rather big house’).
There are also restrictions on the cooccurrence of degree words on the adjec-
tive on one hand and the type of noun determiner on the other (in particular,
its definiteness/specificity). As noted in Saltarelli (1988), Euskaltzaindia (1985)
and de Rijk (2008), degree words like oso/nahiko are often at odds with definite
determiners and demonstratives. This is shared with Spanish:
34 These preposed degree words resemble similar constructions in English like so/how/too/
that beautiful a house, where the degree element seems to pied-pipe the adjective phrase
to a pre-determiner position (in fact, always indefinite a).
38 ortiz de urbina
The difference between the two languages in (a) is smaller than it appears
to be when one bears in mind that the Basque phrase is interpreted as indefi-
nite (but specific in appropriate contexts) rather than definite. As Pastor (2011)
shows, there are also interactions between some degree words and the indefi-
nite determiner in Spanish, since the non-specific interpretation of indefinite
determiners one can find in a phrase like una casa grande ‘a big house’ becomes
less difficult to obtain in (b).
We mentioned in 2.2.2.1 that some degree words follow the adjective, so
they do not intervene between the noun and the adjective if the latter is not
accompanied by any other complement or adjunct. The expected order is then
actually unmarked (N [Adj Deg] Det) and no extraction or preposing can be
found, so we only have etxe handi samar bat ‘a rather big house’, but no *samar
etxe handi bat. One of the degree words that follow the adjective is asko ‘rather,
fairly’ in Gipuzkoan. When the adjective+asko phrase is used attributively, the
degree word ends up in final position of the nominal phrase and it will then be
followed by the case or agreement markers for the latter. Thus, in these dialects,
a phrase like telebisio merke asko ‘many cheap TV s’ with the standard quanti-
fier asko contrasts with telebisio merke askoa ‘a very cheap TV’. The indefinite
quantifier asko in the first phrase occurs without a determiner, as is general
when a noun is quantified by an indefinite quantifier. However, askoa in the
second phrase acts as a degree word to the adjective. This means that the noun
in that phrase is not quantified, so a determiner -a is attached at the end, cor-
responding to the function of the whole nominal phrase in the clause. In the
following examples, this determiner stands for an agreement marker on the
predicate NP in (a) below and for an indefinite determiner in an argumental
position (b,c):
Notice that in the latter the definite determiner actually has an indefinite
interpretation. In neither case is it possible to use the indefinite determiner
bat even though the same phrase with standard degree word oso ‘very’ would
admit it (oso traje dotore bat/*traje dotore asko bat ‘a very elegant suit’).
Degree words can actually be extracted from the whole nominal expression
and placed in emphatic position before the verb. As de Rijk (2008: 243) indi-
cates, this may be found only with copulative verbs like izan, egon ‘be’ or their
transitive counterparts:
Definiteness effects are reinforced under focus extraction; compare the ex-
tracted and basic phrasal patterns:
while 2.3.2 identifies some of the information packaging strategies which can
be seen to be operative in the two languages. The difference between main and
embedded orders is briefly mentioned in section 2.3.3, and weight and heavi-
ness are dealt with in 2.3.4.
35 The other orders are far less common: 11,5% (OSV), 10,6 % (OVS), 8,6 % (VSO) and 3,8 %
(VOS). Notice however that these are raw counts which do not take into account the pres-
ence of topicalizations or focalizations in the samples. SVO orders are also shown to be
more common in the dialect of Ondarroa, as described in Rotaetxe (1978).
word order 41
These preferences have as a consequence that Basque does not display the
types of SV/VS asymmetries in neutral orders which are found in Spanish and
other Romance languages, mostly with unaccusative verbs. The importance of
verb initial basic orders led Contreras (1976) to postulate a basic VSO pattern
in Spanish and has produced an important body of literature which assumes
subjects need not be in Spec of IP in this language or actually correspond to
topics in Clitic Left Dislocation structures when preceding the subject (Alexi-
adou & Anagnostopoulou 1998; Ordoñez & Treviño 1999 among many others).
Thus, indefinite subjects of unaccusative verbs like caer ‘fall’ or morir ‘die’ are
introduced into the conversation more neutrally in the postverbal position (a).
In preverbal position (b), the indefinite phrase seems to receive a marked, con-
trastive topic type of interpretation. With definite subjects, however, the most
neutral position is preverbal (d), rather than postverbal (c):
36 The postverbal position is obligatory for bare subjects (and objects) in Spanish, another
type of situation that does not arise in Basque, where bare nominals are virtually absent,
mostly restricted to N egin verbs like lan egin ‘work, lit. do work’, amets egin ‘dream, lit. do
dream’, etc.:
i. Cayeron rocas sobre el pueblo vs. *Rocas cayeron sobre el pueblo
fell rocks on the village rocks fell on the village
ii. Murieron soldados jóvenes en esa guerra vs. *Soldados jóvenes murieron en esa
died soldiers young in that war soldiers young died in that
guerra
war
iii. *Cigüeñas anidan aquí vs. Aquí anidan cigüeñas ‘Storks nest here’
iv. *Refugiados trabajan aquí vs. Aquí trabajan refugiados ‘Refugees work here’
37 In literary contexts such as novels, direct speech quotations are followed by V-Subject
orders (esan zuen Jonek ‘said John’), even if relatively ‘neutral’ in that neither the quo-
tation nor the subject (or the verb, for that matter) are emphasized. The presence of the
same unexpected order in English may indicate this is a literary stylistic cliché.
38 Michelena (1978) says that, at least in narrative texts, the (X)VO pattern is traditional rather
than innovative.
word order 43
2003 or Elordieta 2013). Further left periphery phenomena affect tensed verbs
in negative clauses, which are attracted to the left. This is more salient where
the verb is complex, since the auxiliary, which usually follows the verb as
expected in an OV language, appears to the left of the main verb, as well as to
the left of the clause (34c), as described in Laka (1994):
The immediate preverbal, left peripheral position is general for wh-words, and
quite general in western dialects as well as the standard language in the case
of foci, whether new information, contrastive or corrective. Thus, contrary to
Spanish, the answer to a wh-question will have narrow focus in the preverbal
position of the answer, as shown in (b) above, whereas this new information
will be postverbal, and probably final in Spanish:
b. Ha salido Juan.
has left Juan
‘Juan has left.’
44 ortiz de urbina
The mirative type of focus in a sentence like (39) would provide unexpected
and surprising new information, in this example perhaps because Angela is
46 ortiz de urbina
known not to like cakes at all or to eat extremely little, or any other contex-
tually defined presupposed knowledge which would make her eating three
slices unexpected. Finally, Quantifier Fronting in (40) does not convey any con-
trastive interpretation for the fronted element. Leonetti & Escandell (2014)
analyze it as an instance of Verum Focus Fronting, with polarity emphasis.
Jiménez-Fernández (2015) reviews the different types of pragmatic situations
where these constructions can be used, as well as some of the syntactic dif-
ferences such as adjacency with the verb, availability in embedded contexts,
incompatibility among different types of fronting etc. As indicated above, these
seem to be relatively marked structures in Spanish, often with overlapping
and unclear features. All of them could (and probably would) be expressed
through the focalization strategy described above for Basque, with verbal adja-
cency, availability in embedded contexts, incompatibility with further focal-
ized material, etc., even though much remains to be done on the fine details in
central and eastern dialects.40 Those that are well-described from the point of
view of prosody, such as the dialect of Lekeitio (Hualde, Elordieta & Elordieta
1994) indicate that, far from being marked sentence types, all sentences must
include a focally accented element in the preverbal position. From this per-
spective, and abstracting away from the actual focalization structures, Basque
can be qualified as a discourse-configurational language in the sense of Kiss
(1998).
There are further focalization strategies in Basque present in the colloquial
language and often branded as resulting from interference with Spanish. Thus,
Etxepare (1998) describes some minor, marked and not fully grammaticalized
constructions which he considers contrastive, but which Irurtzun (2016: 259)
takes as instances of mirative focus. An example is given in (42):
40 Irurtzun (2016: 244), however, considers the standard, preverbal strategy to be available
across all dialects of the language.
word order 47
Irurtzun (2016: 255) considers the Basque corrective structure as more con-
trastive and presuppositional than the standard preverbal focalization strategy,
with an import akin to that of cleft constructions. As in Spanish, a previous
proposition is assumed to be negated, that somebody other than Peter bought
the magazine.
Corrective focalization is found in southern dialects, in very similar ways
for the two languages among the same speakers. In contrast, there is another
‘reinforced’ focalization strategy, found in northern dialects and probably not
connected to Romance patterns, which has been considered ‘standard’ (though
dialectal) since the first descriptions. It differs from the regular preverbal strat-
egy in that the focalized constituent precedes not the whole verbal complex,
but only the auxiliary, which has been separated from the lexical verb:
In this split-verb strategy, the lexical verb need not be final but may follow the
auxiliary or other constituents to its right, probably resulting from whichever
processes are responsible for the apparent dislocation of constituents to the
right, always possible in Basque, as will be discussed below. The split verb strat-
egy, as Irurtzun (2016) points out, contains a focalized element with reinforced
focal features, and Lafitte (1944: 48) claimed it must be rendered in French
with ce que cleft constructions. We can also interpret Altube’s (1929: 15) remark
that the focus preceding the auxiliary acquires ‘a particularly marked relief’ as
pointing in the same direction.
48 ortiz de urbina
(47) a. [San Juan egunean / Jonek / sua] egunkari batez piztu zuen.
Saint John day.on Jon fire newspaper a.with light aux
‘John lit the fire on Saint John’s day with a newspaper.’
(48) (El día de San Juan) Jon encendió la hoguera (el día de San Juan)
the day of S. Juan Jon lit the bonfire
con un periódico (el día de San Juan).
with a newspaper
‘John lit the fire on Saint John’s day with a newspaper.’
It seems then that there is no basic asymmetry in this respect between Span-
ish and Basque, contrary to what Altube claimed. If we turn to typically heavy
elements such as clauses (as opposed to normal nominals), it is fairly clear that
they prefer a rightward position in both languages. Given basic SOV and SVO
orders as terms of the comparison, this preference will in fact be also observ-
able in the postverbal position of object and, quite generally, verbal dependent
clauses in Basque. Thus, in the following examples, the (non-focal) sentential
subject of the transitive verb and of the copulative predicate is less common
in the preverbal position (a) than in the postverbal one (b) in both languages,
just as in the English translation:
object the fact that, unlike a nominal object, they can be postverbal even if
neutral. There is, moreover, some evidence that such postverbal clauses in non-
canonical position ‘count’ as preverbal. Thus, in the Lekeitio dialect described
in detail in Hualde at al. (1994), the preverbal element must bear focal stress,
and in the basic SOV pattern, objects can be interpreted as neutral or focal-
ized, always bearing focal stress. Unlike neutral nominal objects, however,
non-emphatic object clauses are not preverbal but postverbal. In the same
way, whereas sentence initial root tensed verbal forms are usually ruled out in
Basque, first person tensed forms like nago ‘I am [of the opinion], I think’, or
daukat ‘I have [the opinion], I think’ can be found in some dialects with clausal
complements to their right and occurring at the beginning of the sentence:
In any event, subject and object clauses in Basque pattern with Spanish subject
clauses in their ability to be located in the right periphery of the clause with a
neutral interpretation.
There are other contexts where end-weight effects on clauses can be checked
in the two languages independently of the verb. Thus, in ditransitive construc-
tions heaviness can be checked relative to each of the two verbal complements,
while a small clause complement can in turn contain a clausal subject heavier
than normal small clause predicates. We will only exemplify with the first type
of situation. In the following Spanish ditransitive examples, the clausal con-
stituent will occupy the final position regardless of its function (DO in (a) and
IO in (b), which also corresponds to the canonic order):
41 This example is in fact consciously built to sound like an Axular-like sentence, i.e, a sen-
tence in 17th century Basque. On top of the spelling, notice the use of ezen as an intro-
ductory completive marker, added to the final complementizer -ela; only the latter is
harmonic with OV languages, another expected difference between Spanish and Basque.
52 ortiz de urbina
The order in (55b) corresponds to the canonic DO-IO, and only a very heavily
emphatic DO would be able to appear to the right. There is therefore a differ-
ence between the IO-OD orders with clausal and with nominal IO s, as in the
following examples:
(56) a. atribuir [a que haya sabido jugar bien sus bazas] su suerte
b. atribuir al azar su suerte
While the deviation from the canonical order with the short IO in (56b) pro-
duces at most only a mild, presentational type of emphasis, the same devia-
tion over a clausal IO in (56a) would most likely require a heavy, corrective-
like emphasis, since it involves a deviation from both the unmarked DO-IO
order and the expected light-heavy sequencing. Similar heaviness facts can
be found for Basque, modulo the ‘neutral’ final position of the verb and pre-
ferred focalization patterns. Since tensed clauses in this language can only
be found in absolutive positions, the dative constituent will be exemplified
with a nominalized clause (57a). An independent focal element is used to
prevent the DO/IO from being focalized; this also fronts the verb, produc-
ing an S-V left peripheral pattern superficially similar to the Spanish neutral
order:
Again, the heavy clausal constituent is construed as most neutral when it fol-
lows the lighter nominal one, regardless of grammatical function. In the heavy-
light sequences in (57b) and (58b) above, the presence of a preverbal focus bars
the final, light element from receiving a corrective focus interpretation, and
these sentences are therefore less natural than the corresponding ones in (57a)
and (58a).
3 Final Remarks
As indicated at the outset of this article, the OV and VO labels, taken descrip-
tively as sets of expected orderings of certain (mostly) pairs of constituents,
fit relatively well as general characterization of the major word order profile
of Basque and Spanish, respectively. It is at the phrasal level (NP s, PP s, AdjPs,
etc.) that differences are more marked, but even there confluence is occasion-
ally found. At the clausal level, SOV and SVO stand as grammatically neutral,
pragmatically bleached orders, and, at least in the case of Basque, certainly not
as the most common sequence types. Over these, there are numerous informa-
tion packaging mechanisms which account for the majority of actual orders,
and some of these, old and new, also produce similar orderings of clausal con-
stituents in the two languages. Even though Basque does not have the right
peripheral information focus of Spanish, the latter does have as an option a
preverbal focus akin to the general purpose focalization strategy of Basque,
and clause final corrective focus is also increasingly used in Basque. With vir-
tually universal bilingualism with either Spanish or French and, in the western
area, an important mass of new speakers that make up a good portion of Basque
speakers in heavily populated areas, it would be surprising if a certain degree
of grammatical confluence did not take place. The current grammar of Basque
presents therefore an interesting case for language contact studies. In this con-
text, even though this article only aims at presenting the ground description of
the main data which may be relevant for those studies, it may be worth point-
ing out that in the case of word order, this confluence is found in the apparently
54 ortiz de urbina
increasing use of orders which exploit possibilities, traditional or not, which are
compatible with the Spanish grammar of the same users. As a simple example
of the type of word order developments one can find in contemporary Basque
and which will have to be systematically researched, we can consider example
(59), uttered by a young native speaker of a Biscay variety. A variety of factors,
standard some and substandard others, result in a sentential order of the ele-
ments in this sentence closely aligned with Spanish ¿Podemos repetirlos todos?:
The substandard use of semiauxiliary ahal ‘can’ (cf. standard repetidu ahal
doguz), which produces a verbal sequence linearly identical to Spanish and
English poder repetir/can repeat, and the final informational focus as opposed
to a preverbal focalized position of the quantifier combine to produce a more
Spanish-like order of clausal constituents in this example. Probably, the declar-
ative counterpart would still differ greatly from the Spanish structure even for
this type of speaker, so that, in general, the conditions, extent and specific
mechanisms of these accommodation phenomena will have to be examined.
On top of the interest of this line of research for the analysis of Basque, both
from a synchronic and a diachronic perspective, we believe the specific socio-
linguistic situation of Basque vis-à-vis its Romance neighbors can make of this
a potentially fruitful research area for grammar contact studies.
Acknowledgements
This work has been supported by the Spanish Ministry of Economy and Com-
petitiveness (FFI2014–51878-P). I would like to thank Ane Berro and Beatriz
Fernández for multiple comments on this and related issues.
References
Abney, Stephen. 1987. The English Noun Phrase in its Sentential Aspect. Ph.D. disserta-
tion, MIT.
Alexiadou, Artemis & Elena Anagnostopoulou. 1998. “Parametrizing AGR: Word order,
V-movement and EPP-checking”. Natural Language and Linguistic Theory 16: 491–
539.
word order 55
Altube, Seber. 1929. “Erderismos”. Euzkerea 10. [Repr. 1975, Bilbao: Indauchu]
Arregi, Karlos. 2002. Focus on Basque Movements. Ph.D. dissertation, MIT.
Artiagoitia, Xabier. 2002. “The Functional Structure of the Basque Noun Phrase.” In
Xabier Artiagoitia, Patxi Goenaga & Joseba Lakarra (eds.), Erramu Boneta: Festschrift
for Rudolf P.G. de Rijk. Vitoria-Gasteiz: ASJU 73–90.
Bennett, Paul A. 1979. “On Universal 23”. Linguistic Inquiry 10: 510–511.
Bianchi Valentina, Giulio Bocci & Silvio Cruschina. 2014. Focus fronting, unexpected-
ness, and the evaluative dimension. Ms. [Available at http://ling.auf.net/lingbuzz/
002166]
Bosque, Ignacio. 1989. Las categorías gramaticales. Madrid: Síntesis.
Bosque, Ignacio. 1999. “El sintagma adjetival. Modificadores y complementos del adje-
tivo. Adjetivo y participio”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática
descriptiva de la lengua española. Madrid: Espasa. 217–310.
Cinque, Guglielmo. 1993. “A null theory of phrase and compound stress”. Linguistic
Inquiry 24: 239–297.
Cinque, Guglielmo. 2013. Typological Studies. Word Order and Relative Clauses. New
York: Routledge.
Contreras, Heles. 1976. A theory of word order with special reference to Spanish (Vol. 29).
Amsterdam: North-Holland Publishing Company.
Cruschina Silvio. 2012. Discourse-related features and functional projections. New York/
Oxford: Oxford University Press.
De Rijk, Rudolf P.G. 1969. Is Basque an SOV language? Fontes Linguae Vasconum 3: 319–
351.
De Rijk, Rudolf P.G. 1993. Basque Hospitality and the Suffix -ko. In Jose Ignacio Hualde
& Jon Ortiz de Urbina (eds.), Generative Studies in Basque Linguistics. Amsterdam/
Philadelphia. 145–162.
De Rijk, Rudolf P.G. 2008. Standard Basque: A Progressive Grammar. Cambridge, MA:
The MIT Press.
Demonte, Violeta. 1999. “El adjetivo: clases y usos. La posición del adjetivo en el sin-
tagma nominal”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva
de la lengua española. Madrid: Espasa. 129–215.
Dryer, Matthew. 1992. “The Greenbergian Word Order Correlations”. Language 68: 81–
138.
Dryer, Matthew. 2013. “Relationship between the Order of Object and Verb and the
Order of Relative Clause and Noun”. In Matthew Dryer & Martin Haspelmath (eds.),
The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evo-
lutionary Anthropology. [Available online at http://wals.info/chapter/96. Accessed
on 2016-03-02.]
Duguine, Maia. 2008. “Silent Arguments without pro”. In Theresa Biberauer (ed.), The
Limits of Syntactic Variation. Amsterdam: John Benjamins. 311–320.
56 ortiz de urbina
Elordieta, Arantzazu & Bill Haddican. 2016. “Strategies of verb and verb phrase focus
across Basque dialects”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Micropa-
rameters in the Grammar of Basque. Amsterdam/Philadelphia: John Benjamins. 221–
242.
Elordieta, Arantzazu. 2001. Verb movement and constituent permutation in Basque. Ph.D.
dissertation, Holland Institute of Generative Linguistics/Leiden University.
Elordieta, Arantzazu. 2013. “On the Relevance of the Head Parameter in a Mixed OV
Language”. In Theresa Biberauer & Michelle Sheehan (eds.), Theoretical Approaches
to Disharmonic Word Order. Oxford: Oxford University Press. 306–329.
Escandell-Vidal, Victoria & Manuel Leonetti. 2014. “Fronting and irony in Spanish”.
In Andreas Dufter & Alvaro S. Octavio de Toledo (eds.), Left Sentence Peripheries
in Spanish: Diachronic, Variationist and Comparative Perspectives. Amsterdam: John
Benjamins. 309–342. [doi: 10.1075/la.214.16esc]
Etxeberria, Urtzi. 2005. “Contextual Restriction and Quantification in Basque”. In Svet-
lana Vogeleer (ed.), Bare Plurals, Indefinites and Weak-Strong Distinction. Belgian
Journal of Linguistics vol. 19. Amsterdam/Philadelphia: John Benjamins.
Etxepare, Ricardo & Myriam Uribe-Etxebarria. 2008. “On negation and focus in Spanish
and Basque”. In Xabier Artiagoitia Beaskoetxea & J. Lakarra Andrinua (eds.), Gra-
matika jaietan: Patxi Goenagaren omenez. Bilbo/Donostia: UPV/Gipuzkoako Foru
Aldundia. [ASJU supplement LI]. 287–310.
Etxepare, Ricardo & Jon Ortiz de Urbina. 2003. “Focalization”. In Jose Ignacio Hualde
& Jon Ortiz de Urbina (eds.), A Grammar of Basque. Berlin/New York: Mouton. 459–
516.
Etxepare, Ricardo. 1998. “A case for two types of Focus in Basque”. In Elena Benedicto,
Maribel Romero & Satoshi Tomioka (eds.), UMOP 21: Proceedings of the workshop of
Focus. Amherst: GLSA, 65–81.
Euskaltzaindia [Academy of the Basque Language]. 1985. Euskal Gramatika. Lehen
Urratsak I. Iruñea: Euskaltzaindia.
Euskaltzaindia [Academy of the Basque Language]. 1999. Euskal Gramatika. Lehen
Urratsak V. Bilbo: Euskaltzaindia.
Fernández, Beatriz & Ibon Sarasola. 2010. “Marinelei abisua: izen ondoko datibo sin-
tagmak izenburuen sintaxian” [A warning to sailors: postnominal dative phrases in
the sintax of nouns]. Lapurdum, 14:55–75.
Goenaga, Patxi. In progress. Euskal gramatika: egiturak eta osagaiak [Basque Grammar:
Structures and Constituents]. Euskaltzaindia.
Hidalgo, Bittor. 1995a. “Ohar estatistiko garrantzitsuak euskararen hitz ordenaren
inguru. Euskara S.O.V.?”. Fontes Linguae Vasconum 70:401–420.
Hidalgo, Bittor. 1995b. Hitzen ordena euskaraz [Order of Words in Basque]. Ph.D. disser-
tation, University of the Basque Country.
Hualde, Jose Ignacio, Gorka Elordieta & Arantzazu Elordieta. 1994. The Basque Dialect
of Lekeitio. Bilbo: Universidad del País Vasco. [ASJU Supplement XXXIV]
word order 57
Irurtzun, Aritz. 2016. “Strategies for argument and adjunct focalization in Basque”. In
Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the Grammar of
Basque. Amsterdam/Philadelphia: John Benjamins. 243–263.
Jiménez-Fernández, Angel L. 2015. “Towards a typology of focus: subject position and
microvariation at the discourse-syntax interface”. Ampersand 2:48–60.
Kayne, Richard. 1984. “Unambiguous Paths”. In Connectedness and Binary Branching.
Dordrecht: Foris Publications. 129–164.
Kayne, Richard. 2017. “Antisymmetry and Morphology. Prefixes vs. Suffixes”. Ms.
Kiss, Katalin É. (ed.). 1998. Discourse Configurational Languages. New York/Oxford:
Oxford University Press.
Lafitte, Pierres. 1979 [1944]. Grammaire Basque. Navarro-Labourdin Litteraire. Donostia:
Elkar.
Laka, Itziar. 1994. On the Syntax of Negation. New York/London: Garland Publishing Co.
Lehmann, Winfred P. 1975. Proto-Indo-European Syntax. Austin, TX: Univ of Texas Press.
Leonetti Manuel & Victoria Escandell. 2014. “Fronting and verum-focus in Spanish”. In
Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages.
Amsterdam/Philadelphia: John Benjamins. 155–204.
Michelena, Luis & Angel Yrigarai. 1955. “Nombres vascos de persona”. ASJU II: 107–
127.
Michelena, Luis. 1977. “Notas sobre compuestos verbales vascos”. Revista de dialectolo-
gía y tradiciones populares XXXIII: 245–271 [Repr. in Joseba Lakarra & Iñigo Ruiz
Arzalluz. 2011. Luis Michelena. Obras Completas VII. Donostia: Universidad del País
Vasco. [ASJU Supplement LX]
Michelena, Luis. 1978. “Miscelánea filológica vasca”. Fontes Linguae Vasconum. 205–228.
Ordoñez, Francisco & Esthela Treviño. 1999. “Left dislocated subjects and the pro-drop
parameter: A case study of Spanish”. Lingua 107:39–68.
Ortega, Ivan. 2016. Focus Related Operations at the Right Edge in Spanish: Focus and Ellip-
sis. Amsterdam/Philadelphia: John Benjamins.
Ortiz de Urbina, Jon. 2002. “Focus of correction and remnant movement in Basque”.
In Xabier Artiagoitia Beaskoetxea, Patxi Goenaga Mendizabal & Joseba A. Lakarra
Andrinua (eds.), Erramu Boneta. Festschrift for Rudolf P.G. de Rijk. Donostia: Univer-
sidad del País Vasco. 511–524. [ASJU Supplement XLIV]
Ortiz de Urbina, Jon. 2003. “Word Order”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A Grammar of Basque. Berlin/New York: Mouton. 448–459.
Oyharçabal, Bernard. 2003. “Relative Clauses”. In Jose Ignacio Hualde & Jon Ortiz de
Urbina (eds.), A Grammar of Basque. Berlin/New York: Mouton. 762–823.
Pastor, Alberto. 2011. “Degree Modifiers, Definiteness, and Specificity”. In Luis A. Ortiz-
López (ed.), Selected Proceedings of the 13th Hispanic Linguistics Symposium. Somer-
ville, MA: Cascadilla Proceedings Project.
Pérez Gaztelu, Elixabete, Lluisa Gràcia & Igone Zabala (eds.), 2004. Estructuras morfo-
58 ortiz de urbina
1 Introduction
Whereas the morphology of finite forms in Basque differs greatly from that of
the Romance languages, in the organization of the Tense/Aspect/Mood (TAM)
system, on the other hand, we find interesting similarities, some of which
are likely due to language contact. These similarities are particularly appar-
ent when we consider diachronic processes of grammaticalization. Thus, it has
long been noticed that some of the analytical or periphrastic constructions of
Basque have a direct corresponding structure in Romance, suggesting parallel
developments in Basque and the neighboring Romance languages (Michelena
1981; Trask 1997: 237; Gómez & Sainz 1995). There are, however, important dif-
ferences as well.
When comparing the present-day TAM system in Basque and Romance lan-
guages like Spanish and French, we notice the following common features:
(a) Regarding mood, there is an indicative/subjunctive contrast, in addition
to conditional and imperative forms. (In Basque there are some addi-
tional forms, see below).
(b) Regarding tense distinctions, in the indicative, there is a future/present/
past contrast, whereas in the subjunctive only present and past are dis-
tinguished. (In Spanish there is an obsolescent future subjunctive, which
did not exist in Latin. In French, the past subjunctive is literary).
(c) Regarding aspect, there is a morphological contrast between perfective
and imperfective forms, but only in the past indicative.
Differences between Romance and Basque are most apparent in the morphol-
ogy of verbal forms:
(a) In the Romance languages, all verbs have both simple and compound
forms. In Basque most verbs only possess compound forms.
(b) In Basque all transitive verbs and a class of intransitive verbs (unerga-
tive verbs) take the auxiliary ‘have’ and all other intransitive vebs (unac-
cusative) take the auxiliary ‘be’. In French, most verbs take ‘have’ and
only a small class of intransitives are conjugated with ‘be’ in their com-
pound forms. Finally, in modern Spanish ‘have’ has been generalized as
an auxiliary for all verbs (and has been lost as a main verb). The choice
‘the day is gone’ (El Cid). In Modern Italian, a relatively large class of intransi-
tive verbs—known as unaccusative verbs—still take ‘be’, cf. It ho comprato un
libro ‘I have bought a book’ (transitive), ho lavorato ‘I have worked’ (unergative)
vs. sono arrivato/-a ‘I have arrived’ (lit. ‘I am arrived, masc/fem.’, unaccusative).
Notice that when the ‘be’ auxiliary is used, the participle shows gender (and
number) agreement with the subject.
The development of the present perfect created a contrast that did not exist
in Latin, where the perfect had both values of simple past and present per-
fect; e.g.: Lat amāvī ‘I loved; I have loved’ vs Sp amé ‘I loved’, he amado ‘I have
loved’. In French the contrast in meaning was later lost again, with the disap-
pearance of the simple past from the spoken language. In present-day French,
the form deriving from the Latin perfect (i.e. the passé simple, e.g. Fr j’ aimai)
has been relegated to the written language and nowadays only the compound
form (passé composé, e.g. Fr j’ai aimé) is used both as a present perfect and as
a perfective past in the spoken language.
Another major development in the verb system from Latin to Romance,
which is relevant for the comparison for Basque that we want to make here,
was the loss of the synthetic future (e.g. Lat cantābō ‘I will sing’) and its replace-
ment with a periphrasis of obligation formed with the present of ‘have’ and
an infinitive. This periphrasis was later morphologized as a single word; e.g.:
Lat cantāre habeō ‘I have to sing’ > OSp cantar (h)e > Sp cantaré ‘I will sing’
(see, e.g., Fleischman 1982). The transitional phase can be observed in the fol-
lowing examples from El Cid: doblarvos he la soldada ‘I will double your salary’
(ModSp os doblaré el sueldo); El Campeador dexar las ha en vuestra mano ‘The
Campeador will leave them in your hand’ (ModSp las dejará en vuestra mano)
The same construction, but with the past imperfect of ‘have’ gave rise to the
Romance conditional; e.g.: Lat cantāre habēbam ‘I had to sing’ > OSp cantar ia
> Sp cantaría ‘I would sing’.
With these Romance diachronic facts in mind, let us now consider the struc-
ture of verbal forms in Basque.
The nonfinite forms used in the conjugation of verbs include the perfective,
imperfective and prospective participles, in addition to the radical. The per-
fective participle is the most basic form of the verb in the sense that, knowing
the perfective participle, it is possible to derive all other forms. The nonfinite
forms of erosi ‘to buy’, erori ‘to fall’, eman ‘to give’ and hartu ‘to take’ are shown
in Table 3.1.
The prospective participle bears a suffix -ko that can be identified, at least
diachronically, with the relational suffix, which, among other functions, serves
to form the locative genitive as in mendi-ko ‘of the mountain’. In the standard
language, this suffix -ko competes with -en for verbs whose perfective partici-
ple ends in -n. In northeastern dialects, however, the suffix -(r)en is also used
with other verbs; e.g. harturen, erosiren instead of hartuko, erosiko. From a
diachronic perspective, the latter is the same suffix that forms the regular gen-
itive when attached to noun phrases, e.g. mutil-a-ren ‘of the boy’, mutil-en ‘of
(the) boys’. Both suffixes -ko and -en may have benefactive or destinative value.
Although these usages of these suffixes are restricted in the modern language
(e.g. bihar-ko ‘for tomorrow’), they were prevalent in earlier centuries, which
may explain their use in the prospective participle (Mounole 2011: 312–315).
Consider the examples in (1), from B. Etxepare (1545):1
1 Abbreviations used in the glosses in this chapter: loc = locative, gen = genitive, rel = rela-
tional.
tense, aspect and mood 63
been confined to the subjunctive and other non-indicative moods) and other
periphrases, now obsolete, which we will consider below (see section 5.4).
Given this, it seems sensible to consider the possible role of language contact
in the development of the system of forms illustrated in Tables 3.2 and 3.3.
The present perfect form erosi dut ‘I have bought’ has the same structure as
in the Romance and Germanic languages, as it combines a perfective participle
with the present tense of the auxiliary.
In both Peninsular Spanish and Basque, the present perfect, in addition to
its perfect meaning, functions as a hodiernal or recent past; e.g. Esta mañana
he visto a Mikel/ Gaur goizean Mikel ikusi dut ‘I saw Mikel this morning’ (vs Ayer
vi a Mikel/Atzo Mikel ikusi nuen ‘Yesterday I saw Mikel’). In dialects spoken in
France, this form can also be used as a non-hodiernal past. Although the latter
development mirrors that in French, it appears that it is due to an indepen-
dent evolution, as Mounole (2011) argues, since the widening of meaning took
place in those Basque varieties in the 17th century, before it did in French, and,
furthermore, the contact Romance language at the time was not French but
Gascon, which did not undergo this change.
On the other hand, notice that erosi nuen, which combines the perfective
participle with the past of ‘have’ does not have a pluperfect meaning, ‘I had
bought’ (Sp había comprado), but, rather it is a simple perfective nonhodiernal
or remote past ‘I bought’.
Both tenses with the perfective participle must have developed from resul-
tative structures through the same grammaticalization path as we find in Ro-
mance and Germanic (see Aldai 2001; Mounole 2011). Nevertheless, some struc-
tural differences should be noted: in Basque there is no singular/plural agree-
ment between participle and direct object, not even in the earliest texts (from
the 15th–16th centuries), when the structure was clearly biclausal and had a
resultative reading (Obviously, there cannot be gender agreement, as Basque
lacks this morphological category). The explanation would be that, when the
tense, aspect and mood 65
structure arose, the use of the definite article had not yet been extended to
predicative constructions (morphological plurality developed from the gram-
maticalization as articles of singular and plural demonstratives, Manterola
2015).
In fact, in 16th century texts, the form we are exemplifying with erosi nuen
could function both as a perfective past and as a pluperfect (Mounole 2011:
55). The change in meaning from pluperfect (‘I had bought’) to perfective past
(‘I bought’) is thus a recent innovation, without a direct parallel in Romance,
although along a well-attested grammaticalization path (Aldai 2002: 2017–218):
pluperfect > pluperfect & perfective nonhodiernal past > perfective nonhodier-
nal past; cf. also the evolution of the Latin pluperfect indicative to past subjunc-
tive in Spanish, e.g. Lat amāveram ‘I had loved’ > Sp amara ‘I loved, subjunctive’.
In Old Basque we also find a structure erosi duket ‘I will have bought’, where
the auxiliary bears the suffix -ke. This form has a future perfect interpretation
and may also express probability (see section 3.4).
Nowadays, to express a pluperfect meaning (i.e. a past of the past), as op-
posed to a remote past, Basque makes use of a “double-compound past perfect”
(Aldai 2002), as in erosi izan nuen ‘I had bought’, where izan is the radical and
past participle of ‘be’ (there is also a parallel double-compound present perfect,
erosi izan dut ‘I have bought’, which usually expresses the experiential perfect).
Double-compound forms have a formal parallel in the formes surcomposées of
French and Occitan (Mounole 2011: 47). A meaning close to that of the pluper-
fect can also be expressed by means of a resultative construction, erosia nuen
~ erosita neukan ‘I had it bought’ vs erosia dut ~ erosita daukat ‘I have it (in the
state of being) bought’, which also has a parallel in Spanish (e.g. las tengo com-
pradas ‘I have bought them, fem.’).
The forms with the prospective participle, erosiko dut ‘I will buy’, erosiko nuen
‘I would buy’, also appear to share a diachronic origin with the corresponding
tenses in Spanish and most other Romance languages (Michelena 1981). The
dialectal alternation between suffixes in Basque, e.g. erosiko dut ~ erosiren dut
‘I will buy’ supports this identification between the Romance and Basque struc-
tures to a certain extent. There is a more exact parallel with a construction
of obligation in Spanish with the preposition de ‘of’, e.g. Sp he de comprar ‘I
have to buy’ (although this Spanish construction retains a meaning of obli-
gation and has not been grammaticalized as a future). Nevertheless, notice
that the Basque prospective participle is built on the perfective participle. A
literal Spanish translation of erosiko dut would be *he de comprado. The par-
allel, thus, is not perfect. It is more likely that the grammaticalization of the
suffixes -ko and -en in the prospective participle actually arose from their use
in phrases with benefactive or destinative value (Mounole 2011: 312–315). Struc-
66 hualde and mounole
table 3.4 Modern Spanish simple and compound TAM forms and their Basque counterparts
tures like, e.g., bihar-ko dut ‘I have (to do) it for tomorrow’ may have provided
the original model for erosi-ko dut.
The Basque forms with the imperfective participle do not have a Spanish
parallel. The fact that the imperfective participle incorporates a locative suf-
fix suggests a similarity with, e.g., Eng I am on buying > I’m a-buying. In this
respect, we may note that Gascon has a similar progressive construction with
‘be’ and the infinitive preceded by the preposition a/en, e.g. Gasc èste a fer quau-
còm ‘s/he is doing something’ (Haase 1992: 93), (cf. also Port estou a comprar
‘I am buying’). If the Basque present indicative form is the result of a simi-
lar process of grammaticalization starting from a progressive construction, it
may originally have taken the ‘be’ auxiliary with all verbs, with a later shift
in auxiliary for transitive and unergative verbs, following the choice of aux-
iliary in other tenses (erortzen naiz ‘I am falling, I fall’ and also *erosten naiz
> erosten dut ‘I am buying, I buy’) (see Mounole 2008, 2011). Western Basque
also has a periphrasis with egon ‘be, stay’ for all verbs (e.g. erosten nago ‘I am
buying’) that more closely resembles the Spanish progressive (e.g. estoy com-
prando).
In Table 3.4, a comparison of modern Spanish and Basque indicative forms is
offered, arranged according to Spanish morphology, with simple forms before
compound forms. As can be observed, in spite of the diachronic parallels noted
above, in the present-day languages the only forms that show agreement in
their morphology are those of the present perfect.
Table 3.5 offers the same comparison between modern Basque and Spanish
forms, but using the basic forms of the indicative in Basque (excluding double-
compound forms) as point of comparison and pointing out parallel evolutions.
Forms that are identical in structure and meaning are indicated with an equal
tense, aspect and mood 67
table 3.5 Modern Basque TAM forms and their Spanish formal counterparts
sign, whereas the “approximately” sign (≈) is used for less perfect equivalencies
(with the caveats noted in the text above).
To conclude this section, the most striking difference between Basque and
Romance is the complete lack of synthetic forms in Basque for most verbs, so
that even for the present tense a compound form must necessarily be used. Also
strikingly different is the consistent choice in Basque between the ‘be’ and the
‘have’ auxiliary, depending on the semantics of the verb, across all compound
tenses.
When the contemporary languages are considered, a comparison between
Basque and Romance indicative forms reveals an exact match only in the struc-
ture of the present perfect. When the diachronic evolution of the compound
forms of Basque is also considered, other similarities with Romance develop-
ments become apparent, including in the form of the perfective past, whose
structure mirrors the Romance pluperfect, and, to some extent, the future and
the future of the past or conditional.
Regarding the distinctions that are expressed morphologically, we notice the
existence of present, past and future forms, with a further contrast in both
Romance and Basque between perfective and imperfective in the past. This
restriction of the contrast between perfective and imperfective to the past
tense is cross-linguistically common (Dahl 1985; Bybee et al. 1994:83).
3 Other Moods
3.1 Subjunctive
In both Basque and Romance, we find a contrast in mood between indicative
and subjunctive in subordinate clauses. In Basque, the subjunctive is formed
with the radical and a present or past form of a distinct auxiliary verb. As in the
68 hualde and mounole
3.3 Conditional
In Spanish and French the morphological conditional is used in the apo-
dosis of hypothetical conditional sentences, whereas past subjunctive or past
70 hualde and mounole
3.4 Potential
An aspect in which the Basque TAM system differs markedly from Romance is
in the existence of a set of potential forms. Potential forms bear a suffix -ke or
-teke and are based on the same auxiliary as in the subjunctive. In Romance
potential forms are periphrastic constructions with ‘be able’ and the infini-
tive.
The suffix -(te)ke of potential forms is clearly the same suffix that is found
in the conditional apodosis. In older texts and (rarely) in present-day literary
usage, this suffix may also attach to the indicative auxiliary to indicate either
future time or probability, e.g. ikusten dugu ‘we see’, ikusten dukegu ‘we may see,
we will see’ (also ikusi dukegu).
tense, aspect and mood 71
Basque can also express the potential in periphrases with ahal ‘to be able’
and ezin ‘not to be able’; e.g. erosi ahal dut ~ erosten ahal dut ‘I can buy it’ (with
the perfective or the imperfective participle depending on the dialect, see sec-
tion 6).
As has been mentioned, a handful of Basque verbs deviate from the patterns
illustrated above in that, besides compound or analytical forms, they also pos-
sess synthetic forms. In present-day Basque, in addition to the two auxiliaries
of the indicative, which also function as main verbs with the meanings of ‘have’
and ‘be’, the following verbs have synthetic forms for most speakers:
a) With intransitive morphology: joan ‘go’, etorri ‘come’, ibili ‘walk’
b) With transitive morphology: eduki ‘have’, jakin ‘know’, ekarri ‘bring’, era-
man ‘take’, *io(n) ‘give’
A few other verbs have a more restricted use as synthetic, limited sometimes to
a few forms, e.g. erion ‘flow’, jardun ‘be engaged in’, iraun ‘last, endure’, etzan ‘lie’.
These verbs possess a synthetic present and past indicative and, in some
cases, also conditional, potential and imperative forms; e.g. etorri ‘come’: dator
‘s/he comes, is coming’, zetorren ‘s/he came, was coming’, baletor ‘if s/he came’,
letorke ‘s/he would come’, betor ‘may s/he come’, zetorkeen ‘s/he could come’,
zatoz ‘come!’; jakin ‘know’: dakit ‘I know’, nekien ‘I knew, used to know’, baneki
‘if I knew’, nekike ‘I would know’.
Referring to the contrast between synthetic and compound forms, as in
dator ‘s/he is coming (continuous aspect)’ vs etortzen da ‘s/he comes (gener-
ally or habitually’), Trask (1997:212) writes: “What is striking is that, in contrast
to nearly all other European languages which have developed such an aspec-
tual contrast, in Basque it is the synthetic form which is marked for continuous
aspect […] and this in spite of the fact that the periphrastic form is etymologi-
cally transparent as involving the locative of a verbal noun”.
The online Basque grammar Sareko Euskal Gramatika lists 26 verbs with syn-
thetic forms, but most of the forms listed are not commonly used by a majority
of speakers. In the oldest texts, from the 16th and 17th centuries, we find a much
larger number of verbs with synthetic forms and also additional synthetic TAM
forms (Lafon 1944; Mounole 2011). The progressive loss of synthetic forms does
not seem to be directly attributable to contact with Romance. Influence from
Romance should have resulted instead in a generalization of synthetic forms
for all verbs. What we see in this development is, instead, the application of a
universal tendency to replace synthetic forms with periphrases. This tendency
72 hualde and mounole
operated in Romance, for instance, in the loss of the Latin synthetic passive
and the synthetic future, as well as the recent loss of the passé simple in spoken
French.
As other researchers have noticed, however, all verbs that are or have been
synthetically conjugated in historical Basque belong to the class whose par-
ticipial forms and radical take the prefix e- (Lakarra 2017:77), so that the exten-
sion of the synthetic conjugation to other classes of verbs could not easily
follow an established pattern.
5.2 Resultative
Besides the compound present perfect, Spanish has developed a new resulta-
tive construction with the new ‘have’ verb tener (as mentioned, haber has been
lost in Spanish as a main verb indicating possession and only functions as an
auxiliary and in the ‘there be’ construction). In this resultative construction,
the participle agrees with the direct object in gender and number, cf he escrito
las cartas ‘I have written the letters’ vs tengo escritas las cartas ‘I have the letters
written’ (see Harre 1991). Basque also has a parallel construction, which differs
from the present perfect in that the participle takes the article -a(k), cf gutunak
idatzi ditut ‘I have written the letter’ vs gutunak idatziak ditut ‘I have the letters
written’.
In addition, Basque possesses an experiential construction of identical
structure, with an inflected perfective participle, but with the ‘be’ auxiliary.
This experiential construction can be used with both intransitive and transitive
verbs, e.g. Parisen egona naiz ‘I have been in Paris’, ‘I have had the experience
of being in Paris’ (vs Parisen egon naiz ‘I have been in Paris’), gutun asko idatzia
naiz ‘I have written many letters’ (e.g. ‘I have the experience of having written
many letters’), gorriak ikusia naiz ‘I’ve had a hard time’, lit. ‘I am (having) seen
them red’ (Villasante 1980: 152; Zabala 2003: 431).
Sentences with an inflected participle are perhaps to be analyzed as bi-
clausal constructions, with a larger domain of applicability, beyond their resul-
tative and experiential functions (Ortiz de Urbina & Uribe-Etxebarria 1991).
Basque surpasses Romance in the flexibility of participial predicates.
With the same meaning and functions, Basque also has other structures
where the past participle bears the suffix -ta (probably from the conjunction
eta ‘and’, Krajewska 2013) or -rik (partitive, probably form an earlier ablative
usage of this suffix, Mounole 2008: 291–292) and inflected forms of egon ‘be,
stay’ and eduki ‘have’, e.g. nekatuta/nekaturik nago ‘I am tired’, ikusita/ikusirik
daukat ‘I have seen it’.
The closest Romance parallel is offered by the avoir besoin de ‘need, must,
have need of’ construction in French: j’ai besoin d’un téléphone ‘I need a tele-
phone’, j’ai besoin d’acheter un téléphone ‘I need to/must buy a telephone’ (or
It avere bisogno). Notice, however, that in Basque the object of behar appears in
the absolutive case, not in the genitive, showing a higher degree of grammati-
calization as a verbal construction than in French. In fact, behar ‘need’ can also
be inflected and be used as a regular noun, in which case it takes a genitive
complement: telefono baten (erosteko) beharra dut ‘I have the need of (buying)
a telephone’. In present-day Basque, then, we must distinguish between the ver-
bal construction behar + ‘have’ and the noun behar ‘obligation, need, work’.
Another indication of a high degree of grammaticalization of the verbal con-
struction behar + ‘have’ is that in this construction the word behar can take the
ending -ko of the prospective participle directly. As an alternative, however, the
prospective participle of izan may also be added: telefonoa erosi beharko nuke
~ telefonoa erosi behar izango nuke ‘I should buy the telephone’. On the other
hand, in this construction, behar cannot bear the imperfective suffix (irrele-
vantly, the denominal verb behartu ‘oblige, force’ conjugates like a regular verb).
Both Spanish and French make use of deber/devoir as a modal verb with
deontic (obligation) and epistemic (probability) value, like English must,
should: Sp debes comprarlo, Fr tu dois l’acheter ‘you must buy’, Sp debe haber
caído, Fr il doit être tombé ‘It must have fallen’. In Spanish, to express epistemic
meaning this construction competes with the future (possibility in the present)
and conditional (possibility in the past). As already mentioned, Basque also has
this usage.
To express obligation or need, French and Spanish make use of some addi-
tional constructions. French possesses a construction with falloir ‘be neces-
sary’ used as an impersonal verb with a subjunctive complement, e.g. il faut
que je l’achète ‘I must buy it’. This construction has no parallel in present-day
Spanish or in Basque. On its part, Spanish has constructions with haber de +
infinitive and tener que + infinitive (e.g. he de comprarlo ~ tengo que comprarlo
‘I must buy it’). Above we mentioned the similarity between the haber de con-
struction of Spanish and the Basque future. An older Spanish construction with
caler ‘need’ (lit. ‘to be hot’) is now obsolete in this language (but alive in Judeo-
Spanish and Catalan).
With a parallel structure to the behar ‘need’ + ‘have’ construction, Basque
possesses another construction with nahi expressing the notion of ‘want’. This
construction has resulted from the grammaticalization of the uninflected noun
nahi ‘desire’ with ‘have’. In some Western varieties gura (< Lat gula) is used
instead of nahi. Whereas in standard Basque behar and nahi always take the
‘have’ auxiliary, a further step in the grammaticalization process, found in some
tense, aspect and mood 75
local varieties, is an alternation between ‘have’ and ‘be’ depending on the main
verb; e.g. joan nahi ba duzu > joan nahi bazara ‘If you want to go’, etorri behar du
> etorri behar da ‘s/he must come’ (see Mounole 2010). Here as well we see that
in its tendency to expand the lexically-determined alternation between ‘have’
and ‘be’ across verbal forms, Basque has been moving in the opposite direction
from Romance.
2 Historical examples in Spanish are taken from the Royal Spanish Academy’s CORDE database.
76 hualde and mounole
del dicho pasadizo va cerrado con celosías (Luis Cabrera de Córdoba 1599–1644)
‘Part of that passageway is closed by lattices’. The less frequent corresponding
Spanish expression with llevar ‘carry’ and a past participle also has a resultative
meaning: como llevo declarado ‘As I have already declared’.
The ‘go’ +infinitive construction, which, like in English, has given rise to a
periphrasis with future value in Spanish, French and Portuguese (Sp voy a can-
tar = Fr je vais chanter = Port vou cantar ‘I am going to sing’) has also a parallel in
Basque: kantatzera noa, where the verbal noun takes the allative suffix (but in
some Basque varieties, kantatzen noa). Incidentally, this construction has been
grammaticalized with perfective past value in Catalan, Cat vaig cantar ‘I sang’,
and in some varieties of Gascon, Gasc son mari va revenir à l’ ostel ‘her husband
went back to the hotel’ (Rohlfs 1977: 217–218).
Basque uses a set of particles for modality and aspectual purposes. These par-
ticles are usually placed before the finite verb, i.e. before a synthetic verb or
between main verb and auxiliary in analytical constructions.
There is only one aspectual particle, ohi, which expresses habitual aspect
and has been documented in all dialects since the earliest texts. In Western
dialects it is used in the ‘perfective participle + izan/*edun construction’ and
in Eastern dialects in the ‘imperfective participle + izan/*edun construction’,
instead. The construction with ohi does not have the same status in both dialec-
tal areas, due to different paths of grammaticalization. Whereas in Eastern
dialects ohi is clearly an aspectual marker, in the Western area it behaves as
a semiauxiliary (Ortiz de Urbina 2003: 321–323).3
The use of ohi is not necessary to express habituality. As mentioned above,
Basque, like the Romance languages, can use imperfective forms to refer to
habitual actions, e.g.: goizeko bederatzietan etortzen da/zen ~ goizeko bederatzi-
etan etorri ohi da/zen ‘s-he usually comes/s-he used to come at nine in the
morning’.
The other preverbal particles of Basque belong to the sphere of modality.
Among them, we find ahal ‘desiderative, epistemic modality, yes/no questions’,
bide ‘apparently’, omen ‘hearsay information’ and ei ‘hearsay information’.
3 There are, for instance, syntactic differences in negative sentences; e.g. Eastern etortzen ohi
da ‘s-he usually comes’→ ez ohi da etortzen ‘s-he usually does not come’, where ohi is fronted
with the auxiliary and negative ez ‘not’, vs Western etorri ohi da → ez da etorri ohi.
tense, aspect and mood 77
that, the point is that; certainly, of course’). The main clause or the adverb must
gradually have fallen out of use, and then this use of que was extended to all
affirmative sentences (Rohlfs 1977: 206). In Spanish and Romanian as well, the
particle que is very often employed for emphasizing the clause (Sp ¡que voy!),
and for marking negation (Sp ¡que no quiero!, Rum ca no pot ‘I can’t’). Historical
Basque has no particle of this kind. Nevertheless, Gómez & Sainz (1995) estab-
lished a comparison between the 3rd person past prefix z- and the conjunction
eze, proposing that originally it must have appeared as an enuntiative particle
in narration, which was later reinterpreted as a past morpheme.
Gascon displays a few more particles: be (exclamative particle, be parlas plan
lo gascon! ‘you speak Gascon very well’), e (essentially interrogative particle and
conjunction e sabes parlar gascon? ‘Do you speak Gascon?’) and ja (exclama-
tive and affirmative particle, J’ac sabi ‘I know it’). The latter—also present in
Spanish and Catalan—can be compared to the Basque prefix ba-. It is indeed
the case that the Basque particle originates from bai ‘yes’. In Basque the use
of this preverbal particle was extended and has become obligatory when the
finite verb is located at the beginning of the clause (badakit nik, *dakit nik ‘I
know (it)’, badator ‘s/he’s coming’) (Birabent & Salles-Loustau 1989: 73; Rohlfs
1977: 208–211).
7 Summary
Acknowledgements
The first author acknowledges the research funding for this project received
from the Spanish Ministry of Economy and Competitiveness (FFI2014–51878-
P).
References
Aldai, Gontzal. 2002. The grammaticalization of present and past in Basque. Ph.D. thesis,
University of Southern California, Los Angeles, CA.
Allières, Jacques. 1992. “Gascón y euskera: afinidades e interrelaciones lingüísticas”.
ASJU 26(3). 801–912.
Axular, Pedro. 1643. Gero. Bordeaux: G. Milanges. Edition and Spanish translation by
L. Villasante, 1964, Barcelona: J. Flors; Critical edition by B. Urgell, 2015, Bilbao:
Euskaltzaindia & Nafarroako Gobernua.
Bec, Pierre. 1968. Les interférences linguistiques entre gascon et languedocien dans les
parles du Comminges et du Couserans. Paris.
Birabent Jean-Pierre, & Salles-Loustau Jean. 1989. Memento grammatical du gascon,
Pau: Escòla Gaston Febus & Nosauts de Bigòrra.
Bybee, Joan, Revere Parkins & William Pagliuca. 1994. The evolution of grammar: Tense,
aspect and modality in the languages of the world. Chicago: University of Chicago
Press.
Creissels, Denis & Mounole, Céline. 2012. “Non-canonical valency patterns in Basque:
variation and evolution”. 45th Annual Meeting of the Societas Linguistica Europaea
29 August–1 September 2012, Workshop Contrastive studies of verbal valency in Euro-
pean languages.
Dahl, Östen. 1985. Tense and aspect systems. Oxford: Blackwell.
Etxepare, Bernard. 1545. Lingue Vasconum Primitiae. Bordeaux. Facsimile, critical edi-
tion and translation by P. Altuna, 1995. Bilbao: Euskaltzaindia.
Fleischman, Suzanne. 1982. The future in thought and language: Diachronic evidence
from Romance. Cambridge: Cambridge Univ. Press.
Felischman, Suzanne. 1983. “From pragmatics to grammar: Diachronic reflections on
complex pasts and futures in Romance”. Lingua 60(2–3). 183–214
Gómez, Ricardo & Sainz Koldo. 1996. “On the origin of the Finite Forms of the Basque
Verb”. In Hualde Jose Ignacio, Joseba A. Lakarra & Trask Larry (eds.), Towards a His-
tory of Basque. Amsterdam/Philadelphia: John Benjamins. 235–254.
Haase, Martin. 1992. Sprachkontakt und Sprachwandel in Baskenland: Die Einflüsse der
Gaskgonischen und Französischen auf das Baskische. Hamburg: Helmut Buske Ver-
lag.
80 hualde and mounole
Harre, Catherine. 1991. Tener + Past Participle: A case study in linguistic description. Lon-
don: Routledge.
Hualde, José Ignacio, Gorka Elordieta & Arantzazu Elordieta. 1994. The Basque dialect
of Lekeitio. [Supplements of ASJU XXXIV]. Bilbao/Donostia-San Sebastián: Univ. of
the Basque Country Press/ Diputación Foral de Gipuzkoa.
Krajewska, Dorota. 2013. The diachrony of resultative constructions in Basque. Master’s
Thesis, UPV/EHU.
Lakarra, Joseba A. 1996. Refranes y sentencias (1596): Ikerketak eta edizioa. Bilbo: Real
Academia de la Lengua Vasca/Euskaltzaindia.
Lakarra, Joseba A. 2017. “Basque and the reconstruction of isolated languages”. In Lyle
Campbell (ed.), Language isolates. London: Routledge. 59–99.
Ledgeway, Adam. 2011. “Syntactic and morphosyntactic typology and change”. In Mar-
tin Maiden, John Charless Smith & Adam Ledgeway (eds.), The Romance languages,
vol. 1: Structures. Cambridge: Cambridge Univ. Press. 382–471.
Manterola, Julen. 2015. Euskararen morfologiaren historiarako: artikuluak eta erakus-
leak/ Towards a history of Basque morphology: Articles and demonstratives. Doctoral
dissertation, UPV/EHU, Vitoria-Gasteiz.
Michelena, Luis. 1981. “Lengua común y dialectos vascos”. ASJU XV: 291–303. Reprinted
in: Luis Michelena. 2011. Obras completas [Supplements of ASJU VII]. ed. by Joseba
A. Lakarra & Iñigo Ruiz Arzallus, Donostia-San Sebastián & Vitoria-Gasteiz:
Gipuzkoako Foru Aldundia & UPV/EHU. 517–544.
Mounole, Céline. 2007. “Perifrasi zaharra mendebalde eta erdialdeko euskara zahar-
rean: Azterketa kuantifikatiboa eta proposamen berria”. ASJU 41(1): 67–138.
Mounole, Céline. 2008. “Sintaxi diakronikoa eta aditz multzuaren garapena: inper-
fektibozko perifrasiaren sorreraz”. In Xabier Artiagoitia & Joseba A. Lakarra (eds.)
Gramatika jaietan Patxi Goenagaren omenez [Supplements of ASJU LI]. 548–604.
Mounole, Céline. 2010. “Alineazio aldaketak euskararen diakronian”, In Beatriz Fernán-
dez, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: aldakortasun
sintaktikoa aztergai [Supplements of ASJU LII]. 151–170.
Mounole, Céline. 2011. Le verbe basque ancien: étude philologique et diachronique. Ph.D.
thesis, UPV/EHU. ASJU 48 (2014 [2018]). Bilbao: UPV/EHU.
Ortiz de Urbina, Jon. 2003. “Semiauxiliary verbs”. In José Ignacio Hualde & Jon Ortiz
de Urbina, (eds.), A Grammar of Basque, 300–328. Berlin/New York: Mouton de
Gruyter.
Ortiz de Urbina, Jon & Myriam Uribe-Etxebarria. 1991. “Participial predication in
Basque”. In Joseba A. Lakarra (ed.), Memoriae L. Mitxelena Magistri Sacrum, vol. 2,
[Supplements of ASJU XIV]. Donostia-San Sebastián: Diputación de Gipuzkoa. 993–
1012.
Real Academia Española. Corpus diacrónico del español (CORDE). [Available on the
internet: http://www.rae.es]
tense, aspect and mood 81
Rohlfs, Gerhard. 1977. Le gascon: Etudes de philologie pyrénéenne. Tübingen: Max Nie-
meyer Verlag.
Trask, Robert L. 1997. The history of Basque. London: Routledge.
Urgell, Blanca. 2007. “Para la historia del sustantivo verbal en vasco”. In Joseba A. La-
karra & José Ignacio Hualde (eds.), Studies in Basque and historical linguistics in
memory of R.L. Trask/R.L. Trasken oroitzapenetan ikerketak euskalaritzaz eta hiz-
kuntzalaritza historikoaz [ASJU XL]. 921–948.
Villasante, Luis. 1980. Sintaxis de la oración simple. Oñati: Editorial Franciscana Arán-
zazu.
Vincent, Nigel. 1982. “The development of the auxiliaries habere and esse in Romance”.
In Nigel Vincent & Martin Harris (eds.), Studies in the Romance verb. London: Croom
Helm. 71–96.
Zabala, Igone. 2003. “Nominal predication: copulative sentences and secondary predi-
cation”. In José Ignacio Hualde & Jon Ortiz de Urbina, (eds.), A grammar of Basque,
Berlin/New York: Mouton de Gruyter. 426–448.
chapter 4
1 Introduction
Adjectival participles have been argued to be mixed categories, since they show,
at the same time, properties belonging to verbs and to adjectives. On the one
hand, they are superficially adjectives: they denote a state predicated about an
entity, and can occur in attributive (1) or predicative position (2). On the other
hand, they can also be internally verbal, because, in some cases, participles also
imply a previous event that has as a result the state denoted by the participle.
The different types of participles constitute an area of substantive study for the
analysis of the interface between syntax and the lexicon (Wasow 1977) and syn-
tax and morphology (Marantz 2001, 2007; Anagnostopoulou & Samioti 2014),
as well as for analysis of lexical categories and the projection of syntactic heads
like v, Voice and Aspect. Depending on whether an event is implied or not, and
on the acceptability of different event-related and subject-oriented modifiers,
adjectival participles have been argued to be of different types, with differ-
ent structural compositions, particularly involving different verbal layers (e.g.
Anagnostopoulou 2003; Embick 2004; Alexiadou & Anagnostopoulou 2008;
Bruening 2014; Doron 2014). Those without an event implication have been
considered to be built on top of an acategorial Root (Embick 2004). In con-
trast, others may involve, at different degrees, verbal projections like vP, VoiceP
and even a perfect AspP (Alexiadou et al. 2014; Alexiadou et al. 2015, see also
Sleeman 2011, 2014).
In this chapter, Basque and Spanish adjectival (and adpositional) partici-
ples will be addressed, in the light of the literature on adjectival participles.
As in most of the literature following Kratzer (1994, 2000), I will assume that
there is a differentiation between adjectival and verbal participles: adjectival
participles denote a state, whereas verbal participles describe an event. Addi-
tionally, I will also consider that adjectival participles can be phrasal and that
they can also have event implications. With this picture in mind, in this chapter,
I will explore the properties of phrasal and non-phrasal non-verbal partici-
ples in Basque and Spanish, comparing them with verbal ones, and analyzing
whether the presence/absence of these properties is subject to a classification
of different types of non-verbal participles, such as resultant state / target state
participles (Kratzer 2000) and resultative/stative participles (Embick 2004).
Spanish past participles have been studied in a number of works (among
many others, Luján 1981; Demonte 1983; Bosque 1999, 2014; Marín 1997, 2000,
2004ab, 2009; Gehrke & Marco 2014), but there are not many studies about
Basque non-verbal participles (Ortiz de Urbina & Uribe-Etxebarria 1991; Kra-
jewska 2012, 2013; Berro 2017) and none comparing both Spanish and Basque.
Both languages, although being typologically distant, build and use non-verbal
participles in a strikingly similar way: they are alike in building resultative par-
ticiples by means of a stage-level copula (egon in western and central Basque
varieties and estar in Spanish) and an adjectival (or adpositional) partici-
ple. Additionally, as will be shown, both languages show a similar interaction
between lexical aspect and grammatical aspect, with adjectival (and adposi-
tional) participles interpreted in a non-resultative way when the embedded
event is non-dynamic.
On the other hand, they also differ in a number of aspects. For example,
Basque non-verbal participles are morphologically more complex (showing an
additional resultative morpheme attached to the bare participle) and, depend-
ing on the suffix, they can be adjectival and adpositional. Moreover, they are, in
principle, acceptable in more syntactic contexts, given that Basque non-verbal
participles allow temporal and spatial modification of the event underlying the
participle and, if the copula turns into izan ‘be’ (substituting egon), are com-
patible with all kinds of agent arguments. In Spanish, it has been argued that
the insertion of por-phrases (by-phrases) in adjectival participles is restricted
to those that modify an event kind (Gehrke & Marco 2014). However, as will be
shown, at least, in the Spanish spoken in the Basque Country, temporal and spa-
tial modification of the event is acceptable, and por-phrases with wide scope,
strong determiners and discourse referents are also judged grammatical. These
features have been taken as indicators for the presence of verbal projections
in adjectival participles cross-linguistically (Gehrke 2011, 2015; Alexiadou et al.
2014; Alexiadou et al. 2015) and will be considered in this chapter in order to
approach the structural composition of non-verbal participles in Basque and
Spanish.
Apart from the different degrees of eventiveness implied by the participles,
in this chapter, I will also pay attention to the configuration in which the par-
ticiple occurs, particularly, to whether the participle occurs in the attributive or
84 berro
predicative position, and regarding the latter, to whether the copula is intransi-
tive or transitive. As will be seen, these aspects are sometimes related with the
interpretation of the participle and also with the acceptability of event-related
modifiers.
The structure of the chapter is the following: in section 2, the form and
meaning of non-verbal participles in Spanish and Basque will be analyzed,
paying attention to their morphological and clausal composition, their com-
bination with the copula, their category and their position in the clause. In
section 3, the interaction between the lexical aspect and grammatical aspect
is studied, following particularly Bosque (2014), and also addressing the expe-
riential interpretation that Basque adjectival and adpositional participles can
have (Krajewska 2012, 2013ab). In section 4, the acceptability of non-verbal par-
ticiples with different types of event-oriented modifiers is explored, in the light
of the literature on adjectival participles in other languages such as German
and Greek (Kratzer 1994, 2000; Anagnostopoulou 2003; Alexiadou & Anagnos-
topoulou 2008; Gehrke 2011, 2015; Alexiadou et al. 2014; Bruening 2014; Alexi-
adou et al. 2015). In section 5, the restrictions that hold on event-modification
are addressed, particularly following the event token and event kind distinction
introduced in Gehrke (2011 and subsequent work) and discussing Spanish data
that suggest that adjectival participles in Spanish involve event kinds, and not
events that are instantiated (Gehrke & Marco 2014). I will argue that, in Basque
adjectival and adpositional participles, there is an event token, and that that
may also be the case in Spanish varieties spoken in the Basque Country.
Like in English and German, adjectival participles in Spanish have the same
form as verbal participles used in perfect clauses (4), with the participial ending
attached to the verbal Root [+theme vowel]. Particularly, they look very simi-
lar to verbal passive participles, since adjectival participles show gender and
number agreement with the entity they are predicated about (4), just like ver-
bal passive participles (5).1
1 These are the abbreviations used in the glosses: abl = ablative; abs = absolutive; cl = clitic;
dat = dative; erg = ergative; fem = feminine; imperf = imperfect; ine = inessive adposi-
tion; instr = instrumental; loc = locative (used in the locative copula); masc = masculine;
part = partitive; pl = plural; prt = participle; pst = past; rel = relational adposition; res =
resultative suffix; sg = singular.
non-verbal participles in basque and spanish 85
In Basque, the participle used in perfect clauses is headed by the suffix -tu/-
i/-n/-ø, and, like in Spanish, it does not agree with the arguments of the verb.
Differing from Spanish, adjectival participles in Basque are formed attaching
an additional suffix to the participle used in perfect clauses (glossed here as
Res(ultative), see section 2.1): -a.
When the participle is headed by -a, it agrees in number with the entity is pred-
icated about. As will be explained in the following subsection, Basque has other
suffixes to form resultative participles, but unlike -a ending ones, those do not
agree in number with the entity they are predicated about.
86 berro
-Ta and -rik are dialectally distributed: in western varieties, we find -ta headed
participles, whereas in eastern dialects -rik participles are more prominent. In
central varieties, both types of participles are found (Krajewska 2013). -A par-
ticiples, on the other hand, are used in all dialects.
Participles headed by -a are considered in the Basque linguistic tradition
adjectival, because they can occur in attributive and predicative configurations
and because they agree in number with the absolutive noun they are predicated
about (7), just like individual level adjectives (10b).
The suffix -a is homophonous with the determiner meaning ‘a’ or ‘the’, and has
been analyzed as a resultative aspect marker (Oyharçabal 1992) or a predicative
marker (Ortiz de Urbina & Uribe-Etxebarria 1991; Eguren 2012) in the context
of participles. As shown in example (10), -a also occurs heading individual level
adjectives like altu-a-k [tall-a-pl]. Eguren (2012), in this respect, claims that -a,
in the latter context, is the exponent of a Pred(icative) head (Baker 2003).
On the other hand, -ta and -rik headed participles are usually referred to as
adverbial (Rebuschi 1984), given that they cannot appear in attributive posi-
tion and they do not show number agreement with the absolutive argument.
Their historical origin is adpositional. According to Krajewska (2013), -rik has
evolved from the ablative adposition: it started to be used to mark converbs
non-verbal participles in basque and spanish 87
In this respect, -ta and -rik participles replicate the behavior of adpositions.
They are good adnominal modifiers only in the presence of the relational mor-
pheme. This adnominal use of adpositional resultatives is usually referred to as
a relative form (Hualde et al. 1994).
All in all, Basque seems to have adjectival (ending in -a) and adpositional
(ending in -ta and -rik) participles, apart from the verbal participial form (end-
ing in -tu/-i/-n) used in perfect clauses.
Thus, even if the same root (√open) is used in (13ab) and (14ab), the contrasts
found suggest that it is being used in different configurations, and this is sup-
ported by the participial morphology (-ed) only found in the resultative and
not in the stative adjective. In other cases, however, the same participial form
is used in both (e.g. closed).
In Basque, the participial suffixes -a/-ta/-rik can be used to form resultative and
stative participles. In the case of the Root zabal ‘open, wide’, the stative and the
resultative participles are morphologically differentiated. Zabal-ik, consisting
of the Root and the suffix -rik, is used as a stative adjective, whereas zabal-
du-a/-ta/-rik is used in the resultative participle. In this case, the resultative
involves the Root (zabal), the participial suffix -tu (-du after -n/-l) (the one used
in perfect clauses) and the adjectival/adpositional suffixes -a/-ta/-rik.
non-verbal participles in basque and spanish 89
Contrast like this one are attested for a number of other participles such as huts-
ik ‘empty’ vs. hus-tu-rik ‘emptied’, biluz-ik ‘naked’ vs. biluz-tu-rik ‘undressed’,
poz-ik ‘happy’ vs. poz-tu-rik ‘delighted’ and bakarr-ik ‘alone’ vs. bakar-tu-rik ‘iso-
lated’. The first member of the pairs is not usually considered participial, since
it does not involve the participial suffix -tu that is used in perfect clauses. These
stative elements are referred to as stage-level adjectives (Zabala 1993) or stative
adverbs (de Rijk 2008).
In the case of the Roots ireki ‘open’ and itxi ‘close’, in contrast, the same form
is used as the stative and the resultative participles. It must be noted that, in
these cases, the bare participle has the same form as the bare Root.
In Spanish, we find a similar scenario. Like in English and Basque, Spanish has
also stative adjectives, called perfective adjectives (Bosque 1999, Fábregas &
Marín 2015): lleno (vs. the participial llenado), seco (vs. the participial secado),
limpio (vs. the participial limpiado), sucio (vs. the participial ensuciado), vacío
(vs. vaciado) etc. They are defined as follows (Bosque 1999: 185; translation
taken from Fábregas & Marín 2015).
(20) [Perfective adjectives] denote states reached by the entities they are pred-
icated of.
Bosque (2014) argues that when participial forms have counterpart perfective
adjectives of this sort, the post-nominal participial form can only have eventive
(passive) meaning and not resultative, in Embick’s terms.
vs.
vs.
b. La ropa seca
the.fem clothes dry.fem
‘The dry clothes’
According to Bosque (2014), (21a) and (22a) have a passive eventive meaning,
corresponding to their active counterpart (and do not denote the state reached
non-verbal participles in basque and spanish 91
Both the participle and the perfective adjective denote states. The difference
is that in the perfective participle, an event has not necessarily taken place.
Thus, in (23a), there is not an event of filling up, just like in stage-level adjec-
tives headed by -rik in Basque. In relation to this, Fábregas & Marín (2015)
defended the view that perfective adjectives are the exponents of a Result
subevent (syntactically represented in a First Phase Syntax, Ramchand 2008)
and an adjectival phrase (AP). Thus, according to these authors too, perfective
adjectives correspond merely to a stative layer, with no dynamic event. They
do not involve the Process subevent (responsible for providing dynamicity), so
that the Result is simply interpreted as a stative subevent. They must be then
paralleled to Embick’s stative participles, rather than to resultative ones. Fáb-
regas & Marín (2015) support this analysis by means of several criteria, such as
the following:
The verbs dejar ‘to leave’ and its anticausative counterpart quedar ‘to stay’ focal-
ize the result state and are only possible with telic verbs (García Fernández et
al. 2006). According to Fábregas & Marín (2015), dejar and quedar lexicalize,
subsequently, the Initiation + Process subevents and the Process subevent. The
92 berro
Result subevent is, in turn, spelled out by the perfective adjective. Adjectival
participles like limpiada ‘cleaned’ are not acceptable in this context because
they are blocked by the perfective participle. In the cases where the adjectival
participle entails a change of state, they claim that they necessarily lexicalize
Process.
The configuration involving the verbal passive, with the auxiliary ser ‘be’ in
the present tense, has a habitual interpretation, with the event of throwing
the chairs repeated every morning. The structure with the adjectival partici-
ple, in contrast, only asserts that the state denoted by the participle holds at
the moment of the reference time.
Verbal passive participles occur in a monoclausal configuration, where the
main predicate is the participle and the inflected element stands as an aux-
iliary, whereas resultative participles are bi-clausal: the main predicate is the
copula (estar in Spanish), which takes the adjectival participle as its comple-
ment. Regarding Basque, the bi-clausal analysis of adjectival and adpositional
participles is proposed by Hualde et al. (1994) and de Rijk (2008), and is par-
ticularly analyzed and defended in Ortiz de Urbina & Uribe-Etxebarria (1991).
The monoclausal vs. bi-clausal analysis in Basque has been more controversial,
though, precisely because adjectival participles in Basque are acceptable with
ergative marked agent arguments, just like in monoclausal clauses, and because
they can occur with the copular izan ‘be’, instead of with egon ‘(the stage level)
be’.
non-verbal participles in basque and spanish 93
In both Spanish and Basque, apart from using the intransitive copulas (izan ‘be’
and egon ‘(the stage level) be’ in Basque and estar ‘(the stage level) be’ in Span-
ish), adjectival (an adpositional) participles can occur with transitive copulas
(edun and eduki ‘have’ in Basque and tener ‘have’ in Spanish). This configu-
ration has been called possessive resultative (Nedjalkov & Jaxontov 1988) and
it has been attested that, in the diachronic evolution of many languages, the
perfect has evolved from earlier possessive constructions of this sort (among
others, Detges 2000; Pinkster 1987). Actually, the perfect analytic configuration
of modern Basque, with the participle ending in -tu/-i/-n, used to be employed
as a resultative construction in previous stages of the language (Mounole 2011).
In possessive resultatives, the state denoted by the participle is predicated
about the object of the transitive copula, and the subject may or may not be
interpreted as the agent of the event underlying the participle. For instance, in
(29) and (30), the subject of ‘have’ can be interpreted merely as the possessor
or holder of the state denoted by the object, but not as the agent initiating the
event of cleaning.
Similar constructions can be found in many other languages, like English and
French. In these cases, like in Spanish, the non-pronominal object of ‘have’
takes a pre-participial position, instead of its usual post-participial position,
and the participle agrees in number and gender with the object in the case
of French. In these configurations, the subject is interpreted as the posses-
sor/holder but not necessarily as the agent of the event underlying the par-
ticiple.
According to Iatridou (1996), in (33), the entity ‘the letter’ is not the object
of ‘have’ since (33) does not assert a possessive relation between the subject
(1st person singular) and ‘the letter’. This fact is explained considering that the
whole passive small clause is the complement of ‘have’, and consequently, that
there is no thematic relation between ‘have’ and the entity ‘the letter’, as argued
in Iatridou (1996). Additionally, she argues that the agreement relation between
‘the letter’ and the participle in Modern Greek, and in similar configurations
in French (32), must not be regarded as a case of object agreement, but that
it should be analyzed as subject-predicate agreement inside a passive small
clause.
non-verbal participles in basque and spanish 95
Number and gender agreement between the participle and the predicated
theme cannot be taken as a diagnostic to differentiate adjectival participles
from verbal ones. For instance, as shown in section 3.2, verbal passive partici-
ples also agree with the derived subject.
Bosque (2014) considers that both participles depicted in (34) and (35) are
verbal,2 and not adjectival, and points out that the ability of resultative par-
ticiples to occur with secondary predicates (36) and infinitive complements in
causative sentences (37a) or verbal set phrases (37b) support their verbal status.
2 According to Bosque (2014), the difference between (34) and (35) is that, in (34), the participle
is resultative, whereas in (35), it is eventive. Estar is only compatible with resultative partici-
ples.
96 berro
Even Basque adpositional participles (headed by -ta and -rik) can occur as the
complements of verbs like eman ‘to seem’ and of irudi ‘to seem, to look like’
(40). In contrast, bare participles in Basque—those headed by the participial
ending -tu/-u/-n—are ungrammatical in all these contexts. They cannot be the
complement of eman ‘to seem’ and of irudi ‘to seem, to look like’ and, moreover,
they are incompatible with the copulas egon ‘(the stage level) be’ and eduki
‘have’.
non-verbal participles in basque and spanish 97
Degree modifiers are also compatible with Basque adjectival and adpositional
participles (46) (47), which, once again, contrast with bare participles used in
perfect simple clauses (48).
As can be seen, in Spanish, adjectival participles that are modified by muy ‘very’
can also occur with certain por-phrases (44). Interestingly, the por-phrases
occurring in this context modifies the state denoted by the participle, rather
than the previous event (see Alexiadou et al. 2014 and Alexiadou et al. 2015 for
similar cases in Greek).
The adjectival participle without the ergative argument can modify either a DP
(50a) or a bare noun (51a). In the presence of the ergative argument, the partici-
ple can only modify a DP, cf. (50b) (51b). The same contrast arises with indirect
dative arguments. A dative indirect object, participant of the event underlying
the participle, can only be introduced when the adjectival participle modifies
a DP.
(53) a. Gutun bidal-i-a (ikusi nahi dut.) Attributive with bare noun
letter send-prt-res (I want to see)
‘(I want to see) the sent letter.’
The neutral relative order of the adjective, noun and the determiner in
Basque is depicted in (54). As can be seen, adjectives occur between the bare
noun and the determiner.
But there are also instances where the adjective can modify a DP (55b). Let us
compare the example in (55a), with the adjective modifying the bare noun, and
the example in (55b). There is a subtle meaning contrast between them, which
seems to be similar to that obtained between prenominal and post-nominal
adjectives in English.
In the example in (55a), the subject wants a specific car; it picks up an individ-
ual, which is distinguished from its red color. In the example (55b), in contrast,
the subject expresses which property the car he/she wants must have.
The inability of prenominal adjectival participles to take further modi-
fiers finds a correlate in non-derived adjectives. Actually, as shown by Sadler
and Arnold (1994), prenominal and post-nominal non-derived adjectives show
a clear contrast with respect to their ability to take modifiers and comple-
ments.
Sadler and Arnold (1994) point out several interpretative contrasts between
prenominal and post-nominal adjectives: in prenominal position, the reading
of the adjective is characteristic, timeless or defining property of the noun,
whereas in post-nominal position, the interpretation is of a temporary quality
or property. This is particularly observed in the following example:
102 berro
Nevertheless, Sandler and Arnold do not consider that the analysis of post-
positional adjectives can be reduced to that of predicative ones, since, as they
non-verbal participles in basque and spanish 103
In this section, I will analyze the lexical aspect of the verbs involved in resul-
tative participles and how the lexical aspect interacts with the interpretation
of the adjectival/adpositional participle. I will show that, even though verbs
involving an internal argument are the most frequent verbs giving rise to
adjectival/adpositional participles in Basque and Spanish (and generally, cross-
linguistically), unergative verbs and also transitive verbs where the theme of
predication is co-referent with the subject can also be found in this configu-
ration in Basque (and sometimes also in Spanish). As will be explained, the
occurrence of these verbs and interpretations is closely correlated with the
aspectual meaning of the participle.
3 It must be highlighted that, although there are some semantic correspondences between
attributive participles with bare NP s and attributive participles with DP s in Basque, on the
one hand, and pre-nominal and post-nominal adjectives in English on the other, I do not
want to state that they are syntactically parallel, given that attributive participles with DP s
in Basque can be regarded as being really predicative, but post-nominal adjectives in English
do not need to be so.
104 berro
b. Bazkal-du-ta nago
have.lunch-prt-res be.loc.1sg.abs
‘I am had lunch’
is correferent with the subject of the unergative verb embedded in the adjecti-
val participle. These uses involve consuming verbs similar to that in (63): beber
‘to drink’ and comer ‘to eat’.
The verbs beber ‘to drink’ and comer ‘to eat’ underlying the participle in the
example (65) do not accept direct objects in this configuration. For instance,
the example in (66), which includes direct objects, is not acceptable.
Like in the Basque examples, in the Spanish example of (65), the theme of pred-
ication (los niños ‘children’) corresponds to the unergative subject of the verbs
underlying the participles. Subjects of unergative verbs, as well as subjects of
transitive verbs, are considered to belong not the verbal projection, but intro-
duced by a Voice head projected on top of vP/VP (Kratzer 1996 and following
works). This is why, in an analysis where adjectival participles do not involve
external arguments (Levin & Rapapport 1986, Kratzer 1994 2000 among others),
the fact that predicates of this type are found in resultative participle configu-
rations is unexpected (see section 4.2 for a discussion on this aspect). On the
other hand, the ability of these verbs to occur in this configuration could also be
related to the fact that their subjects, apart from being interpreted as agentive
or initiators, can also be argued to undergo a change (see actually Ramchand
2008, where it is argued that the subject of consuming verbs is both an initiator
and an undergoer), and can, therefore, hold the consequent state denoted by
the participle.
as a consequence of a past event. This is the case of the participles derived from
the verbs castigar ‘to punish’ and atrapar ‘to catch’.
In contrast, when the participle is derived from certain activity verbs (e.g. vigi-
lar ‘to guard’, buscar ‘to seek’, perseguir ‘to chase’), extent verbs (e.g. ocupar ‘to
occupy’) or verbs expressing constituency and other similar physical proper-
ties (e.g. formar ‘to constitute’), the meaning of the participle is unbounded:
the participle denotes a state that holds at the relevant evaluation time, but
does not imply a past event.
The state can have started sometime in the past, e.g. in (68a), but this is irrele-
vant for the meaning of the participle, particularly in (68b). The activity verbs
giving rise to the so-called unbounded past participles in Bosque (2014) also
include the so-called D(avidsonian)-states (Maienborn 2005 2007, Rothmayr
2009, Fábregas & Marín 2017, Berro 2015) (e.g. esperar ‘to wait’, gobernar ‘to gov-
ern’): verbs that share properties with both eventive and stative verbs. They are
eventive and, thus, behave like eventive verbs in some tests (they can occur in
the progressive, they can be infinitive complements of perception verbs, they
can be located in time and space etc.), but non-dynamic, and therefore act like
stative verbs in some other aspects (they give rise to an on-going, non-habitual
interpretation in the simple present, they are bad with dynamic adverbs etc.).
Extent verbs (Gawron 2009; Koontz-Garboden 2010) and verbs express-
ing constituency and other physical property are predicates that denote non-
dynamic eventualities, and which alternate with dynamic interpretations (e.g.
non-verbal participles in basque and spanish 107
obstruct in English, Kratzer 2000). For example, the verb ocupar ‘to occupy’ is
ambiguous between a dynamic and a non-dynamic reading:
4 Note, however, that Bosque (2014) argues that unbounded participles are a subtype of resul-
tative participles (opposed to bounded ones). Embick (2004), in contrast, makes a distinction
between resultative participles and stative participles.
108 berro
claimed to be built at the Root level and not to entail the existence of a prior
event (Alexiadou & Anagnostopoulou 2008). However, certain -tos participles
involve the morphological exponents of verbalizers (e.g. -iz, -on, -a etc.), show-
ing that -tos participles can also be derived from verbs.
Despite the presence of verbalizers, -tos participles do not have eventive mean-
ing and cannot be modified by adverbs (72), contrasting with -menos ending
participles (73).
Spanish unbounded participles (Bosque 2014) derived from activity verbs can
be modified by manner adverbs, a fact which shows that even though un-
bounded participles of this type do not imply a past event, they involve the
relevant functional projection that licenses manner adverbs.
In contrast, participles derived from extent verbs and verbs expressing con-
stituency and other physical property behave differently. Bosque (2014) argues,
building on Gawron (2009), that these classes of verbs should be considered
eventive. Nevertheless, it seems that the participial forms derived from these
predicates do not accept easily manner adverbs, unlike participles derived from
non-dynamic activity verbs.
non-verbal participles in basque and spanish 109
If these predicates involved an event, it is not clear why their participial forms
do not license manner modifiers. Moreover, if they were eventive in their non-
dynamic interpretation, they would be grammatical in the progressive, con-
trary to fact:
The -tu suffix only attaches to the Root when the Root is being used as a verb. In
some cases, the Root changes its final vowel depending on its categorial envi-
ronment. Artiagoitia (2004) addresses this phonological change affecting the
final vowel of certain Roots in verbal environments (where those ending in /e/,
/o/ or /u/ turn to /a/) and proposes that the /a/ ending variants have verbal
category, that is to say, involve the projection of a verbal layer.5
(82) -tu derived verbs (de Rijk 2008: 150–152; Artiagoitia 2004: 151)
neke ‘effort, fatigue’ nekatu ‘to get tired’
aipu ‘citation’ aipatu ‘to cite’
zoro ‘crazy’ zoratu ‘to go crazy’
gorroto ‘hatred’ gorrotatu ‘to hate’
errespetu ‘respect’ errespetatu ‘to respect’
When they are headed by -tu, the Roots end in /a/, showing that they are being
used in a verbal environment, namely, in complement position of a verbalizer.
The two non-resultative (or unbounded, in terms of Bosque 2014) adjectival
participles illustrated in (79) and (80) are also examples of this phonological
change. The verb inguratu ‘to surround’ and osatu ‘to constitute’ are derived
respectively from the nominal Root inguru ‘surrounding’ and the adjectival
Root oso ‘whole, complete’. Both Roots change their final vowel to /a/ when
they are used as verbs.
The adjectival participles in (79) and (80) involve the /a/ ending Roots, as
well as the -tu suffix, showing that both participles are built on verbs, and not
directly on acategorial Roots. In other words, these participles are derived from
verbs, and not from Roots. Thus, these pieces of data show that stative partici-
ples (those without event implications) can be derived from verbs, supporting
the claim made in Anagnostopoulou & Samioti (2014), and against the oppo-
sition between stative and resultative participles in terms of the absence or
presence of v, as presented in Embick (2004).
affecting non-high vowels of bisyllabic Roots. In the context of verbalizing derivational mor-
phemes (80), however, it can also affect trisyllabic Roots and Roots ending in the high vowel
/u/, as can be seen in the examples.
112 berro
meaning where the result of the event holds at the relevant evaluation time.
However, in some cases, adjectival (and adpositional) participles in Basque
(and sometimes also in Spanish) can have a broader use, and cover meanings
pertaining to the experiential perfect, as observed for Basque by Krajewska
(2012 2013ab). The experiential perfect can be built on any type of verb. It
asserts that the subject has a certain experience related to the fact that the pred-
icate has held at one or more intervals in the relevant evaluation time (Comrie
1976; Iatridou, Anagnostopoulou & Izvorski 2001/2003 among others). Finally,
as already advanced in section 3.1.2 and 3.1.3, adjectival (and adpositional) par-
ticiples can also have a continuous interpretation (Bosque 2014), also called
persistent or universal (Comrie 1976; Iatridou et al. 2001/2003; Pancheva 2003),
where the predicate holds from a certain point in the past till the relevant eval-
uation time. In the literature, it has been observed that the universal perfect is
obtained from unbounded predicates such as states and progressives. In con-
figurations involving adjectival (adpositional) participles, I have shown that
it arises when the participle denotes a non-resultative state—unbounded in
terms of Bosque (2014) or stative in terms of Embick (2004)
In Kratzer’s account (see also Alexiadou et al. 2015), the participial morpheme
is meaningless in target state participles; its only function is to categorize the
syntactic object as an adjective and to existentially quantify the Davidsonian
argument of a predicate that has additionally a target state.
Embick (2004) subsumes Kratzer’s target state participles into stative par-
ticiples—those that denote a characteristic state, without event implications.
Nevertheless, Alexiadou et al. (2014) argue, following previous work like Anag-
nostopoulou (2003) and Alexiadou & Anagnostopoulou (2008), that target
state participles can have event implications, and consequently, should be con-
sidered to be derived from vP.
With respect to unbounded participles understood as in Bosque (2014)—
those that are interpreted as underived states, without a causing previous
event, see section 3.1.2—, their aspectual interpretation is not similar to the
perfect of result, but rather seems to be closer to the continuous (Bosque
114 berro
2014), also called persistent (Comrie 1976) or universal perfect (Iatridou et al.
2001/2003; Pancheva 2003). The interpretation of the universal perfect arises
when participles are derived from unbounded predicates like states and pro-
gressives, and especially when a durative temporal adverbial like since-or for-
adverbs are used.
The example in (87) asserts that the property expressed by the predicate (living
here) holds since 2016. Thus, the property does not hold at a time following the
completion of the event denoted by the predicate, as in the perfect of result,
but holds at the same interval as the predicate.
The interpretation obtained in adjectival participles derived from non-
dynamic eventualities like vigilar ‘to guard’ or perseguir ‘to chase’ in Spanish,
or extent verbs and verbs expressing physical properties is similar.
This interpretation is highly predominant when the verb embedded in the par-
ticiple is a transitive or intransitive change of state verbs, like in (89) and (90).
Actually, a participle derived from a change of state verb like hondatu ‘break’ is
quite hard with a frequency adverb like askotan ‘many times’.
In contrast, when the participle is predicated about the subject initiating the
event, the experiential or perfect readings of the participle arises naturally. The
experiential interpretation contrasts with the resultative one in that the out-
come of the event does not necessarily hold at the relevant evaluation time.
Rather than the consequent state, it is the event itself that is focused. These are
some examples taken from Krajewska (2013):
In examples (92), (94) and (95), the theme of predication (the absolutive argu-
ment) corresponds to the subject of the previous event (of ‘listening’ in (92), of
‘seeing’ in (94) and of ‘saying’ in (95)), so that the action affects mainly the state
held by the subject (Krajewska 2013). As a consequence of being the subject of
the event (once or more times), the theme of predication accumulates an expe-
rience. The experiential interpretation is particularly salient when the copula is
in the past tense, and moreover when the copula is *edun or eduki ‘have’ instead
of izan ‘be’. The latter configuration corresponds to the possessive resultative
commented in section 2.3. As mentioned, this configuration has been attested
diachronically in many languages as the source of the perfect. For instance, in
the past tense, the configuration has the meaning of a past perfect (Hualde et
al. 1994: 145–146); a past event completed before some other past event.
The fact that these participles in Basque can have experiential interpretation
can suggest that these are really perfect constructions, where the participles
would be verbal rather than adjectival or adpositional. Krajewska (2012 2013ab)
actually claims that Basque -a/-ta and -rik ending participles are in the mid-
dle of the grammaticalization path towards a new perfect. In her diachronic
study, Krajewska mentions that, in the historical evolution of the Basque lan-
guage in the last centuries, the majority of verbs used in the periphrases involv-
ing -a/-ta and -rik participles are telic change of state verbs where the state
is predicated about the object of the event or the subject of an intransitive
event. Nevertheless, in the cases where other type of predicates are used—
non-verbal participles in basque and spanish 117
Participles occurring in predicative position with the copula estar ‘(the stage-
level) be’ can also have this interpretation in certain cases, although it is not a
generalized tendency.
118 berro
In the examples (98) and (99), the participle has experiential interpretation,
rather than resultative, since a car cannot be in a state of being broken many
times and a computer cannot be in a state of being fixed many times. Still, the
experiential interpretation of participles in this context is not always available
(100). It is clear that the experiential interpretation is not as easily obtained as
in perfect clauses. For instance, in a perfect clause in Spanish, we can get an
experiential interpretation when the participle is derived from a stative verb
(e.g. estar enfermo ‘be sick’)
As in Basque, if the participle combines with the transitive copula tener ‘have’,
an experiential interpretation is much easier to obtain.
McFadden & Alexiadou (2010) analyze the structural differences between ex-
perientials and resultatives, particularly in the context of German and older
English have and be auxiliary alternation (see also Alexiadou 2015). They pro-
pose that, in resultative adjectival participles, the verb is headed by a resultative
Asp(ectual) head that introduces a state that is the result of a prior event. AspP
is, in turn, the complement of the copular verb -be-. In experientials, on the
other hand, the inflected element -have- is an auxiliary located in an aspec-
tual Perf(ect) head, which is below T, and which selects an Asp head. In their
account, this configuration has the meaning of a perfect.
4 Event-Related Modification
In this section, I will deal with the modification of the event underlying the
participle. The ability of adjectival participles to accept different types of event-
related modifiers is used in the literature as a diagnostic to find out the struc-
tural composition of adjectival participles; e.g. to test the presence/absence of
a verb phrase (vP), a Voice phrase and also an aspectual (Asp) phrase (see also
section 5.2).
The fact that there are different classes of adjectival participles has been
noted from early on. Starting from Wasow (1977), deverbal -ed forms in English
have been claimed to be formed at different levels of the language architec-
ture. Wasow (1977) proposed that adjectival passives are created in the lexicon,
whereas verbal passives are created in the syntax. Adding a further distinc-
tion to this two-way classification, Kratzer (1994) claimed that adjectival pas-
sives could be classified into two groups: phrasal passives and lexical passives.
According to her, adjectival participles, which in German are distinguished by
the use of the auxiliary sein ‘be’, are headed by an adjectival passive morpheme
that may be merged into the structure at different positions. In phrasal passives,
the adjectivizing head merges to the verbal phrase (VP in her account), whereas
in lexical passives, it merges to the verb (V). As a consequence, the resulting
participle has different syntactic properties, and in this way, she accounts for
the contrasting behavior between the two, for example in allowing manner
adverbials and the ability to undergo -un prefixation. On the one hand, phrasal
adjectival participles allow modifying adverbs (106a)—unlike underived adjec-
tives (106b)–.
120 berro
On the other hand, lexical participles can be negated by the un- suffix and can-
not be combined with manner adverbs. Thus, lexical participles pattern with
underived adjectives and contrast with phrasal participles.
Another test that is used to explore the structural composition of adjectival par-
ticiples concerns the temporal and spatial location of the event embedded in
the participle. If participles are compatible with temporal and spatial adverbs
that locate the event at a time and place that are different from those of the
state, that means that the event is actually instantiated (Gehrke 2011, 2015). A
modifier like recently or a time-frame adverb like three days ago can be used
with this purpose. Gehrke (2011: 246) shows that in German adjectival partici-
ples, recently can only modify the state rather than the event (108), and that the
adverb three days ago is non-acceptable (109) (von Stechow 1998; Rapp 1996,
1997).
As can be seen, only a temporal adverb that modifies the state, like seit drei
Tagen ‘since three days ago’, is compatible with German adjectival participles.
In the next sections, I will argue that Basque and Spanish are more flexible than
German adjectival participles in this aspect, as they are compatible with this
kind of modification.
The compatibility with manner modifiers show that adjectival and adposi-
tional participles in Basque are actually phrasal, derived from a vP which
denotes an event. Spanish behaves similarly, as can be observed in the examples
below, where the adjectival participle combines with a manner adverb ending
in -mente.
The manner modifiers used in the above examples can be argued to be describ-
ing some aspect of the event that is relevant in the state denoted by the par-
ticiple. Actually, in example (112) the initial se nota que ‘it is noted that’ has
been introduced in order to make the sentence more natural. However, in other
cases, the adverb seems to be modifying the event itself, rather than the state.
Contrasting with German, both Basque and Spanish allow temporal and spatial
modification of the event embedded in the participle.
Thus, both Basque and Spanish allow temporal and spatial modification of
the event. In this aspect, they patter with Greek, which, unlike German, also
accepts temporal adverbs like three days ago and spatial modification.
124 berro
Recent studies on English and German have argued that VoiceP can be involved
also in English and German adjectival participles (e.g. Alexiadou et al. 2014;
Alexiadou et al. 2015 for German; McIntyre 2013; Bruening 2014 and Alexi-
adou et al. 2014 for English). In the literature, the absence of the external
argument in adjectival participles has been analyzed using several diagnos-
tics, such as the inability to control into purpose clauses, and the absence of
non-verbal participles in basque and spanish 125
the disjoint reference effect. Nevertheless, the lack of the disjoint reference
effect as a test to prove the absence of an implicit external argument in Ger-
man (Kratzer 1994) has been called into question in McIntyre (2013), Alexiadou
et al. (2014) and Alexiadou et al. (2015). Particularly, it seems that when non-
reflexive verbs are used, adjectival participles give rise to the disjoint effect,
just like verbal passive participles, so that this diagnostic does not prove that
VoiceP is not projected in German adjectival participles. On the other hand, the
inability of adjectival passives to control into purpose clauses in German and
English has also been counter-evidenced in McIntyre (2013) and Alexiadou et
al. (2014).
In the following two subsections, I will provide several pieces of data that
indicate than an (implicit) external argument is also present in adjectival (and
adpositional) participles in Basque and Spanish. Particularly, I will show that
adjectival (and adpositional) participles are compatible with modifiers that
rely on the presence of an agent.
The data provided clearly indicates that Voice is projected in both Basque and
Spanish. Nevertheless, the acceptability of por-phrases in Spanish is appar-
ently more restricted than the ergative subjects in Basque, as noted by Gehrke
& Marco (2014). According to them, por-phrases are grammatical when they
modify the state denoted by the participle, rather than the event (i.e. when the
participle has unbounded interpretation) or, alternatively, when they modify
an event-kind, rather than an event token (see section 5). These restrictions do
not seem to hold in the Spanish variety spoken in the Basque Country, though.
In section 4, I have shown that Basque and Spanish adjectival (and adposi-
tional) participles are compatible with event-modifiers of different sorts. They
accept manner modifiers that describe the event embedded under the par-
ticiple (such as kontu handiz ‘carefully’ in Basque and cuidadosamente ‘care-
fully’ in Spanish), and allow temporal and spatial adverbs that locate the event
at a different time and place from those of the state denoted by the partici-
ple. Additionally, both Basque and Spanish accept agent-oriented modifiers
and are compatible with controlled purpose clauses, facts that point out that
an implicit external argument is syntactically active in Basque and Spanish
adjectival (and adpositional) participles. Finally, explicit ergative arguments
that are interpreted as the agent of the event embedded in the participle are
also allowed in Basque, although these arguments are much better accepted
when the copula is izan ‘be’ and the participial ending is the adjectival -a
(and not -ta or -rik). In Spanish, too, por-phrases are accepted in both attribu-
tive and predicative adjectival participles, a fact which indicates that Voice
is projected within the participle. Nevertheless, as we are going to see in the
next subsections, the occurrence of por-phrases is not apparently totally unre-
stricted.
non-verbal participles in basque and spanish 129
Gehrke (2011 and subsequent work) observes that by-phrases in German are
possible as long as they belong to the consequent state, rather than to the
underlying event, or if they have an impact on the underlying event that is
still visible in the consequent state. She refers to the latter event as an event
kind. In her analysis, adjectival participles denote consequent states of event
kinds. Event kinds cannot be modified by temporal or spatial adverbs because
event kinds lack spatiotemporal manifestation. Thus, the only modifiers that
are accepted are those which describe the consequent state, or otherwise, are
interpreted as kind-related; modifiers that create a new subkind of the event
by narrowing the event kind, and thus, having an impact on the consequent
state.
The type of manner adverbs allowed in Spanish adjectival participles seem
to point in this direction (section 4.1). Manner adverbs like cuidadosamente
‘carefully’, descuidadamente ‘sloppily’ are adverbs that, although they describe
the event, shape also the form of the consequent state. In the case of rápida-
mente ‘quickly’ in (112), the adverb is better accepted with the introduction of se
nota que ‘it is noted that’ at the beginning of the example, another piece of evi-
dence which shows that the relevance of the event-manner in the consequent
state is crucial for the acceptance of this adverb.
Nevertheless, we have seen that Spanish, like Basque, allows spatiotemporal
modification of the event, something that is unexpected if they allowed only
event-kind modifiers.
Finally, the study also shows that the contrasts found between por-phrases that
belong to the event in verbal and adjectival participles do not find a correlate
in the por-phrases that pertain to the consequent state. Below, an example that
shows that a por-phrase modifying a state can have a strong determiner.
5.2 Structural Motivation for the Event Kind vs. Event Token Distinction
Gehrke (2011) and following works (Gehrke & Marco 2014; Gehrke 2015; Alexi-
adou et al. 2014; Alexiadou et al. 2015) claim that adjectival participles in lan-
guages like Spanish and German are interpreted as consequent states of event
non-verbal participles in basque and spanish 131
kinds because in these languages, the verb does not get further embedded into
verbal projections, but is directly adjectivized. They generalize to verbs the
analysis of nominals (Carlson 1977; Zamparelli 1995; Chierchia 1998; Dayal 2004
among others), where it is considered that nouns start out like predicates of
kinds and get instantiated when they are embedded under Number. In the ver-
bal domain, Gehrke (2015) proposes that events also enter the derivation as
event kinds and become instantiated when they are embedded under further
verbal projections, like Aspect or T. At this level, they obtain the interpretation
of event tokens, that is, events that take place in the actual world. Since in lan-
guages like German and Spanish, participles are not further embedded under
Aspect or T, the interpretation of the event remains in the kind domain, and is
not interpreted as an instantiated event. According to the authors, the restric-
tions on event-modification, for example in languages like Spanish, follow from
the fact that the event remains as an event kind. State-modifiers of a state token
are generally accepted, because the adjectivized participle is further embedded
under T, and thus, the consequent state gets instantiated.
6 In Alexiadou et al. (2015), the PASS head is substituted by an a(dj) head that has the catego-
rizing function.
132 berro
6 Conclusions
In this chapter, non-verbal participles in Basque and Spanish have been com-
pared and contrasted, in light of the literature on adjectival participles. I have
shown that non-verbal participles differ in these languages in their morpholog-
ical make-up: while Spanish adjectival participles are formally similar to verbal
passive participles, in Basque, an additional participial suffix (-a, -ta or -rik)
is added. -A participles have been considered adjectival, following the Basque
linguistic tradition, and -ta and -rik ones have been regarded as adpositional.
Non-verbal participles in both languages also differ in that Basque non-verbal
participles can give rise to experiential interpretations, which is why Krajew-
ska (2012, 2013ab) claims that Basque -a/-ta/-rik “resultative” participles are in
the middle of a grammaticalization path towards a new perfect. Additionally,
Basque adjectival participles also contrast with Spanish ones in that Basque
participles allow ergative subjects to be present, which denote the agent of the
event embedded under the participle.
Despite these differences, this chapter has made visible that Basque and
Spanish non-verbal participles are similar in a number of other aspects. For
instance, the same interaction between the lexical aspect of the event underly-
ing the participle and the aspectual interpretation of the participle is observed
in both languages, with non-dynamic events giving rise to a non-resultative
interpretation of the participle. On the other hand, even though Spanish adjec-
tival participles have been argued to admit only certain type or por-phrases, and
to involve, consequently, an event kind (and not an event token), this chapter
has shown that, at least in the Spanish spoken in the Basque Country, adjecti-
val participles allow temporal and spatial modification of the event underlying
the participle and por-phrases with wide scope, strong determiners and dis-
course referents. These data indicate that this language variety may be similar
to Basque in having non-verbal participles with an event token interpretation.
134 berro
Acknowledgements
I want to thank Ane Odria for sharing with me her linguistic intuitions and
Beatriz Fernández for her comments on the manuscript. The research lead-
ing to these results has received funding from the European Union’s Seventh
Framework Programme for research, technological development and demon-
stration under grant agreement no. 613465. Moreover, this study has been
developed thanks to several projects funded by the Basque Government (the
post-doctoral grant (POS_2015_1_0086 and POS_2016_2_0023), and the project
IT665–13) and by the Spanish Ministry of Economy and Competitiveness
(FFI2014–51878-P).
References
Alexiadou, Artemis. 2015. “On the irrealis effect on auxiliary selection”. In Rolf Kailuweit
& Malte Rosemeyer (eds.), Auxiliary Selection Revisited: Gradience and Gradualness.
Berlin: de Gruyter. 123–144.
Alexiadou, Artemis & Elena Anagnostopoulou. 2008. “Structuring participles”. In
Charles Chang & Hannah Haynie. Somerville (eds.), Proceedings of the WCCFL 26.
MA: Cascadilla Press. 33–41.
Alexiadou, Artemis, Elena Anagnostopoulou & Florian Schäfer. 2015. External Argu-
ments in Transitivity Altenations. A Layering Approach. Oxford: Oxford University
Press.
Anagnostopoulou, Elena. 2003. “Participles and Voice”. In Artemis Alexiadou, Monika
Rathert & Arnim von Stechow (eds.), Perfect explorations. Berlin: Mouton de Gruyter.
1–36.
Alexiadou, Artemis, Berit Gehrke & Florian Schäfer. 2014. “The argument structure of
adjectival participles revisited”. Lingua 149: 118–138.
Anagnostopoulou, Elena & Yota Samioti. 2014. “Domains within words and their mean-
ings: A case study”. In Artemis Alexiadou, Hagit Borer & Florian Schäfer (eds.), The
Syntax of Roots and the Roots of Syntax. Oxford: Oxford University Press.
Artiagoitia, Xabier. 2004. “The case of an enlightening, provoking and admirable
Basque derivational suffix with implications for the theory of argument structure”.
Supplements of ASJU XLVI: 147–183.
Baker, Mark C. 2003. Lexical Categories. Verbs, Nouns and Adjectives. Cambridge: Cam-
bridge University Press.
Berro, Ane. 2015. Breaking verbs. From event structure to syntactic categories in Basque.
Ph.D. diss., UPV/EHU and Université Bordeaux Montaigne.
Berro, Ane. 2017. “Resultative and non-resultative participles in Basque”. Ms, University
of the Basque Country & SFL (UMR 7023, CNRS/Paris 8).
non-verbal participles in basque and spanish 135
Bosque, Ignacio. 1999. “El sistema adjectival. Complementos y modificadores del adje-
tivo. Adjetivo y participio”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática
descriptiva de la lengua española. vol. 1. Madrid: Espasa Calpe. 217–310.
Bosque, Ignacio. 2014. “On resultative past participles in Spanish”. Catalan Journal of
Linuistics 13: 41–77.
Bruening, Benjamin. 2014. “Word formation is syntactic: adjectival passives in English”.
Natural Language and Linguistic Theory 32: 363–422.
Carlson, Greg. 1977. Reference to kinds in English. Ph.D. diss., University of Massachu-
setts at Amherst, published by Garland, New York.
Chierchia, Gennaro. 1998. “Generics and concepts”. Natural Language Semantics 6: 339–
405.
Comrie, Bernard. 1976. Aspect. An Introduction to the Study of Verbal Aspect and Related
Problems. Cambridge: Cambridge University Press.
Dayal, Veneeta. 2004. “Number marking and (in)definites in kind terms”. Linguistic and
Philosophy 27: 393–450.
Demonte, Violeta. 1983. “Pasivas léxicas y pasivas sintácticas en español”. Serta Philo-
logica F. Lázaro Carreter. Madrid: Cátedra. 141–157.
Detges, Ulrich. 2000. “Time and truth: the grammaticalization of resultatives and per-
fects within a theory of subjectification”. Studies in Language 24: 345–377.
Doron, Edit. 2014. “The interaction of adjectival passive and voice”. In Artemis Alex-
iadou, Hagit Borer & Florian Schäfer (eds.), The Syntax of Roots and the Roots of
Syntax. Oxford: Oxford University Press.
Eguren, Luis. 2012. “Predication Markers in Basque”. In Urtzi Etxeberria, Ricardo Etxe-
pare & Myriam Uribe-Etxebarria (eds.), Noun Phrases and Nominalization in Basque:
Syntax and Semantics. John Benjamins Publishing Company. 243–266.
Elordieta, Arantzazu. 2001. Verb Movement and Constituent Permutation in Basque.
Utrecht: LOT.
Embick, David. 2004. “On the Structure of Resultative Participles in English”. Linguistic
Inquiry 35-3: 355–392.
Fábregas, Antonio & Rafael Marín. 2015. “How insertion in non-terminal nodes explains
two participle puzzles”. Paper presented in the 25th Coloquium on Generative Gram-
mar. Bayonne, May 21–23, 2015.
Fábregas, Antonio & Rafael Marín. 2017. “On non-dynamic eventive verbs in Spanish”.
Linguistics 55(3).
García Fernández, Luis, Ángeles Carrasco Gutiérrez, Bruno Camus Bergareche, María
Martínez-Atienza & María Ángeles García-Serrano. 2006. Diccionario de perífrasis
verbales. Madrid: Gredos.
Gehrke, Berit. 2011. “Stative passives and event kinds”. In Ingo Reich, Eva Horch & Den-
nis Pauly (eds.), Proceedings of Sinn und Bedeutung 15. Saarbrücken: Universaar—
Saarland University Press. 241–257.
136 berro
Gehrke, Berit. 2015. “Adjectival participles, event kind modification and pseudo-in-
corporation”. Natural Language and Linguistic Theory 33: 897–938.
Gehrke, Berit & Cristina Marco. 2014. The role of by-phrases in adjectival passives. Lin-
gua 149: 188–214.
Goenaga, Patxi. 1978. Gramatika bideetan. Donostia: Erein.
Hualde, José Ignacio. 1991. Basque Phonology. London and New York: Routledge.
Hualde, José Ignacio, Arantzazu Elordieta & Gorka Elordieta. 1994. The Basque dialect
of Lekeitio. Supplements of ASJU XXXIV. Bilbao/Donostia: UPV/EHU, Gipuzkoako Foru
Aldundia.
Iatridou, Sabine. 1996. “To have and to have not: on the deconstruction approach”. In
José Camacho, Lina Choueiri & Maki Watanabe (eds.), Proceedings of the 14th West
Coast Conference on Formal Linguistics (WCCFL) 14. Stanford: CSLI, 185–201.
Iatridou, Sabine, Anagnostopoulou, Elena & Roumyana Izvorski. 2001. “Observations
about the form and meaning of the perfect”. In Michael. J. Kenstowicz (ed.), Ken
Hale: a life in language. Cambridge, MA: MIT. 189–238.
Iatridou, Sabine, Anagnostopoulou, Elena & Roumyana Izvorski. 2003. “Observations
about the form and meaning of the perfect”. In Artemis Alexiadou, Monika Rathert
& Arnim von Stechow (eds.), Perfect explorations. Berlin/New York: Mouton de
Gruyter. 153–204.
Kennedy, Christopher & Louise McNally. 1999. “From Event Structure to Scale Structure:
Degree Modification in Deverbal Adjectives”. In Tanya Mathews & Devon Strolovitch
(eds.), The Proceedings of SALT IX, 1999. Ithaka, NY: CLC Publications. 163–180.
Koontz-Garboden, Andrew. 2010. “The lexical semantics of derived statives”. Linguistic
and Philosophy 33: 285–324.
Krajewska, Dorota. 2012. “Resultatives in Basque: a diachronic study”. Lingua Pos-
naniensis LIV(2): 56–67.
Krajewska, Dorota. 2013a. The diachrony of resultative constructions in Basque. Master
Thesis, UPV/EHU.
Krajewska 2013b. “Euskarazko egitura erresultatiboaren diakronia”. In Ricardo Gómez
et al. (eds.), 3rd Conference of the Luis Michelena Chair. Vitoria-Gasteiz: UPV/EHU Edi-
torial Services. 264–274.
Kratzer, Angelika. 1994. “The Event Argument and the Semantics of Voice”. Ms., Univer-
sity of Massachusetts at Amherst.
Kratzer, Angelika. 1996. “Severing the external argument from its verb”. In Johan
Roorych & Laurie Zaring (eds.), Phrase Structure and the Lexicon. Studies in Natural
Language and Linguistic Theory, 33. Dordrecht: Kluwer Academic Publishers. 109–
137.
Kratzer, Angelika. 2000. “Building Statives”. Berkeley Linguistic Society 26. 385–399.
Levin, Beth & Malka Rappaport Hovav. 1986. “The formation of adjectival passives”. Lin-
guistic Inquiry 17(1): 623–661.
non-verbal participles in basque and spanish 137
Luján, Marta. 1981. “The Spanish copulas as aspectual indicators”. Lingua 54: 165–209.
Maienborn, Claudia. 2005. “On the Limits of the Davidsonian Approach: The Case of
Copula Sentences”. Theoretical Linguistics 31 (3): 275–316.
Maienborn, Claudia. 2007. “On Davidsonian and Kimian states”. In Ileana Comorovski
& Klaus von Heusinger (eds.), Existence: semantics and syntax. Springer. 107–130.
Marín, Rafael. 1997. “Participios con aspecto de adjetivos: entre la diacronía y la mor-
fología”. Moenia 3: 365–376.
Marín, Rafael. 2000. El componente aspectual de la predicación. Doctoral dissertation.
Universitat Autònoma de Barcelona.
Marín, Rafael. 2004a. Entre ser y estar. Madrid: Arco/Libros.
Marín, Rafael. 2004b. “Sobre pasivas adjetivales”. Verba 31: 455–471.
Marín, Rafael. 2009. “Del adjetivo al participio”. In Elena de Miguel et al. (eds.), Fron-
teras de un diccionario: las palabras en movimiento. San Millán de la Cogolla: Cilen-
gua. 327–348.
McFadden, Thomas & Artemis Alexiadou. 2011. “Pieces of the be perfect in German
and Older English”. In Donald Baumer, David Montero & Michael Scanlon (eds.),
Proceedings of the 25th West Coast Conference on Formal Linguistics. Somerville, MA:
Cascadilla Proceedings Project. 270–278.
McIntyre, Andrew. 2013. “Adjectival passives and adjectival participles in English”. In
Artemis Alexiadou & Florian Schäfer (eds.), Non-canonical Passives, Amsterdam:
John Benjamins. 21–42.
Mounole, Céline. 2011. Le verbe basque ancien: étude philologique et diachronique. Ph.D.
diss., University of the Basque Country (UPV/EHU) and Université Michel de Mon-
taigne-Bordeaux III.
Nedjalkov, Vladimir P. and Sergej Je. Jaxontov. 1988. “The Typology of Resultative Con-
structions”. In Vladimir Nadjalkov (ed.), Typology of resultative constructions. Ams-
terdam: John Benjamins Publishing Company. 3–62.
Ortiz de Urbina, Jon & Myriam Uribe-Etxebarria. 1991. “Participial predication in
Basque”. Supplements of ASJU 14-2: 993–1012.
Oyharçabal, Bernard. 1992. “Structural and Inherent Case-Marking; Ergaccusativity in
Basque”. In Joseba A. Lakarra & Jon Ortiz de Urbina (eds.), Syntactic Theory and
Basque Syntax. Supplements of ASJU XXVII. Donostia: UPV/EHU and Gipuzkoako
Foru Aldundia. 309–342.
Oyharçabal, Bernard. 2010. “Basque ditransitives”. In Maia Duguine, Susana Huido-
bro & Nerea Madariaga (eds.), Argument structure and syntactic relations. Amster-
dam/Philadelphia: John Benjamins. 261–282.
Pancheva, Roumyana. 2003. “The Aspectual Makeup of Perfect Participles and the
Interpretation of the Perfect”. In Artemis Alexiadou, Monika Rathert & Arnim von
Stechow (eds.), Perfect explorations. Berlin/New York: Mouton de Gruyter. 277–306.
Pinkster, Harm. 1987. “The strategy and chronology of the development of future and
138 berro
the perfect tense auxiliaries in Latin”. In Martin Harris, & Paolo Ramat (eds.), His-
torical development of auxiliaries. Berlin/New York: Mouton de Gruyter. 193–223.
Parsons, Terence. 1990. Events in the Semantics of English. Cambridge, MA: the MIT Press.
Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First Phase Syntax. Cam-
bridge: Cambridge University Press.
Rapp, Irene. 1996. “Zustand? Passiv? Überlegungen zum sogenannten “Zustandspas-
siv””. Zeitschrift für Sprachwissenschaft 15.2: 231–265.
Rapp, Irene. 1997. “Partizipien und semantische Struktur: Zu passivischen Konstruktio-
nen mit dem 3”. Status. Tübingen: Stauffenburg.
Rebuschi, George. 1984. Structure de l’énoncé en basque. Paris: SELAF.
de Rijk, Rudolf P.G. 1993. “Basque Hospitality and the suffix -ko”. In José Ignacio Hualde
& Jon Ortiz de Urbina (eds.), Generative Studies in Basque Linguistics. Amsterdam:
John Benjamins. 145–162. [reprinted in Rijk, Rudolf P.G. de. 1998. De Lingua Vas-
conum: Selected Writtings. Supplements of ASJU XLIII. UPV/EHU. 377–390]
de Rijk, Rudolf P.G. 2008. Standard Basque. A Progressive Grammar. Cambridge/Lon-
don: The MIT Press.
Rothmayr, Antonia. 2009. The Structure of Stative Verbs. Amsterdam/Philadelphia: John
Benjamins.
Sadler, Louisa & Douglas J. Arnold. 1994. “Prenominal adjectives and the phrasal/lexical
distinction”. Journal of Linguistics 30(1): 187–226.
Sleeman, Petra. 2011. “Verbal and adjectival participles: Position and internal structure”.
Lingua 121: 1569–1587.
Sleeman, Petra. 2014. “From participle to adjective in Germanic and Romance”. In Petra
Sleeman, Freek Van de Velde & Harry Perridon (eds.), Adjectives in Germanic and
Romance. John Benjamins. 171–198.
von Stechow, Arnim. 1998. “German participles II in Distributed Morphology”. Ms., Uni-
versity of Tübingen.
Wasow, Thomas. 1977. “Transformations and the lexicon”. In Peter W. Culicover, Thomas
Wasow & Adrian Akmajian (eds.), Formal Syntax. 327–360. New York: Academic
Press.
Gawron, Jean Marc. 2009. The lexical semantics of extent verbs. Ms., San Diego State
University.
Zabala, Igone. 1993. Predikazioaren teoriak gramatika sortzailean (euskararen kasua).
Ph.D. thesis, UPV/EHU.
Zamparelli, Roberto. 1995. Layers in the determiner phrase. Ph.D. thesis, University of
Rochester.
chapter 5
1 Introduction
1 Abbreviations: abs = absolutive case, cl = se clitic, det = determiner, erg = ergative case,
gen = genitive, imprf = imperfective, ine = inessive, nom = nominative case, part = parti-
tive, pl = plural, prog = progressive, prt = participle, pst = past.
(in)transitive verbs: unergatives and unaccusatives 141
Unergative verbs, in contrast, have an ergative subject and edun (have) aux-
iliary.
(4)
The subject of unaccusatives, on the other hand, is introduced within the ver-
bal phrase.
The majority of the verbs of change of state and change of location are derived,
in the sense that they are built on elements that exist outside of this ver-
bal complex, like nouns (e.g. ama ‘mother’ and mutil ‘boy’), adjectives (gorri
‘red’, handi ‘big’, txiki ‘small’ etc.) in (5), adverbs (hurbil ‘near’, urrun ‘far’)
and adpositional phrases (e.g. etxe-ra ‘to house’ and lehorr-era ‘to land’) in
(6).
Most of them, and particularly derived ones, can have a causative variant, in
which case the causer argument is marked ergative and the auxiliary switches
to have. The ability to enter the causative alternation is considered a diagnos-
tic for unaccusativity in Levin & Rapport Hovav (1995). Another test that has
been suggested to differentiate unaccusatives from unergatives verbs in Basque
involves the use of the partitive case. Only the subjects of unaccusative verbs
(as well as the objects of transitive verbs) can occur bearing partitive marking
(10) (11), contrasting, this way, with the subjects of unergative verbs (and also
the subjects of transitive verbs) (12) (13) (Levin 1983; Ortiz de Urbina 1989; Sal-
aburu 1992).
b. Ez da haurr-ik etorri.
no be.3sg.abs child-part come
‘No child has come’
(in)transitive verbs: unergatives and unaccusatives 143
Unaccusative subjects and transitive objects pattern alike: they can be marked
with the partitive. In contrast, unergative subjects and transitive subjects also
behave in a similar fashion: they cannot bear partitive marking.
(17) Emission
Distiratu ‘shine’, dirdiratu ‘shine, sparkle’, argitu ‘light’, usaindu ‘smell’.
The verbs shown above are all simplex, consisting of a single phonological
word. See, for example, the sentence in (21). Note, however, that many unerga-
tive verbs in Basque are morphologically complex, consisting of the light verb
egin ‘do’ and a non-verbal element (22).
(23) Emission
a. Sound emission
deiadar egin ‘scream’, oihu egin ‘scream’, barre egin ‘laugh’, negar egin
‘cry’, zaunka egin ‘bark’ etc.
b. Light emission
Dir-dir egin ‘shine’, diz-diz egin ‘glow, sparkle’, nir-nir egin ‘twinkle,
flicker’.
c. Verbal emission
hitz egin ‘talk’, solas egin ‘talk, chat’, marmar egin ‘grunt’, dei egin ‘call’,
otoitz egin ‘pray’, errieta egin ‘reprimand, scold’, burla egin ‘mock, make
fun of’ etc.
2 Weather verbs are different from other complex unergatives in a number of aspects. For
instance, in most of them the non-verbal (nominal) element is headed by the determiner
-a and they can be modified by adjectives like handi ‘big/a lot’ as in bero handia egin [heat big
do] ‘be very warm’. I want to thank Jon Ortiz de Urbina for bringing these facts to my attention.
(in)transitive verbs: unergatives and unaccusatives 147
tences and in focalized structures) and the complement may be marked parti-
tive in negative sentences, as common direct objects.
Furthermore, the bare nominal may also be quantified, like common direct
objects, and egin can be silenced in sentences with negative contrastive focus
(Oyharçabal 2006: 792–793):
Note, however, that the bare noun of complex unergative predicates does not
behave as a common direct object in other aspects: e.g. in its lack of determiner
and its position with respect to manner adverbials. See Oyharçabal (2006) for a
discussion on the (non)-incorporation of the noun, and the variation in dialects
and predicates.
Another aspect in favor of the transitivity of complex unergative verbs is
that, unlike simplex verbs, they are incompatible with a direct object (other
than the non-verbal element selected by egin ‘do’).
Many works have accounted for this contrast claiming that in complex unerga-
tives, the complement of egin ‘do’ is assigned absolutive case, so that another
direct object cannot be case-licensed (e.g. Laka 1993; Fernández 1997). In Berro
(2015a), I additionally claim that the non-verbal element in complex verbs
(dantza in (35b)) occupies exactly the same position of the direct object in
simplex verbs (tangoa in (35a)), accounting in this way for the measuring prop-
erties of the direct objects in simplex verbs.3
The system of event configuration that I am assuming in this paper shares many
similarities with that of Mateu (2002 et seq.) and Ramchand (2008 et seq.). In
my system, the verbal phrase consists of subevents—Events and States—and
their non-subeventive complements, namely Roots. Event provides eventivity
to the predicate, introducing an event argument (Davidson 1967). State is a cen-
tral coincidence relation, which relates a Figure (its specifier) with a Ground
3 Simplex verbs can be bounded if the direct object denotes a bounded scale (e.g. 35a). In a
similar way, the event denoted in complex verbs is also measured by the scalar (aspectual)
properties of its non-verbal complement (the Root). See Harley (2005) and Berro (2015a) for
accounts about the measuring properties of Roots (Berro 2015a about Basque in particular).
(in)transitive verbs: unergatives and unaccusatives 149
The different flavors of the subevents, and the theta roles of the subjects are
derived from the whole event configuration (Wood 2012). Event can have a
cause flavor, like in (36), if it is selected by Voice and if it selects for State. It
can also have a do flavor (37), if it is selected by Voice, and if it selects straight-
away for an Event naming Root (see Harley 2005; Berro 2015a). This is the case
of unergative eventive verbs.
Event can also have a transitional (or go, become) flavor, in the case that
it selects for a State (like in (36)), but Voice is not projected above (38). In this
case, also like in (36), the Root selected by State is a State naming Root. This is
the configuration of unaccusative verbs.4
4 Following Mateu (2002 et seq.), I consider that both telic and atelic unaccusatives have the
configuration depicted in (3).
150 berro
The theta-role of the subjects also varies depending on the configuration. When
Voice selects for Event, the subject introduced in its specifier has an originator
theta role, which can be further specified as causer or agent. In a configura-
tion like (36), the subject has a causer theta-role, and in (37), it may have an
agent role, but only if the entity occupying this position is animate and if the
encyclopedic content related to the Root allows for that. On the other hand,
the subject in the specifier of State has a patient role (undergoer and resultee),
because the State is selected by Event. In the case it is not (as in the stative
verbs that I will present in section 6), it has a holder role. As will be shown, the
subject in Voice can also be a holder, if Voice selects for a State instead of for an
Event.
The verbs that are used in the unaccusative fashion in eastern dialects comprise
volitional verbs expressing speech (mintzatu ‘speak’, solastatu ‘chat’), man-
ner of motion (promenatu ‘stroll’, jauzi ‘jump’), dining (bazkaldu ‘have lunch’,
afaldu ‘have dinner’), and animate activities (borrokatu ‘fight’, jokatu ‘play, act’,
jolastu ‘play’). As I have mentioned above, from this group, verbs of manner of
(in)transitive verbs: unergatives and unaccusatives 151
motion, dining and animate activities occur with an ergative subject and edun
‘have’ auxiliary in western and central varieties.
Aldai (2008 2009) suggests that the dialectal divide within Basque corresponds
to the divide between two case-marking systems: one based on the ergative
alignment of case, represented by eastern dialects, and a semantically aligned
one, represented by western and central dialects. However, Aldai’s characteri-
zation of the eastern case-alignment is questioned by the behavior some verbs,
such as inanimate manner of motion verbs (41ab), verbs directly expressing
duration (41c) and emission verbs (41d), which take an ergative subject and
have auxiliary in eastern dialects (see Berro & Etxepare 2017).
b. Honek ez du funtziona-tzen.
this.erg no have.3sg.erg work-imprf
‘This does not work/is not working’
c. Beran-tzen zuen.
delay-imprf have.3sg.erg.pst
‘He/she was taking long’
Therefore, not all unergative verbs are aligned in the unaccusative fashion in
eastern dialects, so that the distinction between eastern and western/central
dialects in terms of a semantically based alignment versus an ergative align-
ment is not as clear as suggested in Aldai’s proposal.
152 berro
Another line of analysis that has been put forward is that the verbs that take
absolutive subjects in eastern dialects are somehow related to an aspectual
head of telicity or boundedness (Berro 2010 2012). This head would be respon-
sible for the absolutive marking of the subject and would also make it possible
for those unergative verbs to become telic or bounded by selecting an incre-
mental theme.
A third type of explanation, developed in Pineda & Berro (2018), is that
these unergative verbs are just associated to different syntactic-eventive con-
figurations in eastern and western/central dialects. It is a well-known fact that
verbs having very close meaning can correspond to different event configu-
rations across languages. For example, the verb ‘blush’ is a change of state in
Italian (arrossire, lit. ‘become red’) and it is a verb of internal causation in Dutch
(bloezen) (McClure 1990, apud Levin & Rappaport Hovav 1995). Since both lan-
guages have auxiliary alternation, in Italian arrossire combines with essere (be)
auxiliary, whereas in Dutch, bloezen occurs with hebben (have), as expected.
On the other hand, in a closer Romance language like Spanish, the verb sonreír
‘smile’ is subject to intra-linguistic variation: although ‘smile’ is usually consid-
ered an unergative predicate (also in Basque, irribarre egin, lit. ‘do smile’), in
some Spanish varieties it occurs with se clitic in non-reciprocal contexts, like
some unaccusative verbs.5
From this perspective, sonreír and sonreírse are the externalizations of two dif-
ferent eventive configurations. Sonreír is an internally caused verb, whereas
sonreírse is a change of state verb, where the subject undergoes a transition
to a state of having a smile on the face, in the same way as the subject of the
Italian verb arrossire. In Basque, too, we find similar cases, for example with
the verbs leher egin and lehertu, which both mean ‘explode’. Leher egin is an
unergative verb of internal causation, and lehertu, in its intransitive variant, is
an unaccusative verb of change of state. These two different eventive configu-
rations are morphologically externalized: the former having an ergative subject
and have auxiliary, and the latter, an absolutive subject and be auxiliary.
5 This dialectal alternation is also attested in other verbs, like in engordar/engordarse ‘fatten’
and adelgazar/adelgazarse ‘lose weigh’.
(in)transitive verbs: unergatives and unaccusatives 153
I propose that in leher egin, the verbal root is merged to an Event head with an
internal cause flavor (do), whereas in lehertu, the Event head has a transitional
flavor (trans).
b. lehertu
The verbs that take an absolutive subject in eastern varieties could also be
involved in an alternation of this sort. Dining verbs like bazkaldu ‘have lunch’
may be associated to a transition to a state of being finished with lunch. For
instance, some speakers judge these verbs as telic in the progressive-perfect
entailment test.
In telic verbs, the progressive does not entail the perfect, and this is actually
what happens in the verb bazkaldu. Telicity could perhaps lead to the change of
state configuration (with Event + state) of this verb in eastern dialects. On the
other hand, other verbs like mintzatu ‘speak’ and borrokatu ‘fight’ could be con-
sidered similar to some pronominal verbs in Spanish and French (expresarse
and pelearse in Spanish, and s’exprimer and se battre in French), which have
a rather reflexive or reciprocal flavor. In French, particularly, these predicates
occur with être (be) auxiliary, as French is an auxiliary alternation language.
Pineda & Berro (2018) propose that several verbs which show dialectal alter-
nation in Basque and that are aligned in the unaccusative fashion in eastern
dialects are unaccusative verbs in (Middle) French and Occitan. These lan-
guages are actually the Romance languages that are in contact with north-
eastern dialects. The unaccusativity of these verbs is supported by several diag-
nostics, such as the use of the se clitic in previous stages of the language,
occurrence in a causative variant, lack of direct object and available resulta-
tive participles (see Pineda & Berro 2018 for further details).
Similarly to what is found in (Middle) French and/or Occitan, the configu-
ration of these verbs in eastern dialects may be unaccusative (see section 6.1
in relation to this). The fact that this behavior is not generalized to all “unerga-
tive verbs” but that it is, actually, quite restricted would also suggest that we are
dealing with certain lexical exceptions.
With this hierarchy, Sorace (2000 2004) accounts for the delimited and sys-
tematic variation found in intransitive predicates with respect to the auxiliary
selected. According to Sorace, some predicates are consistent in showing aux-
iliary alternation across languages like Italian, French, German or Dutch, and
also in language varieties, while others consistently select for be or have, both
synchronically and diachronically. The verbs at the top and at the bottom of
the hierarchy are the ones having least variation and most consistent behav-
ior (either selecting be or have), whereas the ones in the middle are the most
variable.
Regarding the distribution of auxiliaries in unaccusative verbs in French and
Italian, three groups of verbs can be distinguished (Bentley & Eythórsson 2003:
450–451).
Firstly, change of location verbs select be in both Italian and European French
(some of these verbs take have in Canadian French).
156 berro
Sorace (2000) includes, in this group, verbs that are inherently telic. Neverthe-
less, in certain atelic change of location verbs like ‘descend’, French selects for
avoir (have) auxiliary, whereas Italian selects essere (be).
On the other hand, change of state verbs show internal variation, as they can
be classified in two groups: those selecting be in both Italian and European
French, and those selecting be in Italian and have in European French. The
verbs of the former group, that is to say, the ones that occur with be also in Euro-
pean French, are telic (50a) (51a). This means that, in French, atelic change of
state verbs (e.g. grandir ‘increase, become bigger’, viellir ‘become older’) occur
with avoir (have) auxiliary (51b).
b. Gianni è invecchiato.
John.nom be.3sg become.older.prt
‘John has become older’ (Kayne 2008: 1)
c. Jean a vielli.
John.nom have.3sg become.older.prt
‘John has become older’ (Kayne 2008: 1)
Finally, verbs of continuation of a pre-existing state (e.g. ‘last’, ‘survive’ etc.) and
existence of a state (e.g. ‘be’ ‘exist’, ‘belong’, ‘seem’ etc.) select generally be in Ital-
ian (although, according to Sorace 2000, judgments are weaker and have is not
uncommon), and consistently have in European French.
Other verbs of this group, in contrast, select izan (be) auxiliary, like the verb
gustatu ‘like’ (55) and faltatu ‘lack’, and others such as existitu ‘to exist’ can be
found either with izan (be) (57a) or with edun (have) (57b) (selecting edun
(have) particularly in non-standard uses).
As can be seen, Basque intransitive verbs also conform to the hierarchy pro-
posed in Sorace (2000, 2004), as stative verbs are the ones showing more vari-
ation in the selection of have and be auxiliary. In the next section, I will con-
tinue analyzing the variation existing in stative verbs, particularly focusing on
the role of Romance se clitic and its influence on loan verbs in Basque.
Basque is in permanent contact with Spanish and French, as all Basque speak-
ers are these days Spanish-Basque or French-Basque bilingual. In this section,
two consequences of this contact will be analyzed, namely, the use of Romance
loan verbs, particularly intransitive verbs, and the emergence of new uses
of certain stative predicates. In both cases, we will see that the presence or
absence of the se clitic in the original Romance verbs is closely related to the
subject case marking and auxiliary selection of the verbs in Basque.
160 berro
In relation to this contrast, in the dialects that are in contact with Spanish,
dibortziatu ‘divorce’ is used in the “unaccusative” pattern, selecting an absolu-
tive subject and be auxiliary (59a). In contrast, in the dialects that are in contact
with French, dibortziatu takes an ergative subject and have auxiliary (59b).
(59) a. Miren eta Jon dibortziatu dira. Basque (cont. with Sp.)
Mary.abs and John.abs divorce be.3pl.abs
6 It must be noted that se influences the alignment only in the case of intransitive loan verbs.
Uses of se in transitive clauses (comerse la manzana [eat-se the apple]) do not seem to alter
the alignment in Basque.
(in)transitive verbs: unergatives and unaccusatives 161
A reversed, but similar, pattern is attested with the verb ‘stroll’. In French,
when it is intransitive, it is a pronominal verb (i.e. se promener), whereas in
Spanish, it is non-pronominal (i.e. pasear).7 As expected, in the Basque dialects
that are in contact with French, promenatu ‘stroll’ combines with an absolutive
subject and be auxiliary, whereas in the dialects that are in contact with Span-
ish, paseatu occurs with an ergative subject and have auxiliary.
Another verb that shows (dialectal) variation is entrenatu ‘train’. The verb ‘train’
can be used in the pronominal and non-pronominal form in both Spanish (i.e.
entrenar and entrenarse) and French (s’entraîner). In the dialects that are in
contact with French, speakers prefer to use the intransitive entrenatu in the
unaccusative pattern, namely, with an absolutive subject and be auxiliary. In
contrast, in other dialects (central and western ones), the intransitive entrenatu
is used either in the unergative fashion—with an ergative subject and have
auxiliary—or in the unaccusative one. As in the previous cases, I suggest that
these differences in the subject marking and auxiliary selection are the exter-
nalizations of different event configurations.
7 There is also a variant pasearse in Spanish, with the se clitic, but this one has a slightly differ-
ent meaning, namely, ‘stroll without any clear intention’.
162 berro
As can be seen in the examples, these verbs are used in the unergative pat-
tern (with ergative subject and have auxiliary) and they are morphologi-
cally simplex (they do not consist of the light verb egin ‘do’). Remarkably,
these verbs are stative: they denote a property of their sole argument—like
its weigh, length, temperature etc.—and they behave like stative verbs in a
number of linguistic contexts, such as in their incompatibility with process
adverbs and the degree interpretation of the modifier apur bat ‘a little’ (see
Berro 2015b, 2016). The ergative subject of these verbs is not an agent and
causation is not involved, but the verbs are still aligned in the unergative pat-
tern.
Most of these verbs are used standardly in the language, but with eventive
meanings. In order to express these stative meanings, predicative constructions
are used instead:
b. El suelo resbala.
det floor.nom slip.3sg
‘The floor is slippery.’
b. La sopa quema.
det soup.nom burn.3sg
‘The soup burns’ or ‘The soup is burning hot.’
As it can be observed, in the eventive intransitive use of the verbs, the se clitic is
present. In Basque, in the eventive intransitive use, the verb gets realized in the
unaccusative pattern. Compare the eventive unaccusative verbs in (74a) and
(75a) with the stative unergative uses in (74b) and (75b).
Once again, the meaning that has the se clitic in Romance is related to the unac-
cusative pattern in Basque, and similarly, the se-less meaning in Romance is
related to the unergative pattern. Actually, se is ungrammatical in the stative
uses of the verbs in Spanish.8
I propose that, in these cases too, the difference in the subject marking and aux-
iliary selection is derived from different eventive configurations in the syntax.
The unergative stative verbs have the structure in (75a), whereas the eventive
unaccusative verb has the configuration in (75b).
8 In French, there is a dynamic and intransitive use of the verb gliser ‘slide’ which does not take
the se clitic and which requires have auxiliary.
(i) Pierre a glissé sur la glace (et est tombé)
Pierre has slid on the ice (and is fallen)
Pierre slid on the ice (and fell)
Labelle & Doron (2010) claim that, in this case, the root has merged with a dynamic verbal
head (v), instead of with V. As a consequence, the construction focuses on the process, instead
of on the result.
166 berro
b. irristatu [eventive]
In the stative variant, Event is not projected, and the subject is introduced in the
specifier of Voice. There, it is interpreted as a holder argument. Recall that the
subject of Voice is only interpreted as an originator when Voice selects for
Event. On the other hand, in the eventive variant, the subject is introduced in
the specifier of State, and after Event is projected, it is interpreted as a patient
(undergoer/resultee).
In this section, I will propose that stative verbs are the most variable predi-
cates both in Basque and cross-linguistically because their subject—a holder
argument—can be introduced either in the specifier of Voice or in the speci-
fier of State, without that creating an important semantic contrast. This hap-
pens precisely because, in states, the Event head is not projected. Additionally,
I will argue that, in Romance loan verbs, the presence of the se clitic in the
original Romance language forces an unaccusative configuration of the verb
in Basque, because se-variants necessarily project Event + State (Cuervo 2003,
2014).
In section 4.1 and 5.3, we saw that stative verbs are the most variable group
of verbs in terms of auxiliary alternation (and subject case marking) both in
(in)transitive verbs: unergatives and unaccusatives 167
(78) a. [VoiceP Subj Voice [EventP Event [Root]]] Subj = originator, agent
b. [EventP Event [StateP Subj State [Root]]] Subj = patient
In stative verbs, Event is not projected. Thus, the subject of a stative verb can
be introduced either in the specifier of Voice (79a) or in the specifier of State
(79b), without that creating a significant semantic difference. Note that in even-
tive predicates, having an external or internal subject really makes a difference:
if it is external, it is interpreted as the entity originating the event, whereas if it
is internal, it gets the role of undergoing the event and ending in a final state.
Since there is no Event in state verbs, stative verbs can have external or internal
subjects without much semantic contrast.9
9 There is a contrast that, in Berro (2015a), I claim to be one between containing vs. being
included, or have vs. be In unergative stative verbs, the subject contains or possesses the
168 berro
Regarding the role of the se clitic in Romance and its relation with the
Basque unaccusative pattern, I propose that se unaccusatives in Romance
involve the projection of both Event and State, in the spirit of Cuervo (2003
2014) (see also Cuervo 2003 2014 for Spanish, Folli & Harley 2005 for Italian,
and Labelle 1992, Labelle & Doron 2010 for French).
In the literature, se has been related to telicity in different ways: it has been
argued that se is involved in a configuration having or emphasizing a final state
(Labelle 1992, Cuervo 2003 2014, Folli 2002, Folli & Harley 2005, Legendre &
Smolensky 2009, Labelle & Doron 2010) or a bounded path (Basilico 2010).
Cuervo (2003 2014) analyzes unaccusative verbs that have a se and a se-less vari-
ant (e.g. caer(se) ‘fall’, salir(se) ‘come out/off’, morir(se) ‘die’, ir(se) ‘go’ etc.). On
the basis of several systematic interpretive and formal contrasts between the se
and the se-less variant of these verbs (79), she proposes that these two variants
have different eventive structures, i.e. a bi-eventive vs. a mono-eventive config-
uration. Depending on the structure, the Root is interpreted as Manner or as
Result (Mateu & Acedo-Matellán 2012, see also Labelle & Doron 2010). The se-
less variants are verbs of change or motion (telic or atelic), whereas se variants
are verbs of change of state (necessarily telic). This difference is observed in the
availability of durative and frame modifiers.
According to Cuervo (2003, 2014), the se-less variant of these verbs is mono-
eventive, consisting of only vGO, whereas the se variant is bi-eventive, having
both vGO and vBE.
property denoted by the state, whereas in unaccusative states, the subject is included within
the set of things having the property denoted by the state.
(in)transitive verbs: unergatives and unaccusatives 169
b. se variant
Se salieron tres clavos
‘Three nails came off’ (Cuervo (2014: 51–52))
The light verbs vGO and vBE can be paralleled to the Event and State heads
assumed in this chapter.10 The basic idea is that se-less variants are verbs of
change without entailing a final state, where the subject undergoes the pro-
cess ‘naturally’ (as in the case of snow, rain or leaves in the verb caer ‘fall’). In
contrast, se variants are composed of two subevents, a dynamic subevent and a
result. The subject is the holder the final result. According to Cuervo (2014), the
se clitic is the realization of the unvalued phi-features of vGO, which are valued
with the subject seated in the specifier of vBE.
Assuming Cuervo’s analysis that se variants are bi-eventive, having an Event
and a State subevent, we can understand why originally Romance se-verbs
are analyzed in Basque as unaccusatives (selecting be auxiliary and having an
absolutive subject). In Basque, they are structured in a configuration where the
subject is introduced in the specifier of State, like unaccusatives.
10 Go or become are derived flavors of Event in the assumed event configuration. For
instance, Event is interpreted as trans(ition) if it selects for State and if it is not selected
by Voice.
170 berro
It must be noted, though, that the relation between the unaccusative pattern in
Basque and the presence of the se clitic in Romance is not bi-directional, partic-
ularly in the domain of non-loan verbs. Many verbs that show an unaccusative
alignment in Basque do not have se clitic in their Romance counterparts, for
example, heldu ‘arrive’, etorri ‘come’, handitu ‘increase’, txikitu ‘decrease’, hazi
‘grow’, jaio ‘be born’, gertatu ‘happen’, faltatu ‘lack’ etc. Actually, the contrast
between mono-eventive and bi-eventive unaccusatives proposed in Cuervo
(2003, 2014) does not have a phonological reflex in Basque, since both show the
same alignment pattern (be auxiliary and absolutive subject). The conclusion
that can be drawn is that, for loan verbs, having a se clitic forces an unaccusative
alignment of predicates in Basque, but that, those verbs not having a se clitic
may be subject to other discriminating factors, such as transition, eventivity,
internal causation etc.
8 Conclusion
Acknowledgements
This work has been supported by the Basque Government (the research group
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 Program for research,
technological development and demonstration under grant agreement no.
613465.
References
Alberdi, Xabier. 2003. “Euskal aditz mailegatuen erregimena: hurbilketa”. In Jesus Mari
Makazaga & Bernard Oyharçabal (eds.), P. Lafitteren sortzearen mendemugako bil-
tzarra. Gramatika gaiak. Iker 14 (I). Bilbao: Euskaltzaindia. 37–60.
Aldai, Gontzal. 2006. “Is Basque morphologically ergative? The semantic split-intransi-
tive case marking system of Western Basque”. In Beatriz Fernandez & Itziar Laka
(eds.), Andolin Gogoan. Essays in honour of Professor Eguzkitza. Leioa: UPV/EHU.
117–138.
Aldai, Gontzal. 2008. “From ergative case marking to semantic case marking: The case
of historical Basque”. In Mark Donohue & Søren Wichmann (eds.), The Typology of
semantic alignment. Oxford: OUP. 197–218.
Aldai, Gontzal. 2009. “Euskararen kasu markaketaren aldakortasun dialektala”. In Beat-
riz Fernández, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: aldako-
rtasun sintaktikoa aztergai [Supplements of ASJU LII]. Donostia: UPV/EHU and
Gipuzkoako Foru Aldundia. 5–19.
Aranovich, Raúl. 2003. “The semantics of auxiliary selection in Old Spanish”. Studies in
Language 27-1: 1–37.
Basilico, David. 2010. The se Clitic and its Relationship to Paths. Probus 22: 271–302.
Batlle, M. 2002. L’expressió dels temps compostos en la veu mitjana i la passive pronomi-
nal. El process de subsitutció de l’auxiliar ‘ésser’ per ‘haver’. Barcelona: Institut d’Estu-
dis Catalans / Publicacions de l’Abadia de Montserrat.
Belleti, Adriana. 1988. “The case of unaccusatives”, Linguistic Inquiry 19: 1–34.
Belleti, Adriana. 1999. ““Inversion” as focalization and related questions”, Catalan Work-
ing Papers in Linguistics 7: 9–45.
Bentley, Delia & Thórhallur Eythórsson. 2003. “Auxiliary selection and the semantics of
unaccusativity”. Lingua 114: 447–471.
Berro, Ane. 2010. “Unergative Predicates in Basque Varieties: Consequences for the
Ergative Case Assignment”. DEA research work. University of the Basque Country
(UPV/EHU).
172 berro
Berro, Ane. 2012. “Three Levels of Root Insertion in Basque Intransitive Verbs”. In
Ernestina Carrilho & Beatriz Fernández (eds.), Journal of Portuguese Linguistics.
Special Issue, 11-1. Ediçoes Colibrí-Universidade de Lisboa. 7–22.
Berro, Ane. 2015a. Breaking verbs. From event structure to syntactic categories in Basque.
Ph.D. diss., University of the Basque Country (UPV/EHU) and Université Bordeaux
Montaigne (UBM).
Berro, Ane. 2015b. “Egoera predikatu berriak euskaraz”. In Irantzu Epelde (ed.), Euskal
hizkera eta dialekotoak gaur. Special Issue of Lapurdum 3: 11–25.
Berro, Ane & Ricardo Etxepare. 2017. “Ergativity in Basque”. In Jessic Coon, Lisa M.
Travis & Diane Massam (eds.), The Oxford Handbook of Ergativity. Oxford University
Press.
Bittner, Maria & Ken Hale. 1996. “The structural determination of case and agreement”.
Linguistic Inquiry 27: 1–68.
Bobaljik, Jonathan D. 1993. “On Ergativity and Ergative Unergatives”. In Collin Phillips
(ed.), MIT Working Papers in Linguistics 19: Papers on Case and Agreement II. Cam-
bridge: MIT. 45–88.
Burzio, Luigi. 1981. Intransitive verbs and Italian auxiliaries. Ph.D. diss., MIT. Cambridge,
MA.
Burzio, Luigi. 1986. Italian Syntax. Dordrecht: Reidel.
Castillo, María Elena. 2002. Inacusatividad y aspecto léxico en los verbos de movimiento.
Estudio diacrónico, Series Documenta Universitaria. Scripta I, Girona: Edicions a Peti-
ció SL:
Cuervo, María Cristina. 2003. Datives at Large. Ph.D. diss., MIT.
Cuervo, María Cristina. 2014. “Alternating unaccusatives and the distribution of roots”,
Lingua 141: 48–70.
Davidson, Donald. 1967. “The Logical Form of Action Sentences”. In Nicholas Rescher
(ed.), The Logic of Decision and Action. Pittsburgh: University of Pittsburgh Press.
[Reprinted in Davidson 2006.]
Davidson, Donald. 2006. The Essential Davidson. Oxford: Oxford University Press.
Dixon, Robert M.W. 1979. “Ergativity”, Language 55: 59–138.
Dixon, Robert M.W. 1994. Ergativity. Cambridge: Cambridge University Press.
Donohue, Mark & Søren Wichmann. 2008. The typology of semantic alignment. Oxford:
OUP.
Etxepare, Ricardo. 2003. “Valency and Argument Structure in the Verb”. In José Igna-
cio Hualde & Jon Ortiz de Urbina (eds.), A grammar of Basque. Berlin: Mouton de
Gruyter. 369–465.
Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. Ph.D. thesis, Univer-
sity of the Basque Country (UPV/EHU).
Folli, Rafaella. 2002. Constructing telicity in English and Italian. Oxford: Oxford Univer-
sity Press.
(in)transitive verbs: unergatives and unaccusatives 173
Folli, Rafaella & Heidi Harley. 2005. “Flavors of v”. In Paula Kempchinsky & Roumaya
Slabakova (eds.), Aspectual Inquiries. Dordrecht: Springer. 95–120.
Hale, Ken & S. Jay Keyser. 1993. “On Argument Structure and the Lexical Expression of
Syntactic Relations”. In Ken Hale and S. Jay Keyser (eds.), The View from Building 20.
Cambridge, Massachusetts: The MIT press. 53–109.
Hale, Ken & S. Jay Keyser. 2002. Prolegomenon to a Theory of Argument Structure. Cam-
bridge, Massachusetts: The MIT Press.
Hale, Ken & S. Jay Keyser. 2005. “Aspect and the Syntax of Argument Structure”. In Nomi
Erteschik-Shir & Tova Rapoport (eds.), The Syntax of Aspect: Deriving Thematic and
Aspectual Interpretation. New York: Oxford University Press. 11–42.
Harley, Heidi. 1995. Subjects, Events and Licensing. Ph.D. diss., MIT.
Harley, Heidi. 2002. “Possession and the Double Object Construction”. In Linguistic Vari-
ation Yearbook 2: 29–68. Amsterdam: John Benjamins.
Harley, Heidi. 2005. “How Do Verbs Get Their Names? Denominal Verbs, Manner
Incorporation and the Ontology of Roots in English”. In Nomi Erteschik-Shir &
Tova Rapoport (eds.), The Syntax of Aspect. New York: Oxford University Press. 42–
64.
Kayne, Richard S. 2008. “A note on auxiliary alternations and silent causation”. Ms., New
York University, New York.
Keller, Frank & Antonella Sorace. 2003. “Gradient auxiliary selection and impersonal
passiviation in German: an experimental investigation”. Journal of Linguistics 39:
57–108.
Kratzer, Angelika 1994. “The Event Argument and the Semantics of Voice”. Ms., Univer-
sity of Massachusetts at Amherst.
Kratzer, Angelika. 1996. “Severing the External Argument from its Verb”. In Johan Roo-
rych & Laurie Zaring (eds.), Phrase Structure and the Lexicon. Studies in natural
language and linguistic theory, vol. 33. Dordrecht: Kluwer Academic Publishers. 109–
137.
Labelle, Marie. 1992. “Change of state and valency”. Journal of Linguistics 28: 375–414.
Labelle Marie & Edit Doron. 2010. Anticausative derivations (and other valency alter-
nations) in French. Probus 22–2: 303–316.
Laka, Itziar. 1993. “Unergatives that Assign Ergative, Unaccusatives that Assign Accu-
sative”. In Jonathan D. Bobaljik & Collin Phillips (eds.), Papers on Case and Agree-
ment I, MITWPL 18. Cambridge, MA: MITWPL. 149–172.
Legendre, Géraldine & Paul Smolensky. 2009. “French inchoatives and the Unaccusativ-
ity Hypothesis”. In Donna B. Gerdts, John Moore & Maria Polinsky (eds.), Hypothesis
A/Hypothesis B: Linguistic Explorations in Honor of David M. Perlmutter. MIT Press.
Levin, Beth. 1983. On the Nature of Ergativity. Ph.D. diss., MIT.
Levin, Beth & Malka Rappaport Hovav. 1995. Unaccusativity. At the Syntax-Lexical Se-
mantic Interface. Cambridge, MA: the MIT Press.
174 berro
1 Introduction
1 See Butt (2010) for an overview, and for useful terminological precisions, notably the dis-
tinction between LVC and complex predicate. See also Bowern (2008). On the other hand, as
pointed out by Lin (2001:22), the use of light verb as related to LVC s is not exactly equivalent
to what is known as Chomskyan light verb (Chomsky 1995), that is, a functional verb that takes
the VP as complement and that is responsible for the projection of the external argument and
for the assignment of case to the object (see Larson 1988 and Hale & Keyser 1993 for seminal
proposals).
2 Although we use NVE throughout this work, since it constitutes a good label considering the
kind of LVC s that we encounter in Basque and Romance, it should be noted that in other lan-
guages like Urdu the element containing the conceptual portion of the LVC can be of verbal
category (see Butt 1995, 2010).
other, two constructions sharing the same NVE but being different with respect
to their LV. We can illustrate the former case with the minimal pair involving
the Spanish LVC s dar miedo “give fear” ‘frighten’ and dar vergüenza “give embar-
rassment” ‘embarrass’. These LVC s share the same LV, dar ‘give’, and the abstract
semantics of transferal inherent to this verb is preserved in these constructions.
Thus, both LVC s are construed with a dative understood as the experiencer of
the emotion:3
The difference between the two LVC s is exclusively of conceptual nature, and
corresponds to the difference between the emotions of fear and embarrass-
ment, as encoded in their NVE s. On the other hand, to exemplify the case of
two LVC s differing in their LV but not in their NVE we may keep dar miedo and
compare it with pasar miedo “pass fear” ‘be afraid’:4
In this case the same emotion (fear) is involved, but the argument-structure
properties are different. Thus, pasar miedo requires a subject experiencer and
3 The abbreviations used in glosses are the following ones: abs ‘absolutive’, adv ‘adverbial suf-
fix’, aux ‘auxiliary’, compl ‘complementizer’, dat ‘dative’, det ‘determiner’, erg ‘ergative’,
F ‘feminine’, ines ‘inessive’, inf ‘infinitive’, instr ‘instrumental’, loc ‘locative’, M ‘masculine’,
pl ‘plural’, prs ‘present’, pst ‘past’, ptve ‘partitive’, sg ‘singular’. Also, note that although we
generally provide a full-fledged gloss for Basque auxiliaries, in some instances we will just
gloss aux for reasons of space.
4 See, however, section 2.1, where we provide an account of several pairs of synonymous LVC s
sharing a NVE and differing in their LV s.
178 acedo-matellán and pineda
the stimulus can be expressed as a PP headed by con ‘with’, while dar miedo
appears with a subject stimulus and a dative experiencer. This difference, as
is evident, can only depend on the choice of LV. Interestingly, however, this
choice is not completely free: research has shown that the semantic properties
of the NVE may determine, at least to some extent, which LV it combines with
(Alba-Salas 2002: 51, and references cited therein). At the extreme of seman-
tic cohesion between the LV and the NVE lie, of course, the many LVC s with
idiomatic meaning, that is, a meaning that cannot be computed straightfor-
wardly from the sum of the meanings of the LV and the NVE, as that shown
by the Basque LVC s iskin egin “corner do” ‘elude’ or hanka egin “leg do” ‘escape
rapidly’.
A prominent fact about LVC s is that they tend to correlate with a verb that
expresses more or less the same meaning, as was shown above with the pair
take a walk and walk, and can be further illustrated with the pair dar vergüenza
“give embarrassment” and avergonzar ‘embarrass’ in Spanish. It is important to
emphasize that the semantic equivalence between these pairs is not complete,
as seminally pointed out by Wierzbicka (1982). In any case, LVC s do not always
count with synthetic counterparts. In fact, in languages like Basque most of
them do not (Martinez 2015: 30, Aldai 2009: 823). The generality of the ana-
lytic, LVC-based expression of a large set of events in Basque is, as we will see,
one of the major differences when we compare this language with modern
Romance.
5 In this short introduction we cannot do justice to the vast literature on LVC s, which encom-
passes a considerable number of languages—Asian ones featuring very prominently. Works
providing general remarks on LVC s, or comparing different languages include Butt & Geuder
(2001), Butt (2003, 2010), Bowern (2008), Butt & Lahiri (2013) and Ramchand (2014). For par-
ticular languages, see, among others, Jespersen (1954), Jackendoff (1974), Wierzbicka (1982),
Cattell (1984) and Kearns (1988), on English; Grimshaw & Mester (1988), on Japanese; Ahn
(1991), on Korean; Mohanan (1994), on Hindi; Butt (1995), on Urdu; and Lin (2001), on Chi-
nese. For references on LVC s in Basque and Romance (particularly on French and Spanish)
see the following sections.
light verb constructions in basque and romance 179
of predicational force and only contributes verbhood, and that the argument-
structure properties of the LVC are inherited from the NVE through some mech-
anism. For instance, in an influential paper on Japanese constructions based on
the LV suru ‘do’, Grimshaw & Mester (1988) propose that the LV does not have an
argument structure of its own, and that it’s the nominal NVE that transfers some
or all of its arguments to the LVC. In a similar vein, Cattell (1984) argues that
English LV s are predicationally vacuous. In turn, the lightness of LV s have made
some authors propose that they are the historic result of a process of semantic
bleaching of fully predicational verbs, and consequently, that they can be fur-
ther grammaticalized into auxiliary verbs and, finally, into inflectional markers
(Hook 1991, Hopper and Traugott 1993, 2003).
A different analysis is provided by Kearns (1988) within a lexicalist frame-
work distinguishing between Lexical Conceptual Structure and Syntactic Argu-
ment Structure (Hale & Keyser 1986, Rappaport & Levin 1988, among others).
An LV like give provides a Syntactic Argument Structure, with a particular num-
ber of argument slots in a particular configuration, and also an event argument,
but since the verb is delinked from its Lexical Conceptual Structure, no theta
roles are available. Complementarily, the NVE is associated with a Lexical Con-
ceptual Structure and a thematic array, but it lacks a full Syntactic Argument
Structure. In the LVC these two defective components converge to contribute
their lexical properties, yielding a full predicate.6 In any case, Kearns’s (1988)
work is best known for its contribution to differentiating between T(rue) LV s
like those in Take a walk, Give a groan or Have a bite, and what she calls Vague
Action Verbs, like those heading constructions such as Make an inspection, Give
a demonstration or Do the ironing. The basic difference is that True Light Verbs
combine with NVE s that are not argumental, since they do not pattern with
arguments: they cannot become passive subjects, they cannot be extracted,
they cannot be pronominalized, and they cannot be definite. By contrast, Vague
Action Verbs combine with full-fledged argumental DP s. However, as we will
see below, the range of variation within Basque LVC s (see section 2.2.) casts
some doubts on Kearns’ distinction between the two types of verbs found in
LVC s.
More recently, some of the abovementioned ideas have been to a large extent
disputed. First, the predicational lightness or emptiness of the LV has been
qualified, and standard analyses take the LV to be a predicate and to contribute
to a joint predication with the NVE through some mechanism. For example, in
Butt’s (1995) Lexical-Functional-Grammar approach, an operation called Argu-
ment Fusion combines the LV predicate (for instance, of “give” or “let” seman-
tics) and the NVE predicate to yield a single predicate. The new predicational
unit hosts, on the one hand, argument slots directly contributed by either of
the two basic predicates (e.g., an Agent from the LV, and a Theme from the NVE)
and, on the other, argument slots that are the result of the fusion of two argu-
ments slots of the basic predicates (for instance, a Causee that subsumes a Goal
of the LV and an Agent of the NVE). Even more recently, Ramchand (2014) has
taken a step further in this direction. In her analysis, LV s possess argument and
event structures, which are, moreover, identical to those of their non-LV coun-
terparts. Thus the LV give of LVC s like John gave Mary a cuddle is an item that
lexicalizes an event structure containing the three subeventive heads proposed
in Ramchand (2008)—Init(iation), Proc(ess) and Res(ult)—plus a Poss(essive)
head merged under Res and projecting a phrase whose specifier is the recipient
(Mary) and whose complement is the NVE (a cuddle). This structure provides
the skeletal meaning of caused possession and is actually the very same one
proposed for the non-light use of give (see Ramchand 2008: 103 for details).
The difference between light and non-light give boils down to the nature of
the complement: if the complement involves a bare event noun like cuddle,
the interpretation of the construction is that of an LVC. We witness a type of
approach, therefore, in which the emphasis is put on the similarities, rather
than the differences, between the light and the non-light versions of the same
verb.7
Ramchand’s (2014) work on LV s in English, Bengali, and Persian actually
aims at further exploring and validating an empirical claim made by Butt
(2003) and Butt & Lahiri (2013) on the basis of diachronic evidence from Indo-
Aryan languages. Ramchand (2014: 217) dubs this claim Butt’s Generalization:
“Unlike auxiliaries, which may become grammaticalized over time to have a
purely functional use, light verbs always have a diachronically stable corre-
sponding full or “heavy” version in all the languages in which they are found.”
From the strictly diachronic point of view, Butt & Lahiri (2013:7) state that
they “depart from the received view that the existence of a light verb is due
7 Cf. Ramchand’s (2014:218) Semantics of Structure Conjecture on the Limits of Lightness: “The
meaning of a light verb and its corresponding heavy alternant are in a subset-superset rela-
tion in their conceptual semantics, the light version being a proper subset of the heavy. Only
non-syntactic or conceptual information is systematically negotiable within the “same” lexi-
cal item. Anything that is present in the heavy version but not in the light must therefore be
a species of Type B meaning. At its most pared down, a light verb can only be as light as the
structural semantics corresponding to the type A meaning of the pair.”
light verb constructions in basque and romance 181
While Basque and Romance LVC s share a number of properties that are also
common to LVC s crosslinguistically, they also show some differences, partly
due to independent syntactic properties of the two linguistic systems, like case.
In this section, we offer a description of the morphosyntactic and semantic
properties of LVC s in Basque and compare them to Romance LVC s. We focus
first on the LV and then on the NVE and its cohesion with the LV.
182 acedo-matellán and pineda
(3) agur egin “do goodbye” ‘greet / say goodbye’, alde egin “do side / zone”
‘leave’, aharrausi egin “do yawn” ‘yawn’, aipu egin “do mention” ‘mention’,
aitor egin “do confession” ‘confess’, amets egin “do dream” ‘dream’, argi egin
“do light” ‘light up / enlighten’, arnas(a) egin “do breath” ‘breathe’, barre
egin “do laugh” ‘laugh’, behaztopa egin “do impediment” ‘bump into / com-
plicate’, bekatu egin “do sin” ‘sin’, berba egin “do word” ‘talk’, bultza egin “do
push” ‘push’, dantza egin “do dance” ‘dance’, dei egin “do call” ‘call’, domin-
istiku egin “do sneeze” ‘sneeze’, duda egin “do doubt” ‘doubt’, elurra egin
“do snow” ‘snow’, erregu egin “do plea / prayer” ‘beg / pray’, euria egin “do
rain” ‘rain’, eztul egin “do cough” ‘cough’, gogoeta egin “do reflection” ‘think,
reflect’, hitz egin “do word” ‘talk’, huts egin “do mistake / shortage” ‘fail /
miss’, ihes egin “do escape” ‘run away’, irain egin “do insult” ‘insult’, irrib-
arre egin “do smile” ‘smile’, iruzur egin “do trick” ‘mislead, deceive’, jauzi
egin “do jump” ‘jump’, kasu egin “do attention” ‘pay attention’, keinu egin
“do gesture / sign” ‘gesture, sign’, kontu egin “do care” ‘look after, take care’,
korrika egin “do running (adv.)” ‘run’, lan egin “do work” ‘work’, laztan egin
“do cuddle / caress / kiss” ‘caress, kiss’, lo egin “do sleep” ‘sleep’, min egin “do
pain” ‘hurt’, negar egin “do weep” ‘cry’, oihu egin “do shout, yell” ‘shout, yell’,
ospa egin “do “get out!”” ‘leave’, salto egin “do jump” ‘jump’, tiro egin “do
gunshot” ‘shoot’, topo egin “do stumble” ‘run into’, txalo egin “do applause”
‘applaud’, zin egin “do oath” ‘swear under oath’, among many others.9
However, Basque has also some LVC s with hartu ‘take’ (4), eman ‘give’ (5) and
eduki ‘have’ (Martinez 2015:113–114, Zabala 2004:486–488):10
(4) arnasa hartu “take breath” ‘breathe’, atseden hartu “take repose, rest” ‘rest’,
atsegin hartu “take pleasure” ‘feel happy, enjoy’, damu hartu “take pity,
regret” ‘feel sad, regret’, esku hartu “take hand” ‘participate, take part’, gogo
hartu “take mind, will” ‘decide’, hitz hartu “take word” ‘compromise’, indar
hartu “take strenght” ‘strengthen’, kontu hartu “take care” ‘take care of,
keep in mind’, lur hartu “take land” ‘land’, min hartu “take pain” ‘get hurt’,
parte hartu “take part”, pena hartu “take pity” ‘feel sad’, plazer hartu “take
pleasure” ‘enjoy’,11 poz hartu “take happiness” ‘feel happy’, su hartu “take
fire” ‘burn’, and few others.
(5) amore eman “give love” ‘yield, give in, collapse’, argi eman “give light” ‘light
up / enlighten’, arrazoi eman “give reason” ‘agree with’, aurpegi eman “give
face” ‘face’, begi eman “give eye” ‘look at’, bihotz eman “give heart” ‘encour-
age’, bizkar eman “give one’s back” ‘turn one’s back on’, buru eman “give
head” ‘finish, accomplish / face’, damu eman “give pity” ‘sadden’, errieta
eman “give scolding” ‘tell off’, esker eman “give thanks” ‘thank’, fede eman
“give faith” ‘trust, vouch for’, gogo eman “give mind, will” ‘look, pay atten-
tion / think, consider’, hats eman “give breath” ‘blow’, ikara/beldur eman
“give fear” ‘frighten’, kontseilu eman “give advise” ‘advise’, kontu eman “give
account” ‘give an explanation, give account’, laztan eman “give cuddle,
hig” ‘hug, caress’, leku eman “give place” ‘allow’, lur eman “give land” ‘bury’,
min eman “give pain” ‘hurt’, musu eman “give kiss” ‘kiss’, parte eman “give
report” ‘report, inform’, pena eman “give pity” ‘make someone feel sad’, poz
eman “give happiness” ‘make happy’, su eman “give fire” ‘set fire to’, and few
others.
(6) arrazoi eduki “have reason” ‘be right’, esku eduki “have hand” ‘help, sup-
port’, kontu eduki “have care” ‘be careful’, min eduki “have pain” ‘feel pain’,
on eduki “have good (adjective)” ‘maintain, claim’, pena eduki “have
pity” ‘feel sad’, and few others.
DP, clearly fall beyond the scope of this chapter, which is mainly devoted to LVC s formed
by a LV and a bare noun (hitz egin ‘talk’)—or by a full verb and a bare noun (hitz eman
‘promise’). Moreover, since such cases feature absolutive-marked DP s (adarra ‘(the) horn’,
hitza ‘(the) word’ …), it seems more appropriate to view them as the result of a metaphor-
ical (idiomatic) interpretation of expressions which are not bona fide LVC s.
11 This example and the two preceding ones correspond to psych predicates, which fre-
quently showed a similar pattern in Old Romance languages, as in Old Catalan prendre
delit, Old French prendre plaisir or Old Spanish tomar plazer “take pleasure”, ‘enjoy’. See
Acedo-Matellán & Mateu (to appear) for a detailed description of analytic psych predi-
cates in Old Catalan.
184 acedo-matellán and pineda
There are also some expressions involving eman ‘give’ and hartu ‘take’ that
show an extreme degree of coalescence between the LV and the NVE, since they
are actually not separable. They are arguably not LVC s anymore, but plain, syn-
thetic verbs:
(7) onartu “take good” ‘accept’, atzeman “give hoof” ‘catch, understand’, haute-
man “give choice” ‘realise, notice’ / choose, hitzeman “give word” ‘promise’
For French and Spanish, Alba-Salas (2002: 636–637) provides the following
respective lists, in addition to faire (French) and hacer (Spanish) ‘do’:
(8) asséner ‘deal’ (asséner un coup ‘hit’), avoir ‘have’ (avoir une maladie ‘be
sick’), coller ‘give’ (coller une claque ‘smack’), commetre ‘commit’ (commet-
tre une aggression ‘commit an assault’), donner ‘give’ (donner un conseil
‘give some advice’), dire ‘say’ (dire des compliments ‘pay some compli-
ments’), diriger ‘direct’ (diriger une offensive ‘attack’), effectuer ‘carry out’
(effectuer un paiement ‘make a payment’), entamer ‘initiate’ (entamer une
discussion ‘initiate a discussion’), ficher ‘give’ ( ficher une giffle ‘slap’), flan-
quer ‘give’ ( flanquer un coup de pied ‘kick’), foutre ‘give’ ( foutre une claque
‘smack’), jeter ‘throw’ ( jeter un regard ‘cast a glance’), lancer ‘launch’
(lancer une attaque ‘launch an attack’), livrer ‘carry out’ (livrer une bataille
‘engage a battle’), passer ‘make’ (passer un coup de téléphone ‘make a
phone call’), porter ‘carry’ (porter un coup ‘strike’), poser ‘put’ (poser une
question ‘ask a question’), prendre ‘take’ (prendre une décision ‘make a
decision’), prêter ‘lend’ (prêter attention ‘pay attention’), procéder (à) ‘pro-
ceed (to)’ (procéder a une lecture ‘proceed to a reading’), subir ‘undergo’
(subir une métamorphose ‘undergo a metamorphosis’)
(9) asestar ‘give’ (asestar un golpe ‘hit’), coger ‘catch’ (coger envidia ‘get envi-
ous’), cometer ‘commit’ (cometer un asesinato ‘commit murder’), contraer
‘contract’ (contraer matrimonio ‘marry’), correr ‘run’ (correr peligro ‘be
in danger’), dar ‘give’ (dar un paseo ‘take a walk’), decir ‘say’ (decir un
cumplido ‘pay a compliment’), dirigir ‘direct’ (dirigir una crítica ‘criticize’),
echar ‘throw’ (echar una mirada ‘give a look’), efectuar ‘carry out’ (efectuar
un pago ‘make a payment’), ejercer ‘make’ (ejercer presión ‘put pressure’),
experimentar ‘experience’ (experimentar una mejora ‘improve’), llevar a
cabo ‘carry out’ (llevar a cabo una privatización ‘privatize’), meter ‘put’
(meter una bofetada ‘slap’), pegar ‘give’ (pegar una patada ‘kick’), perder
‘lose’ (perderle el respeto (a alguien) ‘lose respect (for someone)’), pillar
‘catch’ (pillar celos ‘get jealous’), presentar ‘present’ (presentar la dimi-
light verb constructions in basque and romance 185
In section 1.1, we mentioned that the choice of a given LV is not completely free,
but that the semantic properties of the NVE may determine the LV chosen. This
can be seen when comparing, for example, English and Basque LVC s (see (10)):
as Martinez (2015: 78) points out, most English LVC are also found in Basque.
However, this is only a tendency, and English and Basque semantically equiva-
lent LVC s may also bear different LV s—cf. English have and give vs Basque egin
‘do’ in (11)—without any relevant difference regarding event- and argument-
structure properties:
186 acedo-matellán and pineda
This random comparison between English and Basque, and particularly the
mismatches in (11), actually lead us to discuss another important issue that
was introduced in section 1.1: the division of syntactic and semantic labour in
LVC s. Thus, whereas the semantic contribution of the LV is limited to functional
notions related to argument and event structure, the NVE encapsulates the con-
ceptual content of the LVC (the particular event evoked). Interestingly, Basque
offers some examples where the same NVE can combine with different LV s, the
resulting LVC s showing, however, no difference in meaning, at least at the con-
ceptual level:
Moreover, in the Romance area, similar examples are found when compar-
ing Spanish and French:
12 See, for example, the almost systematic equivalence between Spanish dar ‘give’ and Cata-
lan fer ‘do’: dar un beso ~ fer un petó ‘kiss’, dar un abrazo ~ fer una abraçada ‘hug’, dar miedo
~ fer por ‘frighten’, dar pena ~ fer pena ‘provoke sorrow’, dar rabia ~ fer ràbia ‘anger’, dar
lástima ~ fer llástima ‘provoke pity’, dar envidia ~ fer enveja ‘provoke jealousy/envy’, and so
on. Moreover, both verbs turn out to be even more similar, as far as argument structure is
188 acedo-matellán and pineda
Actually, obtaining two LVC s with different meanings and different event-
and argument-structure properties is what normally occurs in Basque when
different LV s are combined with one same NVE (see also Martinez 2015: 114;
Zabala 2004: 457–458, 483). And this is what one expects given the above-
mentioned division of labour between the two components of the LVC. Thus, to
the examples in (14) for the pair egin ‘do’/eman ‘give’, we can add the following
ones with other LV s, including hartu ‘take’ and eduki ‘have’:
(15) a. arrazoi ‘reason’ eman ‘agree with’ ~ arrazoi eduki ‘be right’
b. atsegin eman ‘please’ ~ atsegin hartu ‘enjoy, take pleasure in’
c. damu hartu ‘regret’ ~ damu eman ‘cause regret’
d. esku ‘hand’ hartu ‘participate, take part’ ~ esku eman ‘help’
e. hats hartu ‘breathe’ ~ hats eman ‘blow’
f. kontu hartu ‘take care of, keep in mind’ ~ kontu eduki ‘be careful’
g. min hartu ‘get hurt’ ~ min eman ‘cause pain’
h. pena hartu ‘get sad’ ~ pena eman ‘sadden’
The examples in (15), where verbs other than egin ‘do’ are at play, clearly sup-
port Ramchand’s (2014) approach to LVC s, in the sense that the argument and
event structure of a given LV is identical to those of their non-LV counterpart.
For example, min hartu ‘get hurt’ behaves like liburua hartu ‘get, receive the
book’; and min eman ‘give, cause pain’ behaves like liburua eman ‘give the book’.
In other words, light eman ‘give’ and hartu ‘take’ display a meaning very close
to that of their non-light counterparts, and preserve the number of arguments
and their categorial requirements, generally selecting for NP s.13 By contrast, as
will be shown in section 2.2, egin ‘do’ allows a wide range of NVE s, including
instrumental and locative case-marked DP s, adverbials, and even words not
used outside the particular LVC (Zabala 2004: 481–482). Thus, it seems clear
that there is a difference between instances with egin ‘do’, clearly a light verb,
and instances with eman ‘give’ or hartu ‘take’, which, even when appearing in
this type of expressions, have been considered verbs with full semantic content
(Zabala 2014: 458, 461). Actually, Zabala’s difference between DO-verbs (light)
and GIVE/TAKE-verbs (fully contentful) could account for the many intra- and
concerned, once we adopt the view that the dative, commonly appearing with GIVE, is an
“added argument” in the sense of Pylkkänen (2002, 2008)—see McFadden (2006) for rel-
evant argumentation in the context of Germanic. Thus, under the applicative hypothesis,
it is not the case that GIVE has a slot for the dative object which DO lacks, but instead both
verbs can combine with a dative-introducing head, the applicative.
13 And occasionally DP s: bide(a) eman “give (the) way”, hats(a) eman “give (the) breath”,
gogo(a) hartu “take (the) mind”.
light verb constructions in basque and romance 189
14 We use the term thing following Rodríguez & García Murga’s (2003: 433) terminology:
“Baina barre gauza bati egiten dio erreferentzia, ez egoera bati.” ‘But barre refers to a thing,
not to a situation.’ It seems to us that this concept must be understood in a broad sense,
beyond material objects.
15 The “semantic coercion” of the NVE as denoting an event is reminiscent of Marantz’s
(2005) account of creation predicates like Bake a cake, in which the DP a cake is also taken
to denote an event. An updated view of this type of syntactically driven semantic coercion
is that expounded by Wood & Marantz (2017) in terms of contextual allosemy.
190 acedo-matellán and pineda
Still, it is worth noting that examples such as the next one, while certainly
marked, are possible:
What is worth noticing about examples like (19) is that, while the NVE cer-
tainly shapes the conceptual scene evoked (i.e., the appearance of rain is usually
light verb constructions in basque and romance 191
Moreover, the other contrasts (20)–(22) upon which Zabala bases her defence
of such a “fusion” actually do not differ much from what is found in any transi-
tive construction, where the verb “expresses a range of predicates depending on
the choice of direct object” (Marantz 1984: 25). Thus, it is expected that the LV
egin ‘do’ requires different arguments depending on the NVE that it combines
with: in (20) egin combined with hitz ‘word’ admits a dative argument, in (21)
egin combined with lo ‘sleep’ does not admit a dative argument, and finally in
(22) egin combined with laztan ‘hug, kiss’ generally requires a dative argument.
In a smilar vein, Rafel (2004: 404) also discusses how the valency of tener ‘have’
changes according to its use in a LVC, as has been shown above, or in a run-of-
the-mill transitive construction:
Rafel points out that the licensing of the dative depends on the use of the NVE
miedo ‘fear’, and that the dative is not licensed in a normal transitive predi-
cate such as the one in (24). An extreme case in the position that the argument
structure of the LVC basically depends on that of the NVE is that represented
by Gross (1981) or Danlos (1992: 2), the latter considering that the function of
the LV (“the support verb”) is “mainly to bear the tense and aspect values of the
sentence”.
There are several observations to be made in relation to the above facts,
which we hope to show to conform with, rather than jeopardize, Ramchand’s
(2014) conjecture on the division of labour between the LV (= argument and
event-structure) and the NVE (= conceptual content), as adhered to in this
paper. The first thing to notice is that, for instance, tener miedo may also surface
with a genitive PP encoding the stimulus, undermining the more deterministic
stance taken by Alonso-Ramos (1998):
What the alternation between this example and (23)a shows, we submit, is
that, while the use of the NVE miedo ‘fear’, as opposed to esperanza ‘hope’, does
indeed determine the presence of a stimulus in the conceptual scene evoked,
that “argument” may find different ways of realization in the actual sentence,
either as a genitive or as a dative. Perhaps more importantly, it can be dropped
altogether, the grammaticality of the sentence remaining unaltered:16
Quite crucially, even in (26) the presence of the stimulus does not fail to be
inferred: when we are afraid we are certainly always afraid of something. Gram-
matically, however, there is no need for the stimulus to be expressed in the LVC,
and this, we note, must have to do exclusively with the syntactic properties of
the construction, as headed by the LV.17 It is interesting, in this respect, to con-
trast the LVC at hand with the corresponding synthetic verb temer ‘fear’:
While tener miedo “have fear” ‘be afraid’ and temer ‘fear’ mean more or less
the same, only the former can drop the constituent encoding the stimulus,
suggesting that, indeed, argument structure qua argument realization is a pre-
rogative of the head of the construction, i.e., the verb, either heavy or light,
rather than of the NVE.
With respect to (22), we observe that the dative is not completely out in pred-
icates headed by non-light tener ‘have’:
This shows, again, that the structural licensing of the dative does not really
depend on the complement of tener ‘have’, but on some independent configu-
rational condition.18
It should have become clear, we hope, that we are making a crucial distinc-
tion between the arguments licensed structurally, by the configuration headed
by the LV, and the participants of the event, encoded as a conceptual property
of the NVE, and which are not obligatorily realized.19
b. Gain-ez egin
top-instr do
‘Overfly’
18 See Pineda (2016) for an implementation of the theory of applicatives (Pylkkänen 2002,
2008; Cuervo 2003) to the analysis of datives in Romance and Basque.
19 We follow in this respect the different between structure and content arguments in Grim-
shaw’s (2005 [1993]) terms and structure and content participants in Rappaport Hovav
& Levin’s (1998) terms. And see also, within Construction Grammar, Goldberg (1995).
Notwithstanding our criticism of Alonso-Ramos’s (1998) take on tener miedo/esperanza
“have fear/hope”, we acknowledge that the difference that we are pointing out roughly
corresponds to her distinction between actant syntaxique and actant sémantique.
light verb constructions in basque and romance 195
c. Atze-ra egin
back-adl do
‘Go back’
d. Borroka-n egin
fight-ines do
‘Fight’
e. Aitzurr-ean egin
hoe-det-ines do
‘Dig’
f. Ardatz-ean egin
spindle-ines do
‘Spin’
g. Kart-etan egin
cards-ines do
‘Play cards’
h. Futbol-ean egin
football-ines do
‘Play football’
b. Adar-ka egin
horn-adv do
‘Gore’
c. Aitzur-ka egin
hoe-adv do
‘Dig’
Some LVC s alternate between a version with a bare nominal NVE and a ver-
sion with a locative or adverbial NVE, as the ones shown in (29) and (30)
respectively—note that -ka is an adverbializing suffix. Unlike the versions
involving a bare nominal NVE, those involving a non-nominal NVE show an iter-
196 acedo-matellán and pineda
b. Arraun(-ean) egin
oar/rowing(-ines) do
‘Row’
c. Borroka(-n) egin
fight(-ines) do
‘Fight’
d. Dantza(-n) egin
dance(-ines) do
‘Dance’
e. Ehiza(-n) egin
hunting(-ines) do
‘Hunt’
f. Errieta(-n) egin
dispute(-ines) do
‘Dispute’
b. Borbor(-ka) egin
Bubbling(-adv) do
‘Bubble’
c. Eztul(-ka) egin
cough(-adv) do
‘Cough’
light verb constructions in basque and romance 197
d. Jauzi(ka)/Salto(ka) egin
jump.adv do
‘Jump’
e. Korri(-ka) egin
running(-adv) do
‘Run’
f. Laster(-ka) egin
run(-adv) do
‘Run’
g. Musu(-ka) egin
kiss(-adv) do
‘Kiss’
h. Oihu(-ka) egin
scream-adv do
‘Scream’
i. Zotin(-ka) egin
hiccup(-adv) do
‘Hiccup’
Additionally, the NVE can also be an adjective, as in the LVC on hartu ‘accept’
(lit. “good take”) and a particle, as in goiti egin ‘vomit’ (lit. “do up”), although
these two options are “relatively rare and present a high degree of lexicali-
sation” (Oyharçabal 2006: 787).20 It is also worth mentioning that the NVE of
meteorological LVC s bero egin “hot do” ‘be hot’ and hotz egin “cold do” ‘be cold’
can function as a noun or as an adjective (Martinez 2015: 54–56).
Focusing now on nominal NVE s, they are inanimate nouns, and they are
non-referential (Martinez 2015: 44; Zabala 2004: 470).21 Other than that, even
though the range of nominals that can appear in Basque LVC is a closed set,
there is no semantic property that can apply to all of them: nominals can be
abstract (gogoeta egin “do meditation” ‘meditate’) or concrete (buru egin “do
20 Note that on hartu, actually written onartu ‘accept, admit’, is fully lexicalized and behaves
as a transitive verb: legea onartu ‘accept the law’.
21 A few NVE s, such as turrut (egin) ‘mock / fail’ or laprast (egin) ‘slip’, are not found in syn-
tactic environments other than LVC s (Zabala 2004:449).
198 acedo-matellán and pineda
head” ‘face’), countable (hitz egin “do word” ‘talk’) or uncountable (min egin
‘do harm’ ‘hurt’), and eventive (arrantza egin ‘do fishing’) or resultative (laprast
egin “do slip” ‘slip’) or with no argument or event structure at all (elurra egin
“do snow” ‘snow’) (Rodríguez & García Murga 2003: 420). Thus Basque con-
trasts with Japanese, where, according to Grimshaw & Mester (1988), NVE s are
always event nouns that transfer their argument structure to the LVC, as men-
tioned in section 1.2 (see Fernández 1997 for a comparison between Japanese
and Basque LVC s). However, Grimshaw & Mester’s proposal can be maintained
for Basque under a slightly different view: as argued above, regardless of the
type of noun and the readings it may have in other syntactic environments, the
NVE, in the context of the LVC, always expresses an event. Recall that consid-
ering all Basque NVE s as denoting events goes in line with Rodríguez & García
Murga’s (2003: 432–434) semantic analysis, according to which the NVE may
refer to a thing, but the final interpretation is achieved at the conceptual level
by means of a semantic process (Jackendoff 1990) ensuring that the NVE no
longer refers to a thing, but to an event. In addition, given that events inherently
lack referenciality, the above-mentioned non-referenciality of Basque NVE s is
accounted for (Zabala 2004: 470).
The lack of referentiality is also a property of a small group of Basque NVE s
that are not bare nominals but bear a determiner (Martinez 2015: 33–34, 184–
185). These LVC s mostly refer to meteorological phenomena and their NVE
bears the definite article -a (the only determiner possible in these structures):22
22 Thus, other determiners such as bat ‘a, one’, the demonstratives hau/hori ‘this, that’ (which
induce a necessarily specific interpretation) or the plural definite article -ak are not
allowed in NVE s. Examples can be found where instead of -a the quantifier asko ‘very
much, a lot’ is used, as in (i), but according to Martinez (2015:58, fn. 64) asko refers in this
case to the whole meteorological situation, rather than quantifying the noun euri:
(i) Euri asko egin du.
rain a_lot do aux.prs.erg3sg.abs3sg
‘It has rained a lot.’
This contrasts with its use with common direct objects, such as in (ii), where asko gen-
erally triggers the plural absolutive agreement -it- in the verb (actually, there is dialectal
variation in this point, with some dialects accepting (ii) with the singular auxiliary du):
(ii) Liburu asko irakurri ditu.
book a_lot read aux.prs.erg3sg.abs3pl
‘(S)he has read a lot of books.’
However, note that agreement does not constitute a reliable test here: even if asko quanti-
fied euri, we would not see the morphological effect on the verb, since euri is uncountable
and thus never triggers plural agreement on the verb. Actually, in our view, the null hypoth-
esis involves assuming that asko is quantifying directly over euri in Euri asko egin du. This
scenario is reminiscent of what Bosque & Masullo (1998) argue is involved in Spanish syn-
thetic predicates like sangrar mucho ‘to bleed a lot’. These authors propose that the degree
light verb constructions in basque and romance 199
eguzkia egin “do sun”, elurra egin “do snow”, euria egin “do rain”, haizea egin
“do wind”, izotza egin “do ice”, trumoia egin “do thunder”, and so on (Martinez
2015: 56–60).23 In a few other LVC s, again mostly meteorological, the deter-
miner is optional: hotz(a) egin “do cold” ‘be cold’ and bero(a) egin “do hot” ‘be
hot’, eztul(a) egin “do cough” ‘cough’ and arnas(a) egin “do breath” ‘breath’. In
any case, regardless of whether the NVE is determined or not, it is not refer-
ential, but has a generic interpretation (Zabala 2004: 470; Martinez 2015: 56–
60).
Since the NVE in meteorological LVC s is usually determined but still non-
referential, non-referentiality cannot be directly related to the absence of the
determiner. Another reason that prevents us from establishing this connec-
tion is that in common transitive structures a-marked DP s can be found that
are not referential. If we compare the LVC lan egin “do work” and the com-
mon transitive structure lana egin (lana bearing a determiner -a), the latter
has to be interpreted with a definite reference (“the job, that job”), but if we
look at other common transitive structures such as etxea egin ‘build a/the
house’ or ogia jan ‘eat (the) bread’, whose direct objects cannot appear in an
LVC, then the determined direct objects can be either definite (‘the house,
that house’, ‘the bread, that bread’) or indefinite (‘a house’, ‘bread’) (Martinez
2015: 63).24 In short, Basque arguments always need to be determined, even if
they are generic. Thus, as Himmelmann (2001: 832) points out: “Definiteness,
though undeniably of central importance to the grammar and typology of the
articles, is only one of a number of meanings that can be conveyed by arti-
cles”.
Under a wider conception of LVC s, one can find in Basque some construc-
tions where the LV egin combines with a determined DP, such as bizarra egin
“do the beard” ‘shave’, harrikoa egin “do the washing up”, erroak egin “do the
roots” ‘settle down’, bakeak egin “do the peaces” ‘make peace’. We think that
these few cases are the ones that can be considered constructions involving
Vague Action Verbs in Basque, where the NVE seems to pattern with full-fledged
argumental DP s—see Kearns’s (1988) distinction in section 1.2. In turn, these
modifier mucho ‘a lot’ quantifies over a nominal predicate SANGRE ‘blood’, embedded in
the lexico-syntactic representation of the verbal phrase, involving the abstract LVC-like
structure DO BLOOD (Hale & Keyser 1993, 2002). The difference between Spanish and
Basque would lie in the synthetic vs analytic character of the predicate. On the quantifi-
cation of the NVE in Basque LVC s, see examples (58) below and Oyharçabal (2006: 792).
23 It is not surprising that meteorological LVC s are considered a particular class of Basque
LVC s by several authors (Zabala 2004; De Rijk 2008; Martinez 2015).
24 An anonymous reviewer points out that the contrast may need to be nuanced, as some-
thing like lan-a egin “work-a do” can be non-referential too in central dialects of Basque.
200 acedo-matellán and pineda
(33) a. hacer fiesta “do holiday” ~ festejar ‘celebrate’, hacer mención “do men-
tion” ~ mencionar ‘mention’, hacer juramento “do oath” ~ jurar ‘swear’,
hacer elogio “do praise” ~ elogiar ‘praise’, hacer chantaje “do backmail” ~
chantajear ‘blackmail’, hacer daño “do damage” ‘hurt’ ~ dañar ‘damage’,
hacer ayuno “do fast” ~ ayunar ‘fast’
b. hacer deporte ‘do sport’, hacer mérito “do merit” ‘mention’, hacer juerga
“do fun” ‘have fun’, hacer luto “do mourning” ‘mourn’, hacer turismo ‘do
tourism’, hacer dinero ‘make money’
(34) a. faire mention “do mention” ~ mentionner ‘mention’, faire (de) la cuisine
“do (of) the cooking” ~ cuisiner ‘cook’, faire la recolte “do the harvest”
~ recolter ‘harvest’, faire du progrès “do of the progress” ~ progrésser
‘progress, advance’, faire le nettoyage “do the cleaning” ~ nettoyer ‘clean’,
faire des études “do of the studies” ~ étudier ‘study’, faire de l’ ombre “do
of the shadow” ~ ombrager ‘cast a shadow on’
b. faire du sport “do of the sport” ‘do sport’, faire du tourisme “do of the
tourism” ‘do tourism’, faire la ronde “do the round” ‘go the rounds’, faire
la revue “do the review” ‘review’, faire l’amour “do the love” ‘make love’,
faire la guerre “do the war” ‘wage war’, faire fortune “do the fortune”
‘make one’s fortune’.
As pointed out by Acedo-Matellán (2014) for the avoir cases, these NVE s are
truly bare, since they are caseless (cf. failure of en-substitution) and cannot sus-
tain nominal quantification (cf. availability of très ‘very’, but not of beaucoup de
‘a lot of’):
In the generative tradition, Basque LVC s (48) have often been identified with
syntactically unergative verbs, which show a synthetic correlate in other lan-
guages,25 such as Romance (49)–(50):
(48) hitz egin ‘talk’, barre egin ‘laugh’, negar egin ‘cry’, lo egin ‘sleep’, amets egin
‘dream’, salto egin ‘jump’, dantza egin ‘dance’, lan egin ‘work’
(49) Sp. hablar ‘talk’, reír ‘laugh’, llorar ‘cry’, dormir ‘sleep’, soñar ‘dream’, saltar
‘jump’, bailar ‘dance’, trabajar ‘work’
(50) Fr. parler ‘talk’, rire ‘laugh’, pleurer ‘cry’, dormir ‘sleep’, rêver ‘dream’, sauter
‘jump’, danser ‘dance’, travailler ‘work’
Basing on facts like the analytic realization of unergatives in Basque (and other
languages), Hale & Keyser (1993) postulated a universal transitive nature for
unergative verbs: that is, unergative (i.e. activity) predicates take an internal
argument that is not the subject. Thus, the only difference between the Basque
analytic patterns in (48) and the Romance synthetic patterns in (49)–(50) is
whether overt incorporation of the NVE into the LV takes place or not. Actu-
ally, even in Basque many complex unergatives have a simplex counterpart, as
shown by dantza egin “do dance” and dantzatu, both meaning ‘dance’, and here
too the difference between the two options has to do with whether nominal
incorporation has taken place or not. If it does not take place (dantza egin ‘do
dance’, ‘dance’), the derivation will parallel that of a transitive clause (zerbait
egin ‘do something’).
However, there is another dimension where syntactic incorporation becomes
a controversial issue in the theoretical treatment of Basque and, to some extent,
25 See Martinez (2015: 21–32) for a very detailed semantic classification of Basque LVC s
that takes into account several previous semantic classifications, especially De Rijk’s
(2008), but also Rodríguez & García Murga’s (2003: 423), Etxepare’s (2003: 400–402) and
Zabala’s (2004: 472–476) among others. As a matter of fact, Etxepare (2003) provides a
rich sample of LVC s, distinguishing verbs of emission (sound emission oihu egin ‘yell’,
light emission dir-dir egin ‘shine’, verbal emission errieta egin ‘argue / tell off’), inter-
nal body motion (dar-dar egin ‘tremble’), physical activities (actions against an object
or an individual tiro egin ‘shoot’, motion verbs igeri egin ‘swim’, bodily functions izerdi
egin ‘sweat’), mental activities (duda egin ‘doubt’) and behavioral verbs (axut egin ‘chal-
lenge’).
204 acedo-matellán and pineda
26 The term incorporation has been used to refer to certain phenomena found in a variety
of languages in which verbs overtly host material arguably belonging to their comple-
ments, like prepositions (as in Chichewa, Baker 1988) or nouns (as in Siberian Koryak,
Mithun 1984). Within generative approaches, and since the extremely influential work of
Baker (1988), incorporation also refers to the syntactic mechanism proposed to account
for these phenomena, to wit, the movement of the head of the verb complement to the
immediately c-commanding verbal head (see Haugen 2015 for a recent overview and a
new proposal on the nature of incorporation). Since incorporation, in the standard gen-
erative approaches, involves head movement, it is worth pointing out that this operation
has come under considerable scrutiny at least since Chomsky (1995), who first pointed
out the theoretical inconsistencies associated to this operation and proposed to relegate
it to the purely morphophonological component of grammar. The debate is far from being
settled, with proposals that actual movement (Matushansky 2006) or lack thereof (Brody
2000; Bye & Svenonius 2012) is involved in the phenomena traditionally dealt with under
this rubric. For a recent overview, see Barrie & Mathieu (2014), and for interesting theo-
retical remarks, Haugen (2015).
27 Due to space restrictions, we must refrain from dealing with some other (dis-)similari-
ties between LVC s and common transitive structures, such as the position of NVE s with
respect to manner adverbials such as ondo, ongi ‘well’, or gaizki ‘badly’ (Oyharçabal 2006:
794; Martinez 2015: 51, fn. 53), the (non-)acceptability of adjectives and other type of mod-
ifiers (García Murga 2003: 421; Zabala 2004: 452, Oyharçabal 2006: 796; Martinez 2015:
47–48) or the (im)possibility of ellipsis in different contexts (Oyharçabal 2006: 793; Mar-
tinez 2015: 48–49), among others.
light verb constructions in basque and romance 205
One first argument for the transitive analysis is the incompatibility of LVC s
with a DO: if the NVE (dantza) bears absolutive case, a “second” direct object
(tango) is expected not to be case-licensed (51)a, in contrast with the synthetic
version of the predicate (51)b:28,29
This is also the argument that Acedo-Matellán (2014) wields to defend that the
case-less NVE in the French LVC s dealt with at the end of the previous section
(e.g., avoir faim “have hunger” ‘be hungry’) occupies the object position, thus
constituing an “unincorporated root”.
That said, the incorporation analysis, whereby the NVE is caseless and under-
goes incorporation, can also account for the contrast in (51). Indeed, invoking
Chomsky’s (1981) Empty Category Principle, a “second DO” would fail to be
licensed, since the trace left behind by the incorporated DO cannot be occu-
pied by another argument. Following Martinez (2015: 177), the unfeasibility of
licensing “second DO s” actually shows that, even for those cases in which there
is actually phonological evidence of incorporation (such as hitz egin “do word”
28 Berro (2015) argues that the NVE occupies the same position as DO s in simplex verbs,
basing on the aspectual properties of the predicate. Thus, just like simplex verbs can be
aspectually bounded if the DO denotes a bounded scale, the event denoted in complex
verbs is also measured by the scalar (aspectual) properties of its NVE (i.e., the root), thus
indicating that they occupy the same position—see Harley (2005) for a seminal work on
the parallelism between internal argument DP s and roots as regards their contribution to
Aktionsart.
29 A “second DO” is possible with LVC s expressing mental or communicative activities, when
it is a subordinate clause (Martinez 2015: 74, 376–377; Zabala 2004:480):
(i) Gau-ez, [bere armairuan borrokan zihardutela] amets egi-ten
dream do-inf
zuen.
aux.pst.erg3sg.abs3sg.
‘At night, (s)he used to dream [that they were fighting inside his/her closed].’
However, these apparent double object constructions (which are also found with ditran-
sitive verbs, such as abisatu ‘tell [do-abs me] [do subordinate clause]’) can easily be
accounted for under the assumption that completive subordinate clauses don’t need case,
or are self-sufficient as far as case is concerned (Pesetsky & Torrego 2001).
206 acedo-matellán and pineda
‘talk’, see below), it is a syntactic process, not a lexical one, unlike that argued
to be involved in English laugh, formed in the lexical syntax out of do+laugh
(Hale & Keyser 1993, 2002; Haugen 2015). In this latter kind of incorporation no
trace is left behind allowing for a DO (a cognate object) to be projected—cf.,
e.g., laugh a good laugh.30
Both the proponents of the transitive analysis and those of the incorporation
analysis have put forward many other arguments, based on a variety of syntac-
tic tests (see below). However, since not all Basque LVC s show a homogenous
behavior, most syntactic tests have actually been used to justify both views. In
particular, the similarities and contrasts between NVE s and regular DO s arise
when the focus is put on the degree of cohesion between the LV and the NVE.
As we shall see, the apparently contradictory evidence in fact points out that
while some Basque LVC s may be the result of an incorporation process, others
cannot be. To makes matters worse, the non-homogenous syntactic behaviour
of Basque LVC s does not only depend on the particular LVC, since differences
across dialects and registers are also at play (see Oyharçabal 2006), thus mak-
30 Two LVC s in Basque are exceptional in that they allow a DO, namely huts egin “do fail-
ure/empty” ‘miss’ and hots egin “do noise/rumor” ‘call, propagate’ (Oyharçabal 2006: 791,
fn. 6; Martinez 2015: 172, 377–378):
(i) Ume-a-k eskol-ak huts egin ditu.
kid-det.abs-erg class-abs.pl failure/empty do aux.prs.erg3sg.abs3pl
‘The kid missed the classes’
(ii) Gizon-a-k berrion-a hots egin
man-det.abs-erg good_news-det.abs noise/rumor do
du
aux.prs.erg3sg.abs3sg
‘The man spread the good news’
In these cases a lexicalization process has taken place, so that the argument structure of
the LVC includes a slot for a DO with the Theme-role. That a reanalysis has taken place
in these cases is also evinced by phonology, huts egin and hots egin being generally pro-
nounced as single words with a single accent: hutségin, hotségin. In other words, there is no
longer any trace preventing the merger of a DO. Rather, there has been a reanalysis yielding
a transitive verb able to take a full-fledged DP as complement that can check the structural
absolutive case (Martinez 2015: 172, 377–378). The cases in (i) and (ii) could thus be seen
as instances of Mithun’s (1984:857) case-manipulating type of incorporation, whereby the
case left by the incorporated argument is available for another argument (until then an
oblique), which is thus promoted.
We think that huts egin—but not hots egin—is amenable to a different analysis,
namely, that huts, which can also be an adjective meaning ‘empty’, is a secondary pred-
icate rather than an argument. Specifically, huts would be a predicate of the DO, which
would then be its external argument: trena huts egin “make the train empty, missed”, i.e.,
‘miss the train’. If this analysis is on the right track, huts egin would pattern with the above-
mentioned onartu, whose NVE is also an adjective, on ‘good’.
light verb constructions in basque and romance 207
ing it difficult to find systematic patterns that allow to tip the scales towards an
analysis or the other. It seems therefore that the most adequate way to proceed
is by assuming the existence of syntactically different types of LVC s in Basque.
One first type of Basque LVC involves constructions that behave as com-
mon transitive structures, formed by a transitive verb and its DO. Evidence of
their run-of-the-mill transitive behaviour includes the separability of the LV
and the NVE in interrogative sentences and focus structures (see (53)–(56)),
and availability of partitive case marking on the NVE in negative sentences (57)
(Etxepare 2003: 403):
These LVC s pattern with regular transitive constructions also in that their
NVE, qua incremental theme, can be quantified (58)a just like any incremen-
tal theme of a non-light verb (58)b (Oyharçabal 2006: 792):
An additional piece of evidence for the transitive analysis comes from causative
contexts (Ortiz de Urbina 1989: 47). The NVE behaves as a canonical argument
in the formation of causatives, since the subject of the LVC (59) takes the dative
case in the causative (60), just like the causee of a causativized transitive pred-
icate (61):
However, not all Basque LVC s display the same behaviour in the battery of syn-
tactic contexts just presented. For example, in contrast with (57), not all NVE s
can be partitive-marked, as pointed out by Etxepare (2003: 404) and Zabala
(2004: 451–452) and thoroughly assessed by Martinez (2015: 247–277):
Moreover, even those LVC s showing partitive case-marking do not exactly repli-
cate the behaviour of a common transitive structure; indeed, for LVC s such case
marking is not compulsory (64), whereas for regular transitive structures it is
(65):
Thus, the existence of the case-less alternates above, whether general (63)
or circumscribed to a particular dialect or register (Oyharçabal 2006) (64)a,
favours an incorporation analysis.31
Such an incorporation analysis would also allow us to account for the dif-
ferences regarding the adjacency between the NVE and the LV. As a matter of
fact, in contrast to what was shown in (53)–(56) for dantza egin, some NVE s
cannot be separated from the LV at all, unlike common DO s (Zabala 2004: 452–
453). Thus, the following examples, involving interrogative sentences (66) and
focus structures (67)–(69), should be contrasted with (54)–(56). Also, another
31 There is no exhaustive dialectal survey on this topic. Oyharçabal (2006) observes some
tendencies within dialects (and registers) but, in spite of some remarks on how some other
dialects behave, he mainly focuses on Basque eastern varieties, i.e., those spoken in the
French Basque Country (Iparralde).
210 acedo-matellán and pineda
sign of incorporation in the case of hitz egin (actually often written hitzegin)
comes from its phonological pattern, bearing one single accent: hitzégin, to be
contrasted, for instance, with that of negar egin, negár egín.
The position of manner adverbials such as ondo, ongi ‘well’ or gaizki ‘badly’ also
reveal differences between LVC s and ordinary transitive structures. Whereas
with the latter the unmarked order is S O Adv V (74)a, with the former two
options are possible: in (75)a, available at least in Western dialects, the LVC is
treated like a transitive structure, whereas in the more general (75)b (cf. Oyhar-
çabal 2006: 794), the LVC is treated like an intransitive structure—thus, by
hypothesis, signalling incorporation. Here again, we find variation depending
on the particular LVC. For example, hitz egin is not grammatical if treated as a
transitive structure (76)a, only the option involving incorporation is good (76)b:
So far we have seen that variation depends on the particular LVC (with cases
such as hitz egin favoring the incorporation analysis in general) but also on
geographic variation. In particular, Oyharçabal (2006) finds quite a general ten-
dency in Eastern dialects to avoid the treatment of the NVE as an object, thus
preferring incorporation—which otherwise is also possible in other dialects or
even idiolects. This observation nicely fits with Berro’s (2010:4) claim that “com-
plex unergatives are much more common in southern varieties [that is, those
in contact with Spanish: our clarification, VA&AP], whereas simplex are pre-
ferred in northern ones [that is, those in contact with French: our clarification,
VA&AP].”
The tendency shown by Eastern dialects can also be witnessed in the treat-
ment of LVC s in causative contexts: in contrast to what we saw in (60), Eastern
dialects do not always treat the causee of a LVC as the causee of a transitive
verb (with dative) (77), since it may also appear with absolutive case (78), as
the causee of an intransitive verb (thus indicating that incorporation has taken
place) (Epelde & Oyharçabal 2012):32
32 Epelde & Oyharçabal (2012) propose a contact-based explanation: in French, with which
Eastern dialects are in contact, the causee of unergative predicates is accusative (i); in
turn, Western dialects are in contact with Spanish, and in particular with the Spanish spo-
ken in the Basque Country, which is leísta and thus marks the causee of unergatives with
dative case (Tu le haces reír, Yo le hago bailar). Actually, with simplex unergatives (ii) the
dialectal divide is even more marked: whereas in Western dialects dative is more gener-
ally used (iii), in Eastern dialects absolutive is normally found (iv) (Ricardo Etxepare, p.c.;
Oyharçabal & Epelde 2012).
(i) Tu le fais rire, Je le fais danser
you cl.3sg.m.acc make.prs.2sg laugh, I cl.3sg.m.acc make.prs.1sg dance
‘You make him laugh, I make him dance’
(ii) Aduna-k dantzatu du.
Aduna-erg dance aux.prs.erg3pl.abs3sg
‘Aduna has danced.’
(iii) Aduna-ri dantz-arazi diote.
Aduna-dat dance-cause aux.prs.erg3pl.abs3sg.dat3sg
‘They have made Aduna dance.’
(iv) Aduna dantz-arazi dute.
Aduna dance-cause aux.prs.erg3pl.abs3sg
‘They have made Aduna dance.’
light verb constructions in basque and romance 213
Last, but not least, NVE s present a strong dissimilarity with DO s: whereas com-
mon DO s must bear a determiner (79) NVE s (generally) lack it (80) (Oyharçabal
2006: 793–794).
33 The -a ending of dantza does not correspond to the determiner, but is part of the nomi-
nal root, making ‘dance’ and ‘the dance’ formally equal. Compare: hitz ‘word’ ~ hitz-a ‘the
word’ ~ hitz bat ‘one word’ and dantza ‘dance’ ~ dantza ‘the dance’ ~ dantza bat ‘one dance’.
214 acedo-matellán and pineda
4 Conclusions
We have also paid attention to the NVE and its categorial properties. Basque,
unlike French and Spanish, often allows a type of LVC in which the nominal NVE
appears in the inessive case, or endowed with an adverbial suffix, the result-
ing LVC displaying peculiar aspectual properties. As regards unmarked nominal
NVE s, we have shown that they can be bare, i.e., determinerless and caseless,
in Basque and French—unlike Spanish and Catalan. For Romance, we have
also explored the different kinds of relations—semantic and morphological—
between LVC s and their synthetic counterparts.
Finally, a discussion has been dedicated to the degree of cohesion between
the LV and the NVE. It has been shown that there is empirical evidence pointing
in diverse directions as regards this issue: the NVE is in some cases quite syn-
tactically independent of the LV, in other cases it seems to be somehow fused
with it (i.e., incorporated into it), and, finally, there are cases that lie in between
these two extremes. Dialectal and lexical factors have been pointed out as pre-
dictors of this variation in Basque.
It should be clear from these conclusions that, much in the spirit of Butt’s
(2010) title, Basque (and Romance) LVC s constitute a jungle, and we are still
hacking away.
Acknowledgements
This work has been supported by the postdoctoral research fellowship Beatriu
de Pinós 2014 BP_A 00165 (Secretaria d’Universitats i Recerca del Departament
d’Economia i Coneixement de la Generalitat de Catalunya) (second author),
the postdoctoral research fellowship Juan de la Cierva-incorporación (IJCI-
2016–30474, Ministerio de Economía y Competitividad, Spanish Government)
(second author) and the research project FFI2014–56968-C4–1-P (Ministerio de
Economía y Competitividad, Spanish Government) (first and second author).
The research leading to this work began at Queens’ College, University of Cam-
bridge, in the case of the first author, and at CNRS-IKER, in the case of the
second author. We are grateful to Jaume Mateu and an anonymous reviewer
for very valuable comments. All errors remain our own.
References
Acedo-Matellán, Víctor. 2014. “Can we dance without doing a dance? Two opposite
views on the integration of roots in the syntactic structure of the vP”. In Asaf
Bachrach, Isabelle Roy & Linnaea Stockall (eds.), Structuring the argument. Mul-
216 acedo-matellán and pineda
Brody, Michael 2000. “Mirror theory: Syntactic representation in perfect syntax”. Lin-
guistic Inquiry 31: 29–56.
Butt, Miriam. 1995. The structure of complex predicates in Urdu. Standord, California:
CSLI Publications.
Butt, Miriam. 2003. “The light verb jungle”. In Gulsat Aygen, Claire Bowern, Conor
Quinn (eds.), Harvard Working Papers in Linguistics, vol. 9, 1–49. Papers from the
GSAS/Dudley House Workshop on Light Verbs.
Butt, Miriam. 2010. “The light verb jungle: Still hacking away”. In Mengistu Amberber,
Brett Baker, Mark Harvey (eds.), Complex Predicates: Cross-linguistic Perspectives on
Event Structure, 48–78. Cambridge: Cambridge University Press.
Butt, Miriam, & Wilhelm Geuder. 2001. “On the (semi)lexical status of light verbs”. In
Norbert Corver & Henk van Riemsdijk, H. (eds.), Semi-lexical Categories: On the Con-
tent of Function Words and the Function of Content Words. Berlin: Mouton de Gruyter.
323–370.
Butt, Miriam & Aditi Lahiri. 2013. “Diachronic pertinacity of light verbs”. Lingua 135:
7–29.
Bye, Patrick, & Peter Svenonius. 2012. “Non-concatenative morphology as epiphenom-
enon”. In Jochen Trommer (ed.), The Morphology and Phonology of Exponence. Ox-
ford: Oxford University Press. 427–495.
Cattell, Ray. 1994. Composite Predicates in English. Sydney: Academic Press Australia.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Cuervo, Mª Cristina. 2003. Datives at large. PhD thesis, MIT
Danlos, Laurence. 1992. “Support verb constructions: linguistic properties, representa-
tion, translation”. French Language Studies 2: 1–32.
De Rijk, Rudolf. 2008. Standard Basque: A Progressive Grammar. Cambridge, MA: MIT
Press.
Epelde, Irantzu & Beñat Oyharçabal. 2012. Resultados lingüísticos del contacto francés-
euskera: el cambio dativo > absolutivo en algunas construcciones causativas. Sociedad
Española de Lingüística. Universitat de València.
Etxepare, Ricardo. 2003. “Valency and Argument Structure in the Verb”. In Jose Igna-
cio Hualde & Jon Ortiz de Urbina (eds.), A grammar of Basque. Berlin: Mouton de
Gruyter. 369–465.
Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. PhD thesis, Euskal
Herriko Unibertsitatea / Universidad del País Vasco.
Giry-Schneider, Jacqueline. 1978. “Les Nominalisations En Français: L’opérateur ‘faire’
Dans Le Lexique”. Genève: Droz.
Giry-Schneider, Jacqueline. 1986. “Les noms construits avec “faire”: compléments ou
prédicats?”. Langue française 69, 49–63.
Goldberg, Adele. 1995. Constructions: A Construction Grammar Approach to Argument
Structure. Chicago; London: University of Chicago.
218 acedo-matellán and pineda
Larson, Richard K. 1988. “On the double object construction”. Linguistic Inquiry 19(3),
335–391.
Levin, Beth & Malka Rappaport. Hovav. 1995. Unaccusativity: At the Syntax-lexical Se-
mantics Interface. Cambridge, Mass.; London: MIT, 1995.
Lin, Tzong-Hong. 2001. Light Verb Syntax and the Theory of Phrase Structure. PhD thesis,
University of California, Irvine.
Marantz, Alec. 2005. “Objects Out of the Lexicon: Objects as Events”. Ms., MIT.
Martinez, Arantzazu. 2015. “[Izen + egin]a aditz-lokuzioak: inkorporazio mailak”. PhD
thesis, Deustuko Uniberstitatea.
Matushansky, Ora. 2006. “Head movement in linguistic theory”. Linguistic Inquiry 37,
69–109.
McFadden, Thomas. 2006. “German inherent datives and argument structure”. In
Daniel P. Hole, Abraham Werner & André Meinunger (eds.), Datives and Other Cases:
Between Argument Structure and Event Structure. Amsterdam; Philadelphia: John
Benjamins. 49–77.
Mithun, Marianne. 1984. “The evolution of Noun Incorporation”. Language 60, 847–893.
Mohanan, Tara. 1994. Argument Structure in Hindi. Stanford, California: CSLI Publica-
tions.
Ortiz de Urbina, Jon. 1989. Parameters in the Grammar of Basque. Dordrecht: Foris.
Oyharçabal. Beñat. 2006. “Basque Light Verb Constructions”. ASJU XL: 787–806.
Pesetsky, David & Esther Torrego. 2001. “T-to-C Movement: Causes and Consequences”.
In Michael Kenstowicz (ed.), Ken Hale: A Life on Language. Cambridge, MA: MIT
Press. 355–426.
Piera, Carlos & Soledad Varela. 1999. “Relaciones entre morfología y sintaxis”. In Vio-
leta Demonte & Ignacio Bosque (ed.), Gramática descriptiva de la lengua española,
vol. 3: Entre la oración y el discurso. Morfología. Madrid: Real Academia Española /
Espasa-Calpe. 4367–4422.
Pineda, Anna. 2016. Les fronteres de la (in)transitivitat. Estudi dels aplicatius en llengües
romàniques i basc [published and revised version of the PhD dissertation]. Barcelo-
na: Institut d’Estudis Món Juïc. Col·lecció Cum Laude, 6.
Pylkkänen, Liina. 2002. “Introducing Arguments”. PhD thesis, MIT.
Pylkkänen, Liina. 2008. “Introducing Arguments”. Cambridge, Mass.; London: MIT.
Rafel, Joan. 2004. “Los predicados complejos en espanyol”. In Elixabete Pérez Gaztelu,
Igone Zabala & Lluïsa Gràcia (eds.), Las fronteras de la composición en lenguas
románicas y en vasco. Donostia: Deustuko Unibertsitatea. 393–443.
Ramchand, Gillian C. 2008. Verb meaning and the lexicon: A first phase syntax. Cam-
bridge: Cambridge University Press.
Ramchand, Gillian C. 2014. “On structural meaning vs conceptual meaning in verb
semantics”. Linguistic Analysis 39(1–2), 207–243.
Rappaport, M. & Levin, B. 1988. “What to do with theta-roles”. In Wendy Wilkins (ed.)
Syntax and Semantics 21: Thematic relations. New York: Academic Press, 7–36.
220 acedo-matellán and pineda
Rappaport Hovav, Malka & Beth Levin. 1998. “Building verb meanings”. In Miriam Butt
& Wilhelm Geuder (ed), The Projection of Arguments: Lexical and Compositional Fac-
tors. Stanford, Calif.: CSLI Publications, 97–134.
Rodríguez, Sonia & Fernando García Murga. 2003. “IZEN + EGIN predikatuak euskaraz”.
In Jesús Makatzaga & Beñat Oyharçabal (eds.), Euskal gramatikari eta literaturari
buruzko ikerketak XXI. mendearen atarian, Gramatika gaiak, Iker-14. Bilbo: Euskal-
tzaindia. 417–436.
Sarasola, Ibon. 1977. “Sobre la bipartición inicial en el análisis en constituyentes”,
ASJU XI: 51–90.
Uribe-Etxebarria, Myriam. 1989. “Some Notes on the Structure of IP in Basque”. Ms.,
UConn, Storrs.
Wierzbicka, Anna. 1982. “Why Can You Have a Drink When You Can’t *Have an Eat?”
Language, Vol. 58, No. 4, 753–799.
Wood, Jim & Alec Marantz. 2017. “The interpretation of external arguments”. In Roberta
D’Alessandro, Irene Franco & Ángel Gallego (ed.), The verbal domain. Oxford: Oxford
University Press. 255–278.
Zabala, Igone. 2004. “Los predicados complejos en vasco”. In Elixabete Pérez Gaztelu,
Igone Zabala & Lluïsa Gràcia (eds.), Las fronteras de la composición en lenguas
románicas y en vasco. Donostia: Deustuko Unibertsitatea. 445–534.
chapter 7
Beatriz Fernández
1 Introduction
There is a long tradition of research dealing with datives in general and non-
selected datives in particular. These non-selected datives include external pos-
sessors, benefactives and ethical datives, among others (see for example, Bosse,
Bruening & Yamada 2012). In this chapter, I will analyze some of these datives,
particularly, ethical datives. In doing so, one of the major difficulties to over-
come is the extreme variation of designations and the wide range of properties
(not necessarily coherent with each other) attributed to them. For instance,
ethical datives have been considered to refer to the participants of the speech
act and mostly to the speaker (sometimes also to the hearer). If this were the
case, then ethical datives would be restricted to 1st (or 2nd) person and couldn’t
be doubled by a DP, among other properties (see Jaeggli 1986 for Spanish). Nev-
ertheless, not all the so-called ethical datives seem to be restricted to clitics or
1st and 2nd person. Hence, a more-fine grained typology seems to be necessary
in order to clarify this puzzling scenario.
In this chapter, I will follow Franco & Huidobro (2008) in distinguishing
three types of ethical datives in Spanish. First of all, Class I, reflexive non-
argumental clitics as in me comí la manzana ‘I ate the apple’, alternatively ana-
lyzed as aspectual by Miguel & Fernández Lagunilla 2000, and reminiscent of
personal datives (Horn 2008, 2013). Second, Class II of the ‘non volitional se +
dative clitic type’, as in se me caló el coche ‘the car stalled on me’. Ethical datives
from Class II correspond to affected experiencers in Bosse, Bruening & Yamada’s
(2012) typology or affected datives (dativos de afectación) in Maldonado (1994)
for Spanish and in Fernández (2010) for Basque. Third, Class III, as in me le
pegaron una paliza a mi primo ‘they beat my cousin up’ with a dative clitic clus-
ter including the ethical dative me and the goal dative le. This Class III shows
the most restricted version of ethical datives, as they seem to be necessarily 1st
and 2nd person (me le/te le/*se le) whereas ethical datives from Class I and II do
not. Besides, ethical datives from either Class III or Class I cannot be doubled
(me/*a mí) whereas those from Class II can.
On the other hand, unlike in Hispanic linguistics, ethical datives have been
only briefly mentioned but not analyzed in depth in Basque. The sole exis-
tence of ethical datives in Basque is not uncontroversial, as they are neither
restricted to 1st and 2nd person nor reduced to clitics (Fernández 2010). As a
consequence, some of the properties attributed to ethical datives in general
and Spanish ethical datives in particular are not met in Basque.
In this chapter, I will review classical syntactic tests (as in Borer & Grodzin-
sky 1986 or Jouitteau & Rezac 2007) and discuss more novel criteria proposed
by Bosse, Bruening & Yamada (2012) and Horn (2008, 2013). I will claim that
the only ethical datives attested in Basque are those related to Spanish ethical
datives from the Class II as neither Class I nor Class III is attested in Basque.
I will also claim that only Spanish (and not Basque) exhibits ethical datives in
their most restricted definition (Class III), i.e. 1st and 2nd person clitics referred
to the participants of the speech act with no DP-doubling among other prop-
erties. As we will see, ethical datives of the Class II contribute truth conditional
meaning—or at issue meaning in Potts’ (2005) designation—but contrary to
what is argued by these authors, they do not contribute non-truth conditional
meaning—or not-at-issue meaning—in either Basque or Spanish.
Although no particular analysis will be provided here, I will claim that ethi-
cal datives of Class II are merged in the Specifier of an Applicative head above
v, as argued in Odria (2017) for experiencers of psychological verbs and exter-
nal possessors. Thus, the similarity in the behavior of ethical datives of Class II,
external possessors and also benefactives follows. Besides, ethical datives from
Class I and III need further explanation that go beyond the limits of this
paper.
This chapter will be structured as follows. First of all, I will briefly present
non-selected datives in general and ethical datives in particular (section 2). Sec-
ondly, I will review and discuss some classical and novel tests on Basque and
Spanish ethical datives of Class II (section 3). Section 4 will present the main
conclusions and a brief discussion.
2 Ethical Datives
paper, some of the most salient properties of non-selected datives are pre-
sented and discussed in depth, as far as external possessors and ethical datives
are concerned. The basis of the typology of these non-selected datives and the
properties attributed to each type are still studied as a classical benchmark test.
Other studies such as Jouitteau & Rezac (2007) have explored the literature on
ethical datives in great detail in order to characterize this particular type of
non-selected datives in French.
Paradoxically, one of most influential recent works on datives, i.e. Pylkkä-
nen 2008 [2002] and her well-known high v low applicative distinction does
not solve the issue, as non-selected datives such as ethical datives and bene-
factives can be analyzed as high applicatives, blurring the distinction between
both of them. Besides, although external possessors are treated as low applica-
tives in her proposal, they behave similarly to ethical datives and benefactives
(high applicatives) in many respects.
An attempt to typify non-selected datives properly is due to Bosse, Bruen-
ing & Yamada (2012), an analysis that combines both syntactic and semantic
properties of non-selected datives. In particular, these authors analyze affected
experiencers which seem to correspond to a particular type of ethical datives
in Spanish, i.e. Class II in Franco & Huidobro’s (2008) typology. Unfortunately,
some of the properties attributed to affected experiencers, in particular their
contribution to non-truth conditional meaning and sentientness, are not met
in Spanish and Basque ethical datives of this class.
In this chapter, I will concentrate on ethical datives and I will leave aside
external possessors and benefactives, although I will refer to these non-selected
datives in order to compare the properties of one another. Both external posses-
sors and benefactives are also attested in the languages under discussion—see
for instance Arregi (2003), Fernández (2010), Fernández & Ortiz de Urbina
(2010) and Odria (2017) for Basque external possessors, and Maldonado (1994)
for a discussion on the distinction between benefactives and ethical datives
(named as datives of interest by the author) in Spanish.1 As we will see, the
label ethical datives involves quite different types of datives. Hence, a more fine-
grained typology of ethical datives is necessary in order to characterize them
properly.
2.2 Ethical Datives in Basque and Spanish: Some Data and a Three-Way
Typology
Let us present some basic data on Basque and Spanish ethical datives. In prin-
ciple, an ethical dative is a non-selected one, i.e. a dative not included in the
argument structure of the verb. This ethical dative experiences the positive
or negative effect of an event, thus, it seems to be affected by the event. For
instance, Basque examples in (1) show a dative of this sort:
The sentences in (1) include a 1st person singular dative clitic -t in both the
bivalent unaccusative form (zait) and in the ditransitive one (dit).3 The datives
in these sentences are non-selected arguments, as their counterparts with no
dative are equally available and grammatical in Basque:
2 Abbreviations: abs absolutive, acc accusative, adl adlative, allo allocutive, aux auxiliary,
cl clitic, dat dative, d determiner, ed ethical dative, ed-ii ethical dative of Class II, ed-iii
ethical dative of Class III, erg ergative, fam familiar, fem feminine, mas masculine, nom
nominative, pl plural, psr possessor, pst past, sg singular, 1, 2, 3 for first, second and third
person.
3 Basque is a language with auxiliary alternation: BE is selected with unaccusative predicates
and HAVE with unergative and (di)transitive predicates. Besides, Basque verbal forms include
clitics/agreement markers for absolutive, ergative and dative arguments. For ease of exposi-
tion, I will gloss the verbal form as aux (auxiliary) + cl.dat (dative clitic) when it includes a
dative clitic, and as aux, when it does not. For ease of understanding, further details will be
omitted.
on non-selected datives: ethical datives in basque and spanish 225
The datives in (1) experience the effect of the kid falling asleep (1a) and the kid
breaking the book (1b). The examples in (1) are equivalent to Spanish examples
in (3).
Ethical datives of the ‘(non volitional) se + dative clitic type’ type as the one in
example (3a) are classified as ethical datives within Class II in Franco & Huido-
bro’s (2008) three-way typology. I will name them as Ethical Datives-Class II
(ED-II). Other designations have been used for the very same datives, such as
affected datives (Spanish dativos de afectación) in Maldonado (1994) for Spanish
and Fernández (2010) for Basque. These ED-II seem to correspond to affected
experiencers in Bosse, Bruening & Yamada (2012).
The interpretation of the ethical dative in (1a) and (3a) can be both negative
and positive, whereas it seems to be negative in (1b) and (3b). Thus, the desig-
nation affected experiencer is used instead of malefactive or adversative in some
works, as explicitly said in Bosse & Bruening (2010).
Nevertheless, the malefactive interpretation of datives involved in construc-
tions similar to (1b) and (3b) is not obligatory. For instance, if instead of ‘break
the book’, we say ‘pass the exam’, the interpretation of the dative will be, in
principle, positive. Therefore, I will assume that both benefactive and malefac-
tive interpretations are available for examples such as (1) and (3), an assump-
tion also made by Maldonado (1994) and Franco & Huidobro (2008) for Span-
ish.
It is worth pointing out that that the (di)transitive structure in (3b) is not
included in Franco & Huidobro’s Class II, as it does not involve the clitic se.
Leaving aside this clitic se, the main difference between both constructions
resides in the fact that the former example (3a) is intransitive whereas the
latter is transitive. As far as Basque is concerned, I see no reason to exclude
transitive structures in this ED-Class II. Actually, Maldonado (1994) includes
both intransitive and transitive structures when dealing with ED-Class II. More-
226 fernández
The Spanish example (5a) shows some restrictions that lead Franco & Huido-
bro (2008) to typify it as ED-Class III. However, its Basque counterpart shows
exactly the same behavior of ED-Class II. In this point Spanish and Basque split.
I will come back to this discussion in section 3.2.
on non-selected datives: ethical datives in basque and spanish 227
Compared to ED-Class II, the ED-Class III exemplified in (6) shows the oppo-
site properties: 1) it can only be 1st or 2nd person; 2) it cannot be doubled by a
DP (it is a weak pronoun); 3) it can be freely combined with any other dative
argument; 4) it does not contribute truth conditional meaning, as it cannot be
questioned or bind a variable in the truth-conditional tier; and 5) it only con-
tributes non-truth conditional meaning (an implicature).
Although Basque allocutives (Oyharçabal 1993) fall outside the scope of this
chapter, they share all the properties attributed here to by ED-Class III, with
the particularity that allocutives can only be 2nd person, as in (7). This exam-
ple includes an allocutive marker: a 2nd person singular morpheme that can
be either masculine -k or feminine -n and corresponds to a familiar addressee.
4 The datives involved in Maldonado’s (1994) examples can be argued to be external possessors
(and not ED-Class II), as there is a relationship of inalienable possession between the dative
(the possessor) and a PP including a body part (the possessum), as in (i).
(i) Se me/te/le murió en los brazos.
cl cl.1/2/3SG.dat die in the arms
‘He/she died on me/you/him-her / He/she died in my/your/his-her arms.’
Nevertheless, similar examples including a possessor inside the PP would be also grammati-
cal: en los brazos de Juan ‘in John’s arms’. Hence, the datives involved in (i) can be ED and not
necessarily external possessors.
on non-selected datives: ethical datives in basque and spanish 229
Note that the clitic cluster se le would be grammatical in a context where the
clitic se were co-referential with the subject, that is, ED-Class I (reflexive non-
argumental dative clitic) combined with a ED-Class II. To the contrary, a se clitic
not co-referential with the subject would be out.
Other ED-Class III are not so easily explained, but seem to exhibit the same
constriction:
In Basque, nothing prevents the ED from being 3rd person. Hence, we can spec-
ulate with the idea that the inability for the ED in (11) to be 3rd person cannot be
only due to its nature but to the morphological and syntactic constraints that
govern particular clitic combinations. Be that as it may, Basque and Spanish
seem to separate in this respect.
Summing up, nothing prevents ED-Class II from being 3rd person. With
regard to this, ED-Class II behaves in the same way external possessors, bene-
factives and any other argumental datives such as experiencers do. Being non-
argumental, ED-Class I also aligns with ED-Class II in this respect. It remains an
open question whether ED-Class III is restricted or not to 1st (and 2nd person),
as frequently discussed in the literature. The examples provided in (10) point
in that direction.
As can be seen in Perlmutter and Jaeggli’s classical examples the 1st person
singular dative clitic cannot be doubled (me … *a mí)—note that the 3rd per-
son singular dative clitic le in Perlmutter’s example (11a) can be doubled, as
the presence of the DP (a esa chica) reveals. This clitic corresponds to the ED-
Class III of the typology assumed in this chapter. To the contrary, ED-Class II
does allow doubling dative clitics in Spanish:
on non-selected datives: ethical datives in basque and spanish 231
b. Se le escapó al policía.
cl cl.3sg.dat get away to the police
‘He/She got away from the police.’ (Maldonado 1994: 260)
This property is closely related to the dative being or not an argument. Thus,
for instance, Franco & Huidobro (2010: 218, section 3) assume that the possibil-
ity of having a DP in a case position doubled by a clitic reflects the argumental
status of that clitic. Being this the case, ED-Class II are argumental, as shown
in (12), whereas ED-Class I and ED-Class III are not—(4a) and (4b) respectively
repeated here as (14a) and (14b):
spective taken by this author in her (2015) work and in Bosse, Bruening and
Yamada (2012) among others, ED are non-selected arguments of particular
heads.
In this chapter, I will assume ED-Class II to be arguments in their more gen-
eralized meaning. In particular, I will propose that these ED-Class II behave
exactly as benefactives, experiencers and external possessors. I will follow
Odria (2017) in assuming the very same merging position for experiencers and
external possessors, i.e. the Specifier of an Applicative (Appl) head above v.
Moreover, I will claim that these ED-Class II merge in this position. Hence,
the properties attributed here to ED-Class II follow directly. In particular, as
observed by Franco & Huidobro (2008), it is worth noting that the presence of
the dative clitic doubling the DP is not optional but obligatory in Spanish inver-
sion predicate structures, a particular instance of ED-Class II. The designation
refers to structures with a subject like dative and a nominative theme with an
agreeing verbal form, as in examples (17). The absence of the dative clitic entails
ungrammaticality:
Note that the Basque examples include an absolutive marked argument instead
of the corresponding nominative theme in Spanish counterparts. In order to
be grammatical, the presence of the dative clitic is obligatory, as pointed out by
Fernández (2010) Fernández & Ortiz de Urbina (2010) and recently Odria (2017)
among others. This is not surprising, as leaving aside north-eastern varieties
on non-selected datives: ethical datives in basque and spanish 233
e. Me le rompió la nariz.
ed cl.3sg.dat read break the nose
Intended: ‘He/she broke his/her nose (on me).’ (ED-Class III + External
possessor)
This is particularly interesting as the presence of two dative clitics shows that
ED-Class III, whatever they are, need to be separated from the rest, specifi-
cally, from ED-Class II (19d). On the other hand, ED-Class II, benefactives and
external possessors cannot be combined with one another. Assuming, as I do,
that all these non-selected datives are merged in Spec, Appl, as Odria (2017)
do for Basque experiencers, external possessors and causees,5 the inability of
ED-Class II to be combined with all these non-selected datives follows.
On the other hand, Basque can never combine two dative clitics, whether
they be selected or not (Fernández 2010), as there is a sole slot for a dative clitic
in the morphological configuration. Therefore, the Basque counterpart (20a) to
Perlmutter’s example in (13a) is ungrammatical. Besides, the ungrammaticality
affects both ditransitive auxiliary forms (20a) and intransitive auxiliary forms
(20b): two dative clitics never combine with each other. In (20a) the morpheme
-t can only be interpreted as an ergative and never as a dative.
As a consequence, this test does not apply in Basque. Nevertheless, the exam-
ples in (20) indirectly show that there is nothing comparable to ED-Class III in
Basque.
5 I am deliberately ignoring other double dative constructions, such as those including dif-
ferentially marked objects and goals or any other including derived datives (mainly goals),
exhaustively analyzed by Odria (2017). For an analysis of other double dative constructions
including an experiencer and an applied dative without verb, see in Berro & Fernández (to
appear).
on non-selected datives: ethical datives in basque and spanish 235
Notice that (23b) would be grammatical when asking about the one affected
by the book being broken, but not about the one involved in the book being
broken to him/her. It is worth pointing out that (23b) would be perfect if the
clitic le were coreferential with ‘to whom’.
Besides, ED-Class II can be a quantifier that binds a truth conditional ele-
ment as a variable. Thus, ED-Class II contribute truth conditional meaning with
regard to binding too. Basque examples show the pattern:
If this is on the right track, then the value for both Basque and Spanish is yes,
ED-Class II contribute truth conditional meaning.
As argued by Bosse & Bruening (2010) and Bosse, Bruening & Yamada (2012),
the experience itself cannot be negated: it is the whole verbal event, but not
the experience, that is negated. However, at least in Basque and Spanish, ED-
Class II can be negated. Hence, the meaning of the experience does not go
beyond negation (against what has been observed by these authors for lan-
guages such as Albanian and German among other languages). If this is so,
ED-Class II do not contribute non-truth conditional meaning. Hence, it does
not confirm the yes value attributed to affected experiencers by the authors
with regard to this property. This unexpected value is also attested in ditransi-
tive structures such as (27a) and (27b), negative counterparts of examples (1b)
and (3b). Once again, it seems that not only the verbal event but also the expe-
rience itself can be negated:
238 fernández
The reason for that might be that, contrary to Maldonado (1994) and this chap-
ter, the datives in (1b) and (3b) and their corresponding negation in (26a) and
(26b) do not show an affected experiencer but a benefactive—assuming that
there is a distinction between affected experiencers and benefactives, some-
thing not so obvious attending to our results (see Table 7.1 below). If they
were benefactives and not affected experiencers, then, as in other benefactives
analyzed by Bosse & Bruening (2010), the benefactive relation itself could be
negated, as benefactives are entirely truth conditional.
table 7.1 Typology of non-selected datives inspired by Bosse & Bruening (2010), extended
and modified
extended and modified mainly, but not only, on the basis on Franco & Huido-
bro (2008: 223). Besides, I have maintained the general designation along with
its particular classes, ED-Class I, III and II, instead of subject co-referential, atti-
tude holders and affected experiencers respectively.
The table also includes allocutives, which share the values of the properties
attributed to ED-Class II, although this only covers a part of the intricate nature
of allocutives not necessarily involving dative-like clitics.
One more property could be added to the table, that is, subject co-reference
attested in ED-Class I with a yes value, and with a no value in the remaining
cases. Nevertheless, person restriction distinguishes both ED-Class I and III.
Thus, it has not been included in the table.
Besides, there is a striking (and undesirable) scenario with regard to ED-
Class II, benefactives and external possessors: they show exactly the same
properties and cannot be distinguished from one another. The property ‘NP
must be sentient’ analyzed by Bosse, Bruening & Yamada (2012) could in prin-
ciple distinguish ED-Class II and benefactives, being ED-Class II sentient and
benefactives not necessarily. Unfortunately, ED-Class II do not seem to be nec-
essarily sentient in either Basque or Spanish, as shown in the following exam-
ples:
Acknowledgements
The author would like to thank Jon Ortiz de Urbina for his valuable comments
on a previous version of this paper and to Solveig Bosse for some feedback.
All errors remain my own. The author also acknowledges the research funding
received from 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 Program for research, technological development and demonstra-
tion under grant agreement no. 613465.
References
Arregi, Eurídice. 2003. “On possessor raising effects in Basque”. Ms., UPV/EHU.
Authier, Marc & Lisa Reed. 1992. “On the syntactic status of French affected datives”.
The Linguistic Review 9: 295–311.
Berro, Ane & Beatriz Fernández. To appear. “Applicatives without verbs”. Natural Lan-
guage and Linguistic Theory (NLLT).
Bosse, Solveig & Benjamin Bruening. 2010. “Benefactive versus experiencer datives”. In
on non-selected datives: ethical datives in basque and spanish 241
Mary B. Washburn, Katherine McKinney-Bock, Erika Varis, Ann Sawyer & Barbara
Tomaszewicz. (eds.), Proceedings of 28th WCCFL. 69–77.
Bosse, Solveig, Benjamin Bruening & Masahiro Yamada. 2012. “Affected experiencers”.
Natural Language and Linguistic Theory (NLLT), 30(4): 1185–1230.
Bosse, Solvig. 2015. Applicative arguments. Peter Lang.
Christian, Donna. 1991. “The personal dative in Appalachian English”. In Peter Trudgill
& J.K. Chambers (eds.), Dialects of English. London: Longman. 11–19.
Borer, Hagit & Yosef Grodzinsky. 1986. “Syntactic cliticization and lexical cliticization:
The Case of Hebrew Dative Clitics”. In Hagit Borer (ed.), Syntax and Semantics 19.
New York: Academic Press. 175–215.
Fernández, Beatriz. 2010. “Goi datiboak eta are goragokoak euskaraz: jabe datiboak
eta datibo hunkituak vs. datibo etikoak”. Gogoa 10, 1–2. San Sebastián/Donostia:
UPV/EHU. 1–20.
Fernández, Beatriz & Jon Ortiz de Urbina. 2010. Datiboa hiztegian. Bilbo: UPV/EHU.
Franco, Jon & Huidobro, Susana. 2008. “Ethical Datives, Clitic doubling and the Theory
of pro” Joyce Bruhn de Garavito & Elena Valenzuela (eds.), Selected Proceedings of
the 10th Hispanic linguistics Symposium. Somerville: Cascadilla Proceedings Project.
215–224.
Gutierrez Ordoñez, Salvador. 1999. “Los dativos”. In Ignacio Bosque & Violeta Demonte
(eds.), Gramática descriptiva de la lengua española II. Madrid: Real Academia Espa-
ñola / Espasa Calpe. 1855–1930.
Horn, Laurence R. 2008. “I love me some him: The landscape of non-argument datives”.
In Olivier Bonami & Patricia Cabredo Hofherr (eds.), Empirical Issues in Syntax
and Semantics 7, CSSP, Paris, 169–192. [On-line publication, http://www.cssp.cnrs.fr/
eiss7].
Horn, Laurence R. 2013. “I love me some datives: Expresive meaning, free datives, and
F-implicature”. In Daniel Gutzmann & Hans M. Gärtner (eds.), Beyond Expressives:
Explorations in Use-Conditional Meaning. Leiden/Boston: Brill. 153–201.
Jaeggli, Osvaldo A. 1982. Topics in Romance Syntax. Dordrecht: Foris.
Jaeggli, Osvaldo A. 1986. “Three issues in the theory of clitics: case, doubled NP s and
extraction”. In Hagit Borer (ed.), Syntax and Semantics 19. New York: Academic Press.
15–42.
Jouitteau, Mélanie & Milan Rezac. 2007. “The French ethical dative, 13 syntactic tests”.
Bucharest Working Papers in Linguistics, IX (1): 97–108.
Maldonado, Ricardo. 1994. “Dativos de interés sin intereses”. Verbo y estructuras frási-
cas [Actas do IV Coloquio Internacional de Lingüística Hispánica]. Porto, Facultad
de Letras. 241–264.
Miguel, Elena de & Marina Fernández Lagunilla. 2000. “El operador aspectual se”.
Revista Española de Lingüística 30 (1). 13–43.
Odria, Ane. 2017. The syntax of Differential Object Marking and datives in Basque. Ph.D.
thesis, UPV/EHU.
242 fernández
Ane Odria
1 Introduction
1 DOM is optional for many southwestern Basque speakers and its actual distribution is subject
to dialectal as well as idiolectal variation. Besides, its appearance is highly reduced with third
person objects.
In both Basque and Spanish, DOM objects show the same morphology as
the dative objects of bivalent unergative predicates of the lagundu (Basque) /
ayudar (Spanish) ‘accompany, help’ type: (i) nominal as well as verbal dative
marking in Basque, (ii) a-marking in Spanish, and (iii) cliticization with le(s) in
(Basque) Spanish. However, in the case of bivalent unergatives, the marking of
the object happens to be independent from factors like animacy and specificity
and, semantically speaking, the object patterns more akin to the goal dative of
ditransitive predicates. As a consequence, it has been argued that, contrary to
the direct object configuration of DOM objects, the dative objects in bivalent
unergatives have an indirect object configuration—see Fernández & Ortiz de
Urbina (2010, 2012) and Ortiz de Urbina & Fernández (2016) for Basque, Tor-
rego (2010) and Fábregas (2013) for Spanish, and Pineda (2016) for Romance
languages in general.
This paper makes a further step in the syntactic distinction between Basque
and Spanish DOM on the one hand, and dative objects in bivalent unerga-
tives on the other. It claims that apart from their syntactic configuration, these
objects are also distinguished by their categorical status. While DOM objects
pattern with causee, experiencer and possessor datives in exhibiting a DP
syntactic category, the datives in bivalent unergatives behave more akin to
goal datives and thus show a PP-like behavior. This contrast is evidenced by
the licensing of depictive secondary predication, which—along with the rest
of DP s—is allowed with DOM datives, but—as happens with PP s and goal
datives—is generally rejected with datives in bivalent unergatives (Odria 2015,
2017).
The paper is structured as follows. In section 2, I lay out the main properties
of the DOM found in Basque (section 2.1), in Spanish (section 2.2) and in Basque
Spanish (section 2.3). Section 3 compares DOM objects with dative objects in
bivalent unergatives in both Basque (section 3.3) and Spanish (section 3.4). Sec-
tion 4 deepens on the syntactic distinction exhibited by these types of objects.
Based on the licensing of depictive secondary predication, it argues that, while
DOM objects are DP s categorically, the datives in bivalent unergatives show a
PP-like—i.e., a dual DP/PP—character. This is again proved with both Basque
(section 4.1) and Spanish (section 4.2) data. Finally, chapter 5 closes the paper
by summing up the main conclusions.
differential object marking in basque and spanish dialects 245
In (2), the subject is marked with ergative case (-k), but the object appears with
dative case (-(r)i) rather than with the absolutive. Besides, the two arguments
are coded by their respective markers in the finite verbal form: the second per-
son ergative by -zu and the first person singular dative by -da-.
Contrary to what happens in the DOM of widely studied accusative lan-
guages, in Basque the differential marking co-appears with the ergative mark-
ing of the transitive subject, leading thereby to a configuration with two argu-
ments bearing a morphologically marked case: the ergative and the dative.3
This is an unexpected pattern from a typological point of view, because the
canonical configuration in transitive clauses is supposed to bear a single
2 The following abbreviations will be used in the text: e (ergative), d (dative), a (absolutive),
ine (inesive), acc (accusative), all (allative), dest (destinative), caus (causative), fut
(future), prog (progressive), nom (nominalizer), comp (complementizer), m (masculine),
f (feminine), aux (auxiliary verb), sg (singular), pl (plural), 1/2/3 (first, second and third per-
son).
3 Basque is not an exception in this regard. Other languages like Hindi also display DOM
together with an ergative subject (Mahajan 1990, Bhatt & Anagnostopoulou 1996).
246 odria
In (3), the subject is marked with ergative case (-k), the direct object with the
absolutive (-ø) and the indirect object with the dative (-(r)i). Likewise, the three
arguments are cross-referenced by the finite verbal form, which shows the same
shape as that in (2): the second person ergative is coded by -zu and the first per-
son singular dative by -da-.5
Despite being morphologically identical to indirect objects, DOM objects
share the same thematic as well as argumental relationships with absolutive
direct objects. Hence, it seems reasonable to assume that DOM objects generate
in the same syntactic position as canonical absolutives—i.e., in the comple-
ment position of V. This is common ground in the literature on the topic, given
that, in spite of the differential marking, DOM objects are considered to be
direct objects configurationally—see, among others, Bárany (2018).
4 With few exceptions, Romance and Semitic languages are clear examples of this typolog-
ical tendency, and so are other modern Indo-Aryan languages like Hindi or Punjabi and
Amerindian languages like Guaraní and Aymara (Bossong 1991: 157).
5 In Basque the third person absolutive is not overtly marked in the verbal form.
differential object marking in basque and spanish dialects 247
Besides, even though the main cutting point between DOM and non-DOM
objects is determined by humanness, person deserves special attention too.
In fact, many DOM speakers make a further distinction within human objects
and distinguish between first and second person objects on the one hand, and
third person objects on the other (Fernández & Rezac 2010, 2016; Mounole 2012;
Odria 2012, 2014, 2017; Rodríguez-Ordóñez 2016). As a consequence, it is easier
to find dative marking with first and second person objects than with third per-
son. This distinction is realized in different manners. For some speakers, only
first and second person objects can carry dative marking with a verb like ikusi
‘see’. As shown in (5) and (6), this is the case of a speaker from Zumaia (Central
Basque).7
6 Some DOM speakers consider examples like (4a) quite marginal, and add that the dative mark-
ing of a third person object would be more natural with other verbs. This may be due to the
social stigmatization of the DOM phenomenon with the verb ikusi ‘see’, as it is one of the most
corrected errors at school as well as in Basque academies.
7 The dialectal data presented in this paper were collected in the fieldwork carried out by
myself in previous work—see especially Odria (2017).
248 odria
b. Kali-an ikusizut.
street-ine see.aux[2sg.d-1sg.e]
‘I have seen you in the street.’
For other speakers, the distinction between first and second vs. third per-
son objects is not so rigid. A speaker from Larrabetzu (Western Basque), for
instance, finds DOM with first (7a) and second (7b) person very natural, but
prefers the absolutive for the third one—although in present tense the dative
is acceptable in this case too (8).
The different behavior attested between first and second vs. third person
objects is also mentioned in some descriptions of dialectal varieties. Hualde,
Elordieta & Elordieta (1994: 125–127), for instance, notice that in Lekeitio
Basque (a Western variety), DOM occurs more frequently with first and sec-
ond person than with the third. Likewise, Ibarra (1995: 427) reports that in
Ultzama Basque (a Navarrese variety), first and second person objects are gen-
differential object marking in basque and spanish dialects 249
erally marked dative, while DOM with third person objects is just optional.
Furthermore, it is important to note that, generally speaking, the DOM exam-
ples provided by dialectal grammars involve usually objects of first and second
person.
Along with animacy—or even person—, specificity is also an important
conditioning of Basque DOM. As argued by Mounole (2012), only human
objects with a referential interpretation are able to display the differential
marking—see also Fernández & Rezac (2010, 2016) and Odria (2012, 2014, 2017).
Mounole (2012: 368–369) shows that DOM is mostly incompatible with the
indefinite determiner bat ‘a’, the indefinite quantifier asko ‘many’ or the indef-
inite personal pronoun inor ‘nobody’. This is illustrated by the examples in
(9), (10) and (11), provided by a speaker from Itsasondo (Central Basque). This
speaker admits DOM with first, second and—less commonly—third person
objects. However, in order to bear dative marking with the verb ikusi ‘see’, those
of third person have to be not only human, but also definite. As a consequence,
as happens with many other speakers, human objects containing indefinite
determiners like bat ‘a’ (9), asko ‘many’ (10) and inor ‘nobody’ (11) tend to bear
absolutive marking.
Contrary to Mounole (2012), I thus conclude that instead of specificity per se,
the main factor conditioning the marking of the object along with animacy is in
fact definiteness. Basque DOM is then more restrictive than Romance languages
like Spanish, where specificity rather than definiteness seems to be relevant.
All in all, despite its significant impact, it should be noted that—as happens
in many Romance languages (Bossong 1991: 160–161)—, in Basque definite-
8 Instead of ‘specificity’, Mounole uses the term ‘referentiality’. However, for the sake of sim-
plicity, I continue using the notion ‘specificity’, assuming that both share the same meaning.
9 At least in some southwestern varieties, the sentence in (12b) would sound more natural with
the verb behar ‘need’, which—as happens with bilatu ‘look for’—requires absolutive marking
with indefinite objects (i). In these varieties, bilatzen dut ‘look for.prog aux’ would occur in
a verbal periphrasis formed with the verb ibili ‘be’ (ii):
(i) Idazkari bat/*-i behar dut / *diot.
secretary one.a/bat-d need aux[3sg.a-1sg.e] / aux[3sg.d-1sg.e]
‘I need a secretary.’
(ii) Idazkari bat-en bila nabil.
secretary one-ine look for be.aux[1sg.a]
‘I’m looking for a secretary.’
differential object marking in basque and spanish dialects 251
ness does not condition the marking of the object as rigidly as animacy does.
Whereas DOM is never attested with non-human or inanimate objects, in some
cases indefinite objects can occasionally occur with the differential marking.
This groups the DOM found in Basque with Spanish DOM, where animacy is also
known to be more influential than specificity (Brugger & Brugé 1996, Leonetti
2004).
Besides, although their influence is not as generalized as that caused by the
referential properties of the object, clausal properties like tense and finiteness
affect the marking of the object as well, given that DOM is sometimes reduced
in present tense as well as non-finite contexts (Fernández & Rezac 2010, 2016;
Odria 2017). This is a particular aspect of Basque DOM, as cross-linguistically
clausal factors like tense and finiteness do rarely condition the distribution of
DOM. Be that as it may, the influence of these factors is far from being a system-
atic pattern and seems to be related to morphological facts. Likewise, it should
be noted that the actual distribution of Basque DOM is also determined by the
nature of the verb, since the differential marking is more common with some
verbs than with others (Mounole 2012, Rodriguez-Ordóñez 2016, Odria 2017).
10 Contrary to what happens in Basque Spanish, the leísmo in Standard Spanish refers to the
coding of masculine, singular (usually personal) direct objects by the dative clitic le, sub-
stituting this way the accusative lo—and only exceptionally la (Fernández-Ordóñez 1994:
7, 1999: 1319; Landa 1995: 152). Consequently, the Basque Spanish leísmo is also referred as
real leísmo (Fernández-Ordóñez 1999) or animated leísmo (Urrutia-Cárdenas 1995, 2003).
11 Urrutia-Cárdenas (2003: 292) mentions that the use of le(s) can also be extended to inan-
imate objects. Nevertheless, I agree with Landa (1995: 8) in claiming that such pattern is
hardly acceptable among Basque Spanish speakers.
differential object marking in basque and spanish dialects 253
b. Lo / *le vi aparcado.
3sg.acc / 3sg.d saw.1sg parked
‘I saw it parked.’
b. Le vi contento/-a.
3sg.d saw.1sg happy.m/-f
‘I saw him/her happy.’
12 This is rejected in the Standard leísmo, where the same object would only be differentially
marked in the nominal by means of a-marking.
13 Landa (1995: 162) claims that, instead of specificity, in Basque Spanish clitic doubling with
le(s) is constrained by presuppositionality and gives the example in (i) to support her
hypothesis.
254 odria
object la niña ‘the child’ is a-marked, and the doubling with the dative clitic
le is acceptable. On the contrary, in (18b), the object el libro ‘the book’ is inan-
imate, and thus, non-a-marked. As a consequence, doubling it with a dative
or accusative clitic is ruled out. The same is true for the animate non-specific
object camareros ‘waiters’ in (18c).
Note that this is also the case in Basque DOM, as only those objects bearing
dative case are able to be coded by dative marking in the agreement complex
(19).
Summing up, in spite of their differences, we see that Basque DOM and
(Basque) Spanish DOM coincide in a great extent: (i) DOM is triggered by ani-
macy and specificity, (ii) the differential marking is morphologically identical
to the marking of the indirect object, and (iii) together with the noun phrase,
the differential marking is also realized in the finite verbal form. These shared
aspects are relevant because, as mentioned in section 1, DOM in Basque only
exists in the dialects that are in contact with Spanish.
In the literature on the topic, some authors have suggested that Basque
influences the leísmo in Basque Spanish. As Basque makes no gender dis-
tinction on direct objects, authors like Fernández-Ordóñez (1994, 1999) and
Landa (1995) have suggested that this could have reinforced the use of the
dative le(s) instead of lo(s) for masculine objects and la(s) for feminine ones.
Be that as it may, the so-called contact situation has also been analyzed in
the opposite direction. Mounole (2012), for instance, attributes the spread
of Basque DOM to the influence of Spanish, claiming that Basque DOM is
at least reinforced by the contact with Spanish. Other authors like Austin
(2006, 2015) or Rodríguez-Ordóñez (2013, 2016) go even further and claim
that Basque DOM is not only reinforced, but also induced by the contact with
Spanish—and more precisely, by the leísta Basque Spanish. This is, thus, an
open issue.
14 Along with lagundu ‘accompany, help’, jarraitu ‘follow’ and begiratu ‘look at’ bivalent
unergatives involve other verbs like abisatu ‘notify’, barkatu ‘forgive’, bultzatu ‘push’, deitu
‘call’, entzun ‘hear, listen to’, eskertu ‘thank’, heldu ‘hold’, obeditu ‘obey’ or ukitu ‘touch’—
see Etxepare (2003) and Fernández & Ortiz de Urbina (2010) for a complete classification
of this group of predicates.
256 odria
It is true that in the case of some verbs animacy has a say in the dative marking
of the object. As noted by Fernández & Ortiz de Urbina (2010: 164), this is the
case of itxaron/itxoin ‘wait for’. These authors show that, with few exceptions,
in the General Basque Dictionary Orotariko Euskal Hiztegia (OEH) (Mitxelena &
Sarasola 1989–2005), the object of itxaron/itxoin ‘wait for’ is generally marked
dative when animate (23a), and either dative (23b) or absolutive (23c) when
inanimate.17
However, in spite of its influence in verbs like itxaron/itxoin ‘wait for’, a closer
look to the different types of bivalent unergatives indicates that animacy is not
a conditioning trigger for the dative marking in these verbs. Its influence is only
particular of certain verbs or even dialects/speakers, and it is not as system-
atic as in DOM. Likewise, person is neither determinant in marking the object
dative, and the same thing happens with specificity. Consider now the exam-
ples in (24).
In (24), we see that the same indefinite particles that are generally excluded
with DOM are in fact acceptable with dative objects of bivalent unergatives.
I have already pointed out, northeastern dialects show a preference to mark both animate
and inanimate objects absolutive in bivalent unergative predicates.
differential object marking in basque and spanish dialects 259
rrego 1998, 2010; Leonetti 2004, Fábregas 2013). Similar to the Basque bivalent
unergatives, in Spanish the animate object of certain verbs carries a-marking
even with a non-specific reading. This is exemplified by Leonetti (2004: 85) with
the verbs entrevistar ‘interview’ (25a) and admirar ‘admire’ (25b).18
Likewise, Fábregas (2013: 29) adds that a-marking can also be assigned to inan-
imate objects with other verbs. In (26a), for instance, it is possible for the
inanimate object of acosar ‘harass’ to bear a-marking, and the marking is even
obligatory for the inanimate object of ayudar ‘help’ in (26b) (Fábregas 2013:
29).
Similar to what happens in Basque, verbs like acosar ‘harass’ or ayudar ‘accom-
pany, help’ show a parallelism with pure transitives involving DOM, as both are
bivalent and both contain an a-marked object. Notwithstanding, contrary to
the object of pure transitives, animacy and specificity do not necessarily hold
in these verbs. Therefore, as in Basque, not all a-marked objects that occur in
verbs entailing a bivalent argument structure involve DOM.
18 A more complete list of this group of verbs includes acusar ‘accuse’, admirar ‘admire’, afec-
tar ‘affect’, ayudar ‘accompany, help’, castigar ‘punish’, entevistar ‘interview’, golpear ‘beat’,
insultar ‘insult’, odiar ‘hate’, offender ‘offend’, saludar ‘greet’ and sobornar ‘bribe’ (Fábregas
2013: 27–28).
260 odria
Recall that this is also the main insight in Fernández & Ortiz de Urbina
(2012), Ortiz de Urbina & Fernández (2016) and Pineda (2016). These authors
claim that in Basque the dative object in bivalent unergatives is syntactically
equivalent to the indirect object in ditransitive constructions involving a silent
direct object. In fact, the Spanish correspondences in (28a) and (28b) are
equally obtained in Basque bivalent unergatives. Consider, for instance, the
‘bivalent unergative/ditransitive’ correspondence between deitu/dei egin ‘call’
(29a), bultzatu/bultza egin ‘push’ (29b) or lagundu/laguntza eman ‘help’ (29c).
Hence, it seems reasonable to think that in both Basque and Spanish the single
object that appears with dative morphology is in fact an indirect object in these
cases.
Having coupled the syntactic configuration of DOM objects with canonical
absolutives and that of the datives in bivalent unergatives with indirect objects,
the following section makes a further step in the syntactic distinction attested
between these kinds of objects. It argues that the difference between DOM
and dative objects in bivalent unergatives is basically reduced to their DP vs.
PP-like—i.e., dual DP/PP—categorical status (Odria 2017). While DOM objects
exhibit a DP category, the datives in bivalent unergatives behave more akin to
the rest of PP-like goals. This is evidenced by the possibility or not to license
depictive secondary predication.
19 For non-agreeing datives in northeastern Basque, see Ortiz de Urbina (1995), Albizu (2001,
2009), Fernández (1997, 2010, 2014), Fernández & Landa (2009), Fernández et al. (2009),
Fernández & Ortiz de Urbina (2010), Etxepare & Oyharçabal (2009ab, 2013), Etxepare
(2014) and Ormazabal & Romero (2017).
20 Fernández, et al. (2009: 214–215) report that non-agreeing datives in causative construc-
tions should be analyzed more carefully, given that in this case the lack of dative mark-
ers could be influenced by external factors like indefiniteness, which is known to foster
dative agreement drop (Ortiz de Urbina 1995). Besides, these authors note that, in their
corpus—which is based on classical writings from the 19th and 20th centuries—some
of the non-agreeing datives in causative constructions represent the goal of the embed-
ded predicate, instead of the causee itself. Likewise, they add that the availability to have
non-agreeing datives in causatives could be affected by idiolectal variation as well, as non-
agreeing causees are found in the writings of some but not all north-eastern classical
authors that make use of agreement drop when it comes to goals. In addition, as noted
by Javier Ormazabal (p.c.), in northeastern Basque, dative agreement drop with causees
appeared chronologically later than with goals.
21 See Williams (1980) and Rothstein (1983) for the same pattern in English, McFadden
(2004) for German and Koizumi (1994) for Japanese.
differential object marking in basque and spanish dialects 263
adapted from Zabala (1993: 255, 258)—, where neither the dative (30) nor the
PP (31) is able to control the depictive secondary predicate.
So far, the restriction in question has been mainly tested with goal datives.
Notwithstanding, a closer look at the different nature of dative arguments
reveals that not all kind of datives are equally reluctant when it comes to
licensing depictive secondary predication. In fact, apart from the well-known
exception of the causee, possessor and experiencer datives are also able to
license this kind of predication. This indicates that, instead of the syntactic
configuration—i.e., subject / direct object vs. indirect object—or case mark-
ing—i.e., ergative / absolutive vs. dative—, the restriction on secondary predi-
cation depends basically on the syntactic category of the controller, as DP—i.e.,
causee, experiencer and possessor—but not PP-like—i.e., goal—datives are
able to do so.
Let us first focus on the behavior of the causee. As illustrated in (32) (Zab.ala
1993: 269) and (33) (Demonte 1987: 154), causees are able to control depictive
secondary predication.22
23 Demonte & Masullo (1999: 2466–2467) explain that the licensing of depictive secondary
predication by possessor datives is quite restricted in Spanish. On the one hand, they say
that the depictives allowed in these contexts tend to be replaceable with gerundive verbal
periphrases like estando dormida ‘being slept’ for dormida ‘slept’, or estando anestesiada
‘being anesthetized’ for anestesiada ‘anesthetized’. On the other hand, they note that in
these constructions the possessor tends to be placed in a preverbal position and the depic-
tive in a post-verbal one. Taking these and other facts into account, these authors conclude
that in this case the secondary predicate appears in a position distinct to the predicates
controlled by subjects and direct objects.
differential object marking in basque and spanish dialects 265
zidaten.
aux[3sg.a-1sg.d-3pl.e]
‘In the summer camp, they cut me my hair sleeping.’
Besides, the sentences in (36) illustrate that in Basque experiencers can license
depictives too, a pattern that is again replicated in Spanish subject-like datives
(37) (Fernández-Soriano 1999: 124).24
24 The example in (37a) was gathered by Fernández-Soriano (1999: 124) in La Dorotea by Lope
de Vega.
266 odria
Even though in Basque DOM exhibits a great deal of dialectal as well as idi-
olectal variation, the availability to control depictive secondary predication is
a persistent pattern, given that the results obtained in Larrabetzu are general-
ized among the rest of the consulted speakers from other Basque varieties, such
as Araitz-Betelu (Central Basque) (Arraztio 2010), Lekeitio and Dima (West-
ern Basque) (Fernández & Rezac 2010, 2016), Elgoibar (Transitional Central-
Western Basque), Zumaia (Central Basque), Hondarribia (Transitional Central-
Navarrese Basque), Oñati (Western Basque), Tolosa (Central Basque), Errente-
ria (Central Basque) and Itsasondo (Central Basque) (Odria 2014, 2017).
268 odria
In order to justify the possibility for DOM objects to license depictive secondary
predication, Demonte (1987) claims that, as in causee datives, a is not a true
preposition in DOM objects. Instead, this author takes a in examples like (40b)
to be a dummy preposition, a preposition that does not project a PP maxi-
mal projection and does not therefore impede c-command relations. This way,
Demonte states that the DOM object a María in (40b) can control the depic-
tive borracha ‘drunk’ because there is no P head that blocks the structural c-
command relation between the argument and the predicate.
Interestingly, the possibility for DOM objects to license depictive secondary
predication persists even when the object is coded—or clitic doubled—by the
dative clitic le(s) in the leísta Basque Spanish. Consider, for instance, the exam-
ples in (41), which are the correspondent Spanish examples of those in (39).25
25 In order to avoid interferences, in the Spanish examples in (41) the object bears the same
gender as well as number specification as the subject. Likewise, so as to test the behavior
of the dative clitic le(s), I have changed the objects in (41a) and (41b) to third person.
differential object marking in basque and spanish dialects 269
Overall, assuming that depictives are only incompatible with PP s and PP-like
datives, I conclude that both Basque and (Basque) Spanish DOM objects are
DP s categorically.
26 Even being a bivalent unergative predicate, entzun ‘hear, listen to’ is exceptional allowing
secondary predication (Fernández & Rezac 2010: 134, Odria 2017). This is illustrated in (i):
(i) Ni-ki Mikel-ij mozkortutai/j entzun nion.
I-e Mikel-d drunk hear/listen to aux[3sg.d-1sg.e]
‘I heard to Mikel drunk.’
I believe that the distinction between the two different meanings of entzun ‘hear, listen
to’ might have an influence in this fact. Actually, when entzun means ‘listen to’, dative
marking is generally attested in southwestern dialects, while the absolutive is preferred
in northeastern ones (Fernández & Ortiz de Urbina 2010: 129, 134). On the contrary, when
entzun means ‘hear’, the object can also be absolutive in southwestern varieties, behaving
more akin to a canonical direct object. In fact, Ortiz de Urbina & Fernández (2016: 79–
80) mention that in a sentence like I heard the bossi angryi, the object is usually marked
absolutive, and that dative marking makes the sentence slightly worse. Hence, if entzun
in (i) means ‘hear’ instead of ‘listen to’, its dative object could be closer to a DOM object,
and in that case, it would pattern similar to other perception verbs like ikusi ‘see’, which
would explain its apparently exceptional behavior in (i).
In addition, it is important to note that the dative object of lagundu ‘help, accompany’
can also control secondary predication when pragmatics forces to do so. This is the case
in (ii), where along with the ergative subject, many speakers allow the depictive to modify
the dative object.
(ii)(Haiek)i (ni-ri)j etxe-ra mozkortutai/j lagundu zidaten
they.e I-d house-all drunk help/accompany aux[1sg.d-3plE]
‘They helped/accompanied me home drunk.’
Crucially, some of the consultants add that in (ii) the possibility for the dative to control
secondary predication is logically or pragmatically conditioned. Hence, I believe that the
example in (ii) does not necessarily bring into question the fact that (under normal cir-
cumstances) datives in bivalent unergatives are unable to control secondary predication.
270 odria
Spanish behaves once again like Basque, as it does not allow depictives to depict
of the dative object in verbs like mirar ‘look at’ (43a) or llamar ‘call’ (43b).27
All in all, the impossibility to license depictive secondary predication links the
dative object of bivalent unergatives with PP-like goals of ditransitive construc-
tions, and thus implies that these objects have a dual DP/PP categorical status.
5 Conclusions
In this paper I have compared the syntactic behavior of Basque and (Basque)
Spanish DOM. I have highlighted the main similarities and differences attested
between the two of them and I have distinguished them from the dative object
in bivalent unergative predicates. Contrary to what happens in DOM, in biva-
lent unergatives the dative marking is independent from factors like animacy
and specificity and the object patterns more akin to indirect objects of ditransi-
tive predicates. Moreover, based on the licensing of depictive secondary pred-
ication, I have argued that these datives are additionally distinguished by their
syntactic category. Contrary to DOM objects—which exhibit a DP categorical
status—, datives in bivalent unergatives are generally incompatible with depic-
tives, a pattern that groups them with the rest of PP-like goals.
27 In (43b), borracho ‘drunk’ can also be interpreted as a primary predicate referred to the
dative Pedro. However, when behaving as a secondary predicate, borracho ‘drunk’ can only
be controlled by the subject Juan.
differential object marking in basque and spanish dialects 271
Acknowledgements
This work has been supported by the Basque Government (the research group
IT665–13) and the Spanish Ministry of Economy and Competitiveness
(FFI2014–51878-P). The research leading to these results has also received fund-
ing from the European Union’s Seventh Framework Program for research, tech-
nological development and demonstration under grant agreement no. 613465.
I am grateful to Ane Berro and Javier Ormazabal for their comments and sug-
gestions. Needless to say, all errors are mine.
References
Aissen, Judith. 2003. “Differential object marking: iconicity vs. economy”. Natural Lan-
guage and Linguistic Theory 21: 435–483.
Albizu, Pablo. 2001. “Datibo sintagmaren izaera sintaktikoaren inguruan: eztabaida-
rako oinarrizko zenbait datu”. In Beatriz Fernández & Pablo Albizu (eds.), Kasu eta
komunztaduraren gainean. On Case and Agreement. Bilbo: UPV/EHU Editorial Ser-
vices. 49–69.
Albizu, Pablo. 2009. “Construcciones inacusativas con dativos posesivos y dativos de
interés en vasco: un análisis derivacionalista”. Seminario de Lingüística Teórica, CSIC,
Madrid, May 11.
Arraztio, Kontxi. 2010. “Datibo osagarri bitxiak Araitz-Beteluko hizkeran”. In Beatriz
Fernández, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: aldako-
rtasun sintaktikoa aztergai. [Supplements of ASJU LII]. Bilbo: UPV/EHU. 113–149.
Austin, Jennifer. 2006. “Dative overmarking in Basque”. Euskalingua 9: 136–145.
Austin, Jennifer. 2015. “Transfer and contact-induced variation in child Basque”. Fron-
tiers in Psychology 5: 1–11.
272 odria
Bárany, András. 2018. DOM and dative case. Glossa: a journal of General Linguistics 3(1):
97, 1–40.
Berro, Ane & Beatriz Fernández. 2018. “Applicatives without verbs”. Natural Language
and Linguistic Theory (NLLT).
Bhatt, Rajesh & Elena Anagnostopoulou. 1996. “Object Shift and Specificity: Evidence
from ko-phrases in Hindi”. In L. Dobrin et al. (eds.), Proceedings of CLS 32. Chicago:
Chicago Linguistics Society.
Bossong, Georg. 1985. Empirische Universalienforschung. Differentielle Objekt Markie-
rung in der neuiranischen Sphrachen. Tübingen: Narr.
Bossong, Georg. 1991. “Differential Object Marking in Romance and beyond”. In Dieter
Wanner & Douglas A. Kibee (eds.), New Analyses in Romance Linguistics. Amster-
dam/Philadelphia: John Benjamins. 143–170.
Brugè, Laura & Gerhard Brugger. 1996. “On the Accusative a in Spanish”. Probus 8(1).
1–51.
Cuervo, M. Cristina. 2003. Datives at Large. Ph.D. thesis, MIT.
Demonte, Violeta. 1987. “C-command, Preposition and Predication”. Linguistic Inquiry
18: 147–157.
Demonte, Violeta. 1988. “Remarks on secondary predicates. C-command, extraction
and reanalysis”. The Linguistic Review 6–1: 1–39.
Demonte, Violeta. 1995. “Dative Alternation in Spanish”. Probus 7. 5–30.
Demonte, Violeta & Pascual J. Masullo. 1999. “La predicación. Los complementos pred-
icativos”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva de la
lengua española. Madrid: Espasa Calpe, RAE. 2461–2524.
Etxepare, Ricardo. 2003. “Valency and Argument Structure in the Verb”. In José Igna-
cio Hualde & Jon Ortiz de Urbina (eds.), A grammar of Basque. Berlin: Mouton de
Gruyter. 369–465.
Etxepare, Ricardo. 2014. “Contact and Change in a restrictive theory of parameters”.
In Carme Picallo (ed.), Linguistic Variation in the Minimalist Framework. Oxford:
Oxford University Press. 108–139.
Etxepare, Ricardo & Bernard Oyharçabal. 2009a. “Bi datibo egitura ipar-ekialdeko zen-
bait hizkeratan”. Lapurdum XIII: 145–158.
Etxepare, Ricardo & Bernard Oyharçabal. 2009b. “Datiboa ipar-ekialdeko zenbait hizk-
eratan”. Ms., IKER, April 25.
Etxepare, Ricardo & Bernard Oyharçabal. 2013. “Datives and adpositions in northeast-
ern Basque”. In Beatriz Fernández & Ricardo Etxepare (eds.), Variation in datives: a
micro-comparative perspective. Oxford: Oxford University Press. 50–95.
Fábregas, Antonio. 2013. “Differential Object Marking in Spanish”. Borealis. An Interna-
tional Journal of Spanish Linguistics 2–2: 1–80.
Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. Bilbo: UPV/EHU Edi-
torial Services.
differential object marking in basque and spanish dialects 273
Fernández, Beatriz. 2010. “Goi datiboak eta are goragokoak euskaraz: jabe datiboak eta
datibo hunkituak vs. datibo etikoak”. Gogoa, 10: 1–2: 1–20.
Fernández, Beatriz. 2014. “Goi-eta behe-datiboak eta Datibo-komunztaduraren murriz-
tapena: ipar-ekialdeko hizkerak eta euskara estandarra”. Irantzu Epelde (ed.), Euskal
dialektologia: lehena eta oraina [Supplements of ASJU LXIX]. UPV/EHU. 37–63.
Fernández, Beatriz & Josu Landa. 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.
Fernández, Beatriz & Jon Ortiz de Urbina. 2010. Datiboa hiztegian. Bilbo: UPV/EHU Edi-
torial Services.
Fernández, B. & J. Ortiz de Urbina, J. 2012. “Dative (First) Complements in Basque”. In
Ernestina Carrilho & Beatriz Fernández (eds.), Syntactic microvariation in Westmost
European Languages. Journal of Portuguese Linguistics (Special Issue), 11-1. Edições
Colibrí—Unviersidade de Lisboa. 83–98.
Fernández, Beatriz, Jon Ortiz de Urbina & Josu Landa. 2009. “Komunztadurarik gabeko
datiboen gakoez”. In Ricardo Etxepare, Ricardo Gómez & Joseba A. Lakarra (eds.),
Beñat Oihartzabali gorazarre. ASJU XVIII: 352–380.
Fernández, Beatriz & Milan Rezac. 2010. “Datibo osagarri bitxiak eta Datiboaren Lek-
ualdatzea: ari nai diyot eta kanta egin nazu bidegurutzean”. In Beatriz Fernández,
Pablo Albizu, & Ricardo Etxepare (eds.), Euskara eta euskarak: aldakortasun sintak-
tikoa aztergai. [Supplements of ASJU LII]. Bilbo: UPV/EHU. 113–149.
Fernández, Beatriz & Milan Rezac. 2016. “Differential Object Marking in Basque vari-
eties”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
grammar of Basque. Amsterdam/Philadelphia: John Benjamins. 93–139.
Fernández, Beatriz & Ibon Sarasola. 2010. “Marinelei abisua: izen ondoko datibo sintag-
mak izenburuen sintaxian”. Lapurdum XIV: 55–75.
Fernández-Ordóñez, Inés. 1994. “Isoglosas internas del castellano. El sistema referen-
cial del pronombre átono de tercera persona”. Revista de Filología Española 74-1/2:
71–125.
Fernández-Ordoñez, Inés. 1999. “Leísmo, laísmo y loísmo”. In Ignacio Bosque & Vio-
leta Demonte (eds.), Gramática descriptiva de la lengua española I. Madrid: Espasa
Calpe. 1317–1394.
Fernández-Soriano, Olga. 1999. “Two types of impersonal sentences in Spanish: locative
and dative subjects”. Syntax 2–2: 101–140.
Franco, Jon A. 1993. “Conditions on clitic doubling: the agreement hypothesis”. ASJU 27:
285–298.
Hernanz, Mª Lluisa. 1988. “En torno a la sintaxis y la semántica de los complementos
predicativos en español”. Estudis de Sintaxis, Estudi General, 8 (Colegi universitari
deGirona) 7–29.
Hualde, José Ignacio, Gorka Elordieta & Arantzazu Elordieta. 1994. The Basque dialect of
274 odria
Lequeitio [Supplements of ASJU XXXIV]. Bilbo & Donostia: UPV/EHU & Gipuzkoako
Foru Aldundia.
Ibarra, Orreaga. 1995. Ultzamako hizkera. Inguruko euskalkiekiko harremanak. Iruñea:
Nafarroako Gobernua.
Koizumi, Masatoshi. 1994. “Secondary predicates”. Journal of East Asian Linguistics 3:
25–79.
Landa, Alazne. 1995. Conditions on Null Objects in Basque Spanish and their relation to
Leísmo and Clitic Doubling. Ph.D. thesis, University of Southern California.
Leonetti, Manuel. 2004. “Specificity and differential object marking in Spanish”. Cata-
lan Journal of Linguistics 3: 75–114.
Mahajan, Anoop Kumar. 1990. The A/A-bar distinction and movement theory. Ph.D.
thesis, MIT.
Masullo, Pascual J. 1992. Incorporation and case theory in Spanish. A crosslinguistic per-
spective. Ph.D. thesis, University of Washington.
McFadden, Thomas. 2004. The position of morphological case in the derivation: a study
on the syntax-morphology interface. Ph.D. thesis, University of Pennsylvania.
Mounole, Céline. 2012. “Evolution of the transitive verbs in Basque and apparition of
datively marked patients”. In Gilles Authier & Katharina Haude (eds.), Ergativity,
Transitivity, and Voice. Berlin/New York: Mouton de Gruyter. 355–379.
Odria, A. 2012. What lies behind differential Object Marking: a Basque dialect survey.
Master’s thesis, UPV/EHU. In Pello Salaburu, Patxi Goenaga & Ibon Sarasola (eds.),
Sareko Euskal Gramatika.
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.
Odria, Ane. 2015. “Euskarazko bigarren mailako predikazioaren murriztapen sintak-
tikoez”. In Beatriz Fernández & Pello Salaburu (ed.), Ibon Sarasola gorazarre. Hom-
enatge. Homenaje. Bilbo: UPV/EHU. 449–515.
Odria, Ane. 2017. Differential Object Marking and datives in Basque syntax. Ph.D. thesis,
UPV/EHU.
Odria, Ane. to appear. “DOM and datives in Basque: not as homogeneous as they might
seem”. In A. Pineda & M.A. Irimia (eds.), DOM and datives -a homogeneous class?
Linguisticae Investigationes. Amsterdam/Philadelphia: John Benjamins.
Ormazabal, Javier & Juan Romero. 2013a. “Differential Object Marking, Case and Agree-
ment” Borealis. An International Journal of Hispanic Linguistics 2–2: 221–239.
Ormazabal, Javier & Juan Romero. 2013b. “Non Accusative Objects”. Catalan Journal of
Linguistics 12. 155–173.
Ormazabal, Javier & Juan Romero. 2017. “Historical Changes in Basque Dative Alterna-
tions: evidence for a P-based (neo)derivational analysis”. Ms., UPV/EHU & University
of Extremadura.
differential object marking in basque and spanish dialects 275
Ortiz de Urbina, Jon 1995. “Datibo komunztaduraren gainean”. In Ricardo Gómez &
Joseba A. Lakarra (eds.), Euskal Dialektologiako Kongresua (Donostia, 1991ko Irailak
2–6) [Supplements of ASJU XXVIII]. Donostia: Gipuzkoako Foru Aldundia. 579–588.
Ortiz de Urbina, Jon & Beatriz Fernández. 2016. “Datives in Basque bivalent unerga-
tives”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
grammar of Basque. Amsterdam/Philadelphia: John Benjamins. 67–93.
Oyharçabal, Bernard. 2010. “Basque ditransitives”. In Maia Duguine, Susana Huidobro &
Nerea Madariaga (eds.), Argument structure and syntactic relations: A cross-linguistic
perspective. Amsterdam/Philadelphia: John Benjamins. 233–260.
Pensado, Carmen. 1995. “El complemento directo preposicional: estado de la cuestión
y bibliografía comentada”. In Carmen Pensado (ed.), El complemento directo preposi-
cional. Madrid: Visor. 11–59.
Pineda, Anna. 2016. Les fronteres de la (in)transitivitat. Estudi dels aplicatius en llengües
romàniques i basc. Barcelona: Institut d’Estudis Món Juïc (Col·lecció Cum Laude, 6).
Pylkkänen, Liina. 2008. Introducing arguments. Cambridge/London: MIT Press.
de Rijk, Rudolf P. 2008. Standard Basque. A Progressive Grammar. Cambridge/London:
The MIT Press.
Rodríguez-Mondoñedo, Miguel. 2007. The Syntax of Objects: Agree and Differential
Object Marking. Ph.D. thesis, University of Connecticut.
Rodríguez-Ordóñez, Itsaso. 2013. “Contact-induced phenomena in Gernika Basque: the
case of dative over-marking”. Selected Proceedings of the 15th Hispanic Linguistics
Symposium, Athens, Georgia. 236–251.
Rodríguez-Ordóñez, Itsaso. 2016. DOM in Basque: grammaticalization, attitudes and ide-
ological representations. Ph.D. thesis, University of Illinois at Urbana-Champaign.
Romero, Juan. 1997. Construcciones de doble objeto y gramática universal. Ph.D. thesis,
Universidad Autónoma de Madrid.
Rothstein, Susan D. 1983. The syntactic forms of predication. Ph.D. thesis, MIT.
Torrego, Esther. 1998. The dependencies of objects. Cambridge, Mass.: MIT Press.
Torrego, Esther. 2010. “Variability in the case pattern of causative formation in Romance
and its implications”. Linguistic Inquiry 41: 445–470.
Urrutia-Cárdenas, H. 1995. “Morphosyntactic features in the Spanish of the Basque
Country”. In Carmen Silva-Corvalán (ed.), Spanish in four continents. Washington,
D.C.: Georgetown University Press. 243–259.
Urrutia-Cárdenas, Hernán. 2003. “Los clíticos de tercera persona en el español en el
País Vasco”. Revista de Filología y su Didáctica 26: 517–538.
Williams, Edwin S. 1980. “Predication”. Linguistic Inquiry 11: 203–237.
Zabala, Igone. 1993. Predikazioaren teoriak Gramatika Sortzailean (Euskararen kasua).
Ph.D. thesis, UPV/EHU.
Zabala, Igone. 2003. Nominal Predication: copulative sentences and secondary predi-
cation. In José Ignacio Hualde & Jon Ortiz de Urbina (eds.), A Grammar of Basque.
Berlin: Mouton de Gruyter. 426–446.
chapter 9
1 Introduction
In principle, this initial description sets the Basque data apart from the
Romance ones, if causatives in Spanish and French are of the analytic type
(Comrie 1989; Dixon 2000). Although usually bound and without the ability
to occur freely, however, the causative ‘morpheme’ in Basque is fairly transpar-
ently a verb. This brings the Basque causative construction closer to that with
causative verbs like Spanish hacer or French faire ‘make’. The type of complex
verb formed in all these languages and complementation types are addressed
first in Section 2. Section 3 will be devoted to causee marking and agreement
conflicts resulting from the presence of two datives in the same structure. As
we will see, this is also an area where new patterns converging with Romance
ones seem to be emerging or developing side by side with traditional ones.
1 The following abbreviations have been used in glosses: abs ‘absolutive’, aux ‘auxiliary’, comp
‘complementizer’, dat ‘dative’, det ‘determiner’, comp ‘complementizer’, erg ‘ergative’, fem
‘feminine’, fut ‘future’, gen ‘genitive’, inst ‘instrumental’, masc ‘masculine’, nom ‘nominal-
izing affix’, pl ‘plural’, q ‘question morpheme’, subj ‘subjunctive’.
As indicated above, Basque causative verbs are complex words containing two
verbal forms. There are reasons to treat this combination not as a result of
morphological derivation, but as a syntactic one. Section 2.1 provides some jus-
tification for this analysis, which, in turn, allows for a closer comparison with
Romance in general and Spanish in particular. The complex verb unit in the lan-
guages under discussion is examined in section 2.2, and section 2.3 focuses on
direct and indirect causation in these structures. Impersonal causatives and/or
faire par constructions are discussed in section 2.4.
2 Example from a mid 20th century letter to Basque linguist Azkue, in Roncalese, included in
the OEH under arazi.
278 ortiz de urbina
The latter is the approach taken already in DHM (1989), following Baker
(1988) and also in the spirit of Zubizarreta’s (1985) Complex Verb Hypothesis,
although the choice is made there for theoretical reasons. Some evidence for
this syntactic approach can be derived from the fact that VV compounding is
not found in Basque morphology outside of these cases. Moreover, there is a
correlation between the shape the event verb takes in causatives in any given
dialect and the availability of the root in the syntax of that dialect. That is, while
all dialects use the verbal root in morphological derivation, it is only found as a
free form in the syntax of eastern dialects. Subjunctive and potential forms are
then combined with the root in these dialects (har ‘take’ in 5a), but with the
participial form (which happens to be the neutral, citation form of the verb in
Basque) in western dialects (hartu in 5b):
Similarly, in western dialects that use eragin as the causative verb, this is com-
bined not with the verbal root, as in the standard language, but with the per-
fective participle: we then find hartueragin rather than harrarazi ‘make take’,
or ikusieragin rather than ikusarazi ‘make see’. The use of the ‘syntactic’ verbal
base in causative formation, rather than the verbal root, reserved for morphol-
ogy in western dialects, strongly suggests causatives involve a syntactic combi-
natory process rather than a morphological one.3
An important difference between causative verbs in Basque and Romance
hacer/faire is that the former do not take tensed clausal complements, while
the latter do (6b), generally conveying a more indirect type of causation than
with tenseless ones:
3 According to the OEH, the easternmost use of eragin ‘cause’ can be found in the Labourdin
writer Urte, where, as expected, it combines with the verbal root, productive in the syntax of
that dialect. Central dialects also use this causative verb occasionally, and just like with their
main causative verb arazi, it seems to combine with either the verbal root of the participial
form depending on the syntactic productivity of the former.
280 ortiz de urbina
Similarly, in Sarasola’s (1985) dictionary, many entries for causative verbs with
erazi are provided with definitions which involve the verb egin ‘do’ taking a
subjunctive embedded clause, as ito erazi ‘cause to drown’, glossed as “make
that something drown”, higi erazi ‘cause to move’, given as “make that some-
thing move” or iraun erazi ‘cause to last’ as “make that something last”. The
glosses are as marked in Basque as their translations are in English, and look
more like a metalinguistic device than actual examples of productive lan-
guage.4
Distributionally equivalent to subjunctive tensed clauses in many respects,
tenseless nominalized complement clauses are also very occasionally found as
complements of a causative verb, instead of the general root or participial form.
As before, this possibility does not exist with arazi itself but rather with other
causative verbs, arazo ‘cause’ in (8a), eragin ‘cause’ in (8b) and eginarazi ‘cause
to do’ in (8c):
4 This use is not registered in the very comprehensive General Basque Dictionary (OEH). B. Fer-
nández (p.c.) points out, however, that it is found in the speech of young speakers of Ondarroa
Basque.
complex causative verbs and causees in basque (and romance) 281
Leaving these isolated cases aside, this paper will exclusively focus on the gen-
eral pattern, where the causative verb combines with a verb in its root or par-
ticipial form, as in (1b).
The connection between the two verbs in Basque, however, is more com-
plete than what one finds in Romance. Thus, the type of complement taken by
the causative verb makes it impossible for the complement verb to be negated
(10b)5 or have aspectual modification or passive morphology, unlike Spanish
(10c,d,e):
While perfective in (c) is very marginal (probably due to the temporal sequenc-
ing inherent in situations where someone is forced to do something), passive is
5 The presence of negation has the familiar effect of preventing ‘clause union’ effects, so that
clitic climbing is not possible:
(i) El alcohol te hizo entenderlo. El alcohol te hizo no entenderlo.
the alcohol you made understand.it the alcohol you made not understand it
(ii) El alcohol te lo hizo entender. *El alcohol te lo hizo no entender
the alcohol you it made understand the alcohol you it made not understand
As for Basque, negation blocks clause union in participial complements of nahi ‘want’ and
behar ‘need’: compare ikusi nahi ‘want to see’, with clause union effects and ez ikustea nahi
‘want not to see’, with the nominalized verbal form ikustea instead.
6 Zubizarreta (1985) stars causatives with embedded passives as in this example. As also appar-
ent from the discussion of matrix passives, there are important dialectal differences at work
here.
complex causative verbs and causees in basque (and romance) 283
only slightly marked (e), while progressive seems perfectly acceptable (d). No
Basque counterparts can be formed for (10).
Some differences between Spanish and French causatives on one hand and
Italian on the other are brought forth in Zubizarreta (1985), who accounts
for them claiming Italian fare forms are monoclausal structures while Span-
ish and French causatives are associated “simultaneously with two structures:
one biclausal, the other monoclausal” (1985: 280). In essence, the difference
would imply a more close knit verbal union in the case of Italian as opposed
to French/Spanish. The difference can be observed in the availability of pas-
sivization of the causative verb itself. Zubizarreta’s examples below contrast
the grammatical Italian (11a) and ungrammatical Spanish (11b):
The contrast is not found in Peninsular Spanish, where sentences like (b) are
acceptable. In the analysis presented by Zubizarreta (1985: 274), the causative
verb is a morphophonological word but acts as a bound morpheme from a
morphosyntactic perspective, which brings apparently syntactic causatives
closer to morphosyntactic ones, just as the facts in 1.1 bring the apparently
morphological causatives closer to the syntactic ones. In any event, we would
expect Basque to behave like Italian if, from a descriptive perspective, a mon-
oclausal analysis with a morpheme-like causative verb fare is assumed to
account for the Italian facts. This cannot be checked directly with passives,
very marginal in Basque, but parallel facts can be found in ‘long impersonal’
constructions, similar to Spanish ‘reflexive passives’ or English middles, where
the object of a verb becomes the subject of an intransitivized version of that
verb. As already pointed out in Aissen and Perlmutter (1986), this process can
sometimes apply across clauses, as a result of their proposed clause union or
reduction process. The examples below illustrate this monoclausal-like effect
in both Basque and Spanish using long mediopassives based on impersonal
(causee-less) causative structures, to which we will return below in section
1.4:
284 ortiz de urbina
The logical objects of read (‘those books’) and of polish (‘diamonds’) are now
subjects of the causative (or of the complex verb), as shown by the fact that
they agree with the latter, even though, in principle, they are not its argu-
ments.
Regardless of the actual constraints that play a role in the more or less deviant
status of some of the previous sentences, what is clear is that they disappear
when tensed complement clauses are used, so that (16), for instance, is accept-
able:
The Basque causative with arazi ‘cause’ also expresses a fairly direct type of cau-
sation, less felicitous where pragmatic considerations do not make this type of
causation possible. In this, then, it clearly patterns with Romance cause+infini-
tive structures, rather than with tensed embedded patterns. This is shown in
(17) with stative complements of the causative verb, which Oyharçabal (2003)
identifies as excluded from causative structures. The stative predicate in the fol-
lowing examples is paired with clausal subjects for the causative verb, forcing
a non-agentive interpretation:
The corresponding Spanish (b) examples sound less deviant than the Basque
(a) examples, while the (c) sentences with tensed complements are grammat-
ical in Spanish. Similarly, weather expressions are formed in Basque with exis-
tential constructions which cannot be causativized; alternatives resort to egin
‘do’ for the root verb:
7 A verb like jakinarazi ‘cause to know’ is acceptable in the meaning ‘communicate’, shared
with Spanish, but not in the stative reading in (18).
complex causative verbs and causees in basque (and romance) 287
8 In the case of hotz ‘cold’, egin ‘do’ itself is commonly used instead of izan ‘be’, so the
causative eginarazi is not totally unexpected. It is in the case of euria, though. Whether
because it is a recent use of egin or because it is not a light verb here, unlike in expressions
like negar egin ‘cry, lit. make cry’, the fact is that this egin is never causativized as eragin.
9 There is no tenseless form of the transitive auxiliary; the intransitive auxiliary izan ‘be’ is
used instead. For texts up to the 20th century, the OEH gives three examples, two from
Bible translations and a third one in a modern philosophical book in the context of gods
‘making be’, i.e., creating, without causee. The Ereduzko Prosa Gaur online corpus gives 13
instances from year 2000 to the present, all of them in translations or, two of them, in a
philosophical work. None in the written press.
10 At best, the sentence may be interpreted as the causative of the verb used in the substan-
dard expression logure(a) dut ‘I have sleep’, with a transitive ‘have’ probably patterned after
Spanish tener sueño ‘be sleepy (lit. ‘have sleep’)’ rather than the standard Basque intransi-
tive ‘be’ of logure izan ‘be sleepy’ (lit. ‘be sleep-wanting’).
288 ortiz de urbina
The incremental scale in the comparatives in (22) and (23) has the effect of
making the Basque causatives of izan ‘be’ more acceptable.
As for ‘inversion’ predicates with experiencer datives as in (13c), they are
often states and, in any event, one need not expect the theme subject of such
predicates to act as causee, but rather the experiencer. The theme causee
(example (24a,b), in the intended interpretation) with gustar is as unaccept-
able in Basque as in Spanish, and the experiencer causee is slightly better
(24c,d). Again, the que clause is perfect (25):
arazi ‘cause’ causatives can be directly connected with Spanish and perhaps
Romance make+infinitive causatives.
Examples like (b) are quite marked in Spanish, and are found at best in very
formal contexts.12 Notice that this so-called passive causative does not occur
with passive morphology on the embedded verb; the latter, as indicated above,
is possible with an animate causee:
In contrast, the examples in (26) can have a by-phrase which does not co-occur
with passive morphology. The presence of a passive agent adjunct like par son
11 English does not admit ‘subjectless’ complements of make, so (i) is ungrammatical. This
is why a translation with have is given:
(i) *The City Hall made repair the house
Passive complements of make are of course possible, with overt passive morphology and
intervening NP; in those cases, however, make usually requires a to-infinitive in standard
dialects:
(ii) The City Hall made the house to be repaired
12 In this they resemble ‘reflexive passive’ impersonals with overt agentive phrases like Se
firmó la paz por los embajadores.
290 ortiz de urbina
enfant ‘by his son’ is only one aspect of the characterization of these struc-
tures, and in fact these structures stand out as a separate type even if there
is no overt agent expressed. The reason is that, crucially, the causee forced to
act by the causer is not expressed as a core argument of the faire+verb unit,
so these constructions are impersonal, causee-less in that sense (not necessar-
ily agentless), as discussed in Zubizarreta (1985). Thus, the following example
would also qualify as a ‘passive’ causative of this type:
13 Ortiz de Urbina (2003: 604) brings forth the following example as a possible candidate for
this construction:
(i) Mutilez zain-arazten zitian bere arthalde handiak
boys.inst care-cause aux his flock large
‘He had his large flocks looked after by boys’ (Eliss. P.A.: 80)
The instrumental is used for passive agents in the dialect in question; it is possible, how-
ever, to regard this as a normal impersonal causative with an instrumental reading of
mutilez.
14 Oyharçabal’s example (b) is ambiguous between the intended impersonal interpretation
and an irrelevant interpretation where causative arazi is added to an inchoative unac-
cusative hil ‘die’. The causative version of an inchoative is possible, though less common
than the lexical causative pattern (meaning kill). See Berro (2015).
292 ortiz de urbina
Other pragmatic factors may be involved, since causes are acceptable in other
sentences. In the following sentence the animate noun phrase could be inter-
preted as an agent or perhaps also as a cause, without apparent effect on the
grammaticality of the sentence; the inanimate subject would be a cause:
15 The unaccusative verbs in (32) have a mixture of causes and agents; changing the former
to the latter does not have any effect on the acceptability of these sentences.
16 With respect to direct and indirect causation, Oyharçabal (2003) points out, quoting Dan-
los (2001), that “causality in the real world resembles a chain at the end of which it is always
possible to attach a further link”.
294 ortiz de urbina
17 Notice that in all cases the embedded verb is in an active form, unlike in some of the
translations; the causee is given as a by phrase as a way to make the intended meaning
understood.
complex causative verbs and causees in basque (and romance) 297
The last three examples contain unaccusative verbs, and although judgements
are not so clear for all examples/speakers, their inanimate causees, which
would appear as direct objects in the personal causatives, seem to be able
to become subjects of the mediopassive construction. In any event, leaving
the status of Spanish causativized unaccusatives as faire-par or impersonal
causatives aside, the contrast obtained by the presence of the causee in (a,b)
is clear and may indicate a different status/position for causees in Basque
and Spanish causatives. Summarizing, in spite of the apparent differences in
causativization strategies, if the data described here are correct, Basque and
Spanish (as well as Romance) causatives with tenseless complements share
important fine-grain features, as well as some differences. Accounting for them
goes beyond the limits of the modest descriptive goal of this paper, and the
complexity of acceptability patterns makes the enterprise more difficult. How-
ever, it seems we might be able to gain insights about both languages and about
causative structures in general from the close examination of their grammars.
In the following section we turn to an area where contact may be having a more
distinct impact on the grammar of Basque, namely causee marking, with par-
ticular emphasis on agreement conflicts created in Basque as a result of dative
doubling.
cate. This can be effected in widely different ways depending on one’s theoret-
ical persuasion, but the general layout is fairly clear. As discussed in Comrie
(1989) and Dixon & Aikhenwald (2000), the causer argument of the causative
verb/morpheme/head is aligned with the subject function, and the causee can-
not be a subject unless a fully biclausal structure is formed. If this is not the
case, typologically the most widespread situation is for intransitive event sub-
jects to occur as objects of the complex causative verb and for transitive sub-
jects to be found as indirect objects. The following subsections will focus on
causee case in Basque causatives. In section 3.1 the Standard Basque pattern in
presented, focusing on causee marking and case doubling phenomena. DOM—
(Differential Object Marking) related dialectal or substandard phenomena will
be discussed in section 3.2.
Levin (1983), and since then many alternative analyses have been proposed (see
Laka 1993; Etxepare 2003 or Berro 2015, among many others). Levin also noticed
that some unergative predicates also (or exclusively) exist as simplex verbs, tak-
ing then ergative subjects and transitive auxiliaries: jolastu ‘play’, saltatu ‘jump’,
distiratu ‘glitter’, iraun ‘last’, etc. There is more variation for the causative of
these verbs. De Rijk (2008: 378–380), from where (45b) is taken, states than
in northern dialects they tend to behave like unaccusatives (a), although occa-
sionally now, and more often in the past, dative causees were also found. In
the central dialects (Gipuzkoan mostly, since arazi ‘cause’ gives way to eragin,
arazo ‘cause’ further to the west), unergative causees tend to be marked dative,
behaving then in causatives just like the ergative subjects of transitive verbs
(b):
As also pointed by de Rijk, the normative advice is to use datives with animate
causees and absolutive with inanimate ones, introducing a DOM-like distinc-
tion to which we will return in 3.2. Both unaccusative and unergative verbs
may take yet a second argument in the dative; these will be discussed in the
following section, once dative causees are introduced for regular monotransi-
tive verbs, to which we turn now.
The external argument of a transitive verb appears as a dative causee in the
causative version of that verb, as shown in (46a):
This pattern is general. The embedded verb already includes a direct object and
the logical subject of irakur- ‘read’ cannot appear as a direct object, given that
300 ortiz de urbina
The (b) example illustrates the causative of the basic unergative configuration
of this verb, with an allative complement; as described in the previous section,
the causee appears in the dative. When the verb takes a dative complement, a
dative doubling situation emerges, exemplified here in (c). The auxiliary gloss
has been given in more detail to point out that when two datives vie for agree-
ment, it is the causee that becomes cross-referenced; in this example dative
agreement is with the unexpressed berari ‘to him’, while the goal niri ‘to me’
does not trigger any marker on the auxiliary. If we eliminate the causee by using
complex causative verbs and causees in basque (and romance) 301
an impersonal causative, the original dative is still unable to agree, so that in the
following examples the dative auxiliary is flagged as ungrammatical. If used,
the dative is interpreted as referencing an unexpressed causee:
As indicated, the clitic referring to the indirect object of llamar ‘call’ must
remain attached to the lower verb, and may not climb to the matrix causative
verb, as if ‘clause union’ had not taken place (Aissen & Perlmutter 1983).
cal/interest datives may also be found with most verbal configurations and will
be addressed briefly below.
Beginning with inversion predicates, we already mentioned that the dative
argument can only marginally be taken to be a causee, as in (24c), repeated
here as (51a); attested examples are given in (51b,c):
Dative is expected here for the causee, since there is already an absolutive
element, and this case also fits its experiencer status. In order to check for
dative doubling configurations, we would need an example where the theme is
interpreted as a causee, a situation which, as indicated in 1.3 above, is usually
unacceptable in Spanish with infinitival causatives. The type of interpretation
sought is similar to that in English (52), where the intervening NP would corre-
spond to the causee of the structures in question in Basque and Spanish:
There is no thematic link between make and the following NP in English, while
the causee does have one with the complex verb in both Spanish and Basque;
these sentences are unacceptable:18
18 There are nonetheless occasional attested examples which may correspond to this inter-
pretation, without an auxiliary:
(i) [ikasleei] hizkuntza hurreratu behar zaie, gustarazi
students.dat language draw need aux like.cause
‘language must be brought over to the students, made be likeable’
The interpretation seems to be that something must be done to make the language more
appealing to students, rather than simply making the students like the language, and to
the extent that this is a possible (and different) interpretation, it would perhaps be an
(isolated) example of theme ‘causee’.
complex causative verbs and causees in basque (and romance) 303
The Spanish sentence is interpretable (with some effort) thanks to the two cli-
tics and their placement. The single agreement marker in the Basque examples
makes interpretability more complicated.
The second type of ‘bivalent unaccusative’ is that where a goal dative is
added to an intransitive verb (54a).20 While the directional complement is
more commonly expressed in the allative, the dative exists as a formal register
possibility. Given ergative marking, the absolutive intransitive subject remains
absolutive, now as object of the transitive causative verb. The combination with
the ergative causer and the dative goal would produce a pattern identical to that
of ditransitive predicates (54b); however, the structure is considered deviant by
speakers:
19 Using a dativeless transitive auxiliary like dute does not improve the sentence.
20 See Rezac (2009), who follows Ormazabal & Romero’s (2007) derivational approach for
Double Object Constructions, deriving goal datives from PP s.
304 ortiz de urbina
all causees are marked dative in the standard language, but all datives seem
to be interpreted as causees, and this is not the interpretation in (54b).21 In
(54c) an inanimate causee is intended, with a goal dative; the sentence is only
interpretable as the causative of a monotransitive use of hurbildu, again with
a causee dative. An indication that unacceptability arises from the unintended
connection between dative and causee is that if dative agreement is eliminated,
the sentence is perfectly interpretable and acceptable for many speakers in the
intended sense. The connection between dative and causee is also made evi-
dent by looking at impersonal, causee-less structures. Again, their dative is not
interpreted as goal but as causee, so (55) is unacceptable in the intended inter-
pretation:
The scant use of the basic goal dative configuration in non-formal contexts only
adds to the complexity of factors which may help us understand speaker judge-
ment variability in these cases.
21 This applies to argumental datives. Ethical datives can be found in causative structures
without receiving a causee interpretation. We return to them below.
complex causative verbs and causees in basque (and romance) 305
22 See Fernández (this volume) for a comparison between Spanish and Basque ethical and
other non-selected datives. See also Fernández & Ortiz de Urbina (2010: 179) for Basque
ethical and interest/possessive datives.
23 Interest datives associated with ergatives are felt, right or wrongly, to show the influ-
ence of Spanish on Basque. The association with the absolutive would be expected in a
treatment of ethical datives as low applicative heads, as in Pylkkänen (2008), where the
dative introduces a source, goal, benefactive, or, in this case, interested argument related
to the object. This approach is taken in Oyharçabal (2007) and Albizu (2009), among oth-
ers. In the case of (d), even association with the absolutive object (‘my beans’) may be
excluded in eastern dialects, as ethical datives are excluded from stative contexts (Oyhar-
çabal 2007).
306 ortiz de urbina
priority for agreement with the cause excluding the ethical marking, in ways
similar to what has been noted above for dative causees; we return to similar
cases in the following section:
Although variability is rampant and the situation for each dialectal area must
be checked, it seems that we can arrive at the generalization that the exam-
ples clearly ruled out are those where ethical dative agreement on the verb is
given priority over agreement with a dative causee. This is the case in (58c)
and (58d), in dialects where the unergative subject surfaces as a dative cause.
Those examples which are acceptable for some speakers and dubious for others
(58a,b and 58d with absolutive causee) are those where no such conflict arises;
the variability in the acceptability of (58a,b) would then be unrelated to these
agreement conflict considerations.24
24 Albizu (2009) provides a technical analysis in the line of Pylkkänen (2008) and Rezac
(2009). In western dialects with DOM effects in causatives, all causees in (58) will be
marked dative, as discussed in the following section. See footnote 28 for some comments
on the dative doubling situation created by ethical datives.
complex causative verbs and causees in basque (and romance) 307
with a causee and a goal is possible; if not the result is probably more unac-
ceptable because of the impossibility of disambiguation than ungrammati-
cal:
Similar acceptability issues are found in the dative doubling examples (47c),
(48) and (49) examined above, where the causee dative corresponds to the log-
ical subject of an ergative verb.
308 ortiz de urbina
As discussed with regard to the (lack of) interaction between ethical datives
and causee datives in unaccusative predicates, any other dative preempts the
use of ethical datives. In a ditransitive sentence, in particular, the dative may
not be interpreted as ethical. Thus, in the basic ditransitive (62) the dative may
be interpreted as an ethical dative only if the verb eraman ‘take’ is used as a
monotransitive one:
We can now introduce dative causees into the picture. A sentence like (65) with
a single overt dative and a verb like eraman ‘take’, which can have both mono-
transitive and ditransitive uses, is several ways ambiguous:
25 Speakers can be made to understand the meaning of this clause if the first dative corre-
sponds to the interested participant and the dative closer to the verb corresponds to the
goal.
complex causative verbs and causees in basque (and romance) 309
The first and most prominent interpretation is one where the dative is inter-
preted as a causee, the agent of a monotransitive use of eraman ‘take’. A second
interpretation, more formal, corresponds to an impersonal causative of ditran-
sitive eraman ‘take’ where the dative is interpreted as a goal, i.e., making some
unspecified causee take the son to Mikel. The third interpretation is one where
the dative is interpreted as an ethical dative, and again, this is available only
if there is no causee, i.e., making some unspecified agent take Mikel’s son. If
we try to include two of these datives in the same sentence the result is quite
complex:
In both cases, the agreeing dative is interpreted as causee. The other dative may
be interpreted as goal or ethical, perhaps with different likelihood depending
on ordering alternatives which remain to be studied.
310 ortiz de urbina
(68) a. Bide pendiz bat artu eragin eta goraiño joan eragin eutsan.
path steep one take cause and up go cause aux.3dat
‘[S/he] made him take a steep path and go all the way up.’ (Aug. Zubika-
rai)
26 In one of the earliest attested occurrences of eragin, from the second half of the 16th cen-
tury, this causative actually appears not only with a participial complement verb, but even
in a tensed form, suggesting it is not fused with the embedded verb, in a structure similar
to English make:
(i) Azaiteau galdu lerait / ardura usatzakeak.
nourishment lose cause often use.without.erg
‘Not using it often would make me lose this nourishment.’ (Lazarraga 1196v)
The causee is dative, again possibly as an argument of the causative verb itself.
complex causative verbs and causees in basque (and romance) 311
The (a) and (b) examples above contain animate causees which would have
been marked absolutive in the standard language, as described in 3.1.2.1 above,
marked dative in published texts. The (c) examples present the types of causees
which, being inanimate or non-specific, would usually take absolutive case also
in these dialects.27 This dative interacts and overlaps with other datives in ways
which may but need not be similar to those described above for multiple dative
situations, but, like DOM patterns in general, this is an understudied area in
Basque linguistics.28
27 Fewer examples are given of the standard-like pattern precisely because it is the norma-
tive one, found in more monitored texts in the western and central area. The qualification
‘types of animate causees’ in the text is introduced because examples with inanimate
absolutive causees should belong to the same author and, preferably, text, as those with
animate causees, something which has not been attempted here.
28 For instance, double datives will also emerge with unaccusative absolutive dative config-
urations:
312 ortiz de urbina
b. La / lo vi en la calle.
it.fem / it.masc saw in the street
‘I saw it in the street.’
Masculine Feminine
+Animate -Animate +Animate -Animate
Direct object le lo le la
Indirect object le le le le
The clitic le in (a) can refer both to a feminine and masculine direct object,
while the standard feminine and masculine direct object clitics la/lo in (b) can
only refer to inanimate objects with that gender. The resulting system is illus-
trated in Table 9.1, adapted from Fernández-Ordóñez (1999); for simplicity, only
the singular forms are given.
A result of this clitic/agreement system is that animate direct objects are
marked in the same way as indirect objects,29 a distribution essentially identi-
cal to that of the preposition a with objects.30 The presence of DOM patterns
in both preposition marking and clitic form in Spanish compounds the pres-
sure of DOM marking on the Basque system, perhaps helping to account for
the causative DOM patterns discussed above.
There is a further aspect of DOM which highlights the degree of conver-
gence with Spanish. As noted earlier in this section, DOM in Basque is more
widespread and mainstream in causatives than in basic direct object mark-
ing, where it is largely stigmatized. Similarly, le as a feminine direct object
clitic, although apparently spreading, is largely restricted to colloquial speech
and lower socio-economic status (Fernández-Ordóñez 1999). However, DOM in
Spanish causatives is common even in educated speech in the Spanish of the
Basque Country and neighboring areas, so that the patterns in (71) are standard:
29 In addition, as Odria (this volume) points out, DOM contexts are often sensitive to speci-
ficity, and both the appearance of the clitic and of the preposition a are more likely to
occur with definite, known, direct objects.
30 It is doubtful that the emergence of this clitic pattern in Basque Spanish may in turn be
due to Basque influence. Animacy does play a role in the shape of locative cases, but these
facts are quite different from DOM.
314 ortiz de urbina
b. La / lo hice caer.
it.fem / it.masc made fall
‘I made it fall.’
The feminine and masculine direct object clitic in (b) are preferably interpreted
as inanimate.
Extensive bilingualism and the growing impact of ‘new speakers’, bilingual
speakers with Spanish as the dominant language who have acquired/learned
the language in a school environment (Ortega et al. 2015) are most likely having
a deep impact on Basque. The complex interactions among the systems of dif-
ferent varieties of Basque and between Basque and the surrounding Romance
languages, as also seen in other contributions to this volume, make of this con-
tact a promising ground for linguistic research.
Acknowledgements
Research for this paper was partially funded by the Spanish MINECO project
FFI2014–51878-P. Special thanks to Ane Berro and Beatriz Fernández for their
patience and comments on this paper. Examples taken from the General
Basque Dictionary (OEH) have been left with the code given there identifying
their source.
References
Aissen, Judith & David Perlmutter. 1983 [1976]. “Clause Reduction in Spanish”. In Studies
in Relational Grammar. Chicago: Chicago University Press. 360–403.
Albizu, Pablo. 2009 “Construcciones inacusativas con dativos posesivos y dativos de
interés en vasco: un análisis derivacionalista”. Talk presented at the Seminario de
Lingüística Teórica, CSIC, Madrid, May 11, 2009.
Baker, Mark. Incorporation: A Theory of Grammatical Function Changing. Chicago: Uni-
versity of Chicago Press.
Berro, Ane. 2015. From Event Structure to Syntactic Categories in Basque. Ph.D. disserta-
tion, University of the Basque Country-University Bordeaux Montaigne.
Bordelois, Ivonne. 1974. The Grammar of Spanish Causative Complements. Ph.D. disser-
tation, M.I.T.
Comrie, Bernard. 1989. Language Universals and Linguistic Typology. Chicago: Chicago
University Press.
De Rijk, Rudolf P.G. 2008. Standard Basque. A Progressive Grammar. Cambridge, MA:
MIT Press.
complex causative verbs and causees in basque (and romance) 315
Dixon, Richard M.W. 2000. “A Typology of Causatives: Form, Meaning and Syntax”. In
Richard M.W. Dixon & Alexandra Y. Aikhenwald (eds.), Changing valency. Case stud-
ies in transitivity. Cambridge: Cambridge University Press, 30–83.
Dunlos, Laurence. 2001. “Event coreference in causal discourses”. In Pierrette Bouillon
& Federica Busa (eds.), The Language of Word Meaning. Cambridge: Cambridge Uni-
versity Press.
DHM. 1989. [Deustuko Hizkuntzalaritza Mintegia: Andolin Eguzkitza, Javier Ormaza-
bal, Jon Ortiz de Urbina, Koldo Sainz & Myriam Uribeetxebarria] “Inkorporazioa
perpaus kausatiboetan” [Incorporation in causative clauses]. In Pello Salaburu (ed.),
Sintaxi Teoria eta Euskara. San Sebastián: EHU Argitarapen Zerbitzua. 87–108.
Etxepare, Ricardo. 2003. “Valency and argument structure”. In José Ignacio Hualde &
Jon Ortiz de Urbina (eds.), A Grammar of Basque. Berlin/New York: Mouton De
Gruyter. 362–426.
Fernández, Beatriz & Milan Rezac. 2016. “Differential Object Marking in Basque Vari-
eties”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
Grammar of Basque. Amsterdam: John Benjamins. 93–138.
Fernández, Beatriz. This volume. “On non-selected datives: Ethical datives in Basque
and Spanish”.
Fernández-Ordóñez, Inés. 1999. “Leísmo, laísmo y loísmo”. In Ignacio Bosque & Vio-
leta Demonte (eds.), Gramática Descriptiva de la Lengua Española. Madrid: Espasa
Calpe, 1317–1397.
Folli, Raffaella & Heidi Harley. 2007. “Causation, obligation, and argument structure:
On the nature of little v”. Linguistic Inquiry 38(2). 197–238.
Laka, Itziar. 1993. “Unergatives that Assign Ergative, Unaccusatives that Assign Accusa-
tive”. In Jonathan D. Bobaljik & Colin Phillips (eds.), Papers on Case and Agreement
I, MITWPL 18. Cambridge, MA. 149–172.
Landa, Alazne. 1995. Conditions on Null Objects in Basque Spanish and their Relation to
leismo and Clitic Doubling. Ph.D. dissertation, Univesity of Southern California.
Levin, Beth. 1983. On the Nature of Ergativity. Ph.D. dissertation, MIT
Odria, Ane. This volume. “Differential Object Marking in Basque and Spanish Dialects”.
OEH [Orotariko Euskal Hiztegia-General Basque Dictionary]. Available online at www
.euskaltzaindia.eus/oeh.
Ormazabal, Javier & Juan Romero. 2007. “The Object Agreement Constraint”. Natural
Language and Linguistic Theory 25. 315–347.
Ortega, Ane, Jacqueline Urla, Estibaliz Amorrortu, Jone Goirigolzarri & Belen Uranga.
2015. “Linguistic identity among new speakers of Basque”. International Journal of
the Sociology of Language 231. 85–105.
Ortiz de Urbina, Jon. 2003. “Causatives”. In José Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A Grammar of Basque. Berlin/New York: Mouton De Gruyter. 592–607.
Ortiz de Urbina, Jon & Beatriz Fernández. 2016. “Datives in Basque Bivalent Unerga-
316 ortiz de urbina
tives”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
Grammar of Basque. Amsterdam: John Benjamins. 67–92.
Oyharçabal, Bernard. 2003. “Lexical Causatives and Causative Alternation in Basque”.
ASJU Supplement XLVI. 223–253.
Oyharçabal, Bernard. 2007. “Basque Ditransitives”. Ms, IKER-CNRS.
Perlmutter, David. 1978. “Impersonal Passives and the Unaccusative Hypothesis”. In
Fourth Annual Meeting of the Berkeley Linguistics Society, vol. 4. 157–190.
Pylkkänen, Liina. 2008. Introducing Arguments. Cambridge, MA: The MIT Press.
Rezac, Milan. 2009. “Person restrictions in Basque intransitives”. Lapurdum 2009, 13.
305–322.
Treviño, Esthela. 1994. Las causativas del español con complemento infinitivo. México: El
Colegio de México.
Tubino, Mercedes. 2011. Causatives in Minimalism. Amsterdam/Philadelphia: John Ben-
jamins.
Zubizarreta, M.L. 1985. “The relation between morphophonology and morphosyntax:
The case of Romance causatives”. Linguistic Inquiry, 16:2. 247–289.
chapter 10
1 Introduction
Properties Examples
internal causal rhematic content The ground in wet [because it has rained]
In the Romance languages, as in English, causal clauses, like other types of sub-
ordinate clauses, may be headed by a conjunction (e.g. Sp porque ‘because’,
como ‘as’, que ‘that’, pues ‘as, so’) or by more complex expressions, which show
different degrees of grammaticalization in this function (e.g. Sp ya que ‘since’,
puesto que ‘since’, dado que ‘given that’). In Basque, on the other hand, there is
more structural heterogeneity. Finite clauses with causal value in Basque may
contain verbal suffixes, verbal prefixes and/or clause initial conjunctions, as
shown in (1) with the example ikusi dugu ‘we have seen (it)’:
In spite of all the obvious differences between Basque and Romance causal
elements, there are also many similarities, especially when we consider the his-
torical origin of causal connectors. In Table 10.2, we make explicit some of these
similarities by comparing Basque causal subordinators with etymologically
equivalent expressions in the main Romance languages with which Basque
has been in contact, Spanish, Gascon and French.1 Since contact with French
is relatively recent, we indicate only the most usual forms for this language.
(Another historically important contact language was Navarro-Aragonese
Romance, which in all relevant respects coincides with Castilian Spanish). Sub-
ordinators that have become obsolete are signaled with “†”.
As shown in Table 10.2, in both Basque and Romance we find causal sub-
ordinators that have developed historically from question words, from gen-
eral complementizers (‘that’) and from time and modal expressions, as we
also find in many other language (Kortmann 1997, 1998; Pérez Saldanya 2014:
3479–3483). On the other hand, Basque shows at least one grammaticaliza-
tion phenomenon that has no parallels in Romance: the development of a
causal construction involving the coordinating conjunction eta ‘and’ as a verb
enclitic.
We also find other differences in the specific meanings and uses of different
causal structures. In modern Basque a direct answer to the question ‘why’ gen-
erally requires the construction with -(e)lako ‘because’ attached to the inflected
verb. This suffix is thus the basic equivalent of Sp porque, Fr parce que or Eng
because. However, unlike Sp porque, French parce que and Eng because, the
Basque connector is usually only employed in internal causal clauses. Other
elements such as preverbal bait- ‘since’, clause-initial zeren ‘since, because’ and
postverbal eta have a larger set of causal usages in Basque.
1 We have obtained the information about Occitan from Alibèrt (1935–1937) and Fernández
González (1985). For Gascon, we have consulted Palay (1974) and the Bearnais dictionaries in
Lexilogos (2002–2017).
320 hualde and pérez saldanya
table 10.2 Causal subordinators in Basque and Romance classified by etymological origin
Interrogative (zer ‘what’ >) (Lat. quia >) †ca (Lat. quāre >) †car car
zeren … bait-
zeren … -(e)n
ze (colloquial)
General bait- (Lat. quid >) que que que
complementizer ezen
‘that’+ ‘of ’ (Bq) -(e)lako porque per çò que parce que
‘for’ + ‘that’ (Rom) perquè
Locutions with nouns dela kausa/bide/medio a/por causa de que a causa que à cause que
and ‘that’ † por razón que per (a)mor/pramor
que
Time adverb -(e)nez gero (Lat.post >) pues (que) † pus (que), puix que puisque
(or other temporal ya que
items) puesto que vist que vu que
del moment que
Modal adverb nola … bait- (Lat. quomodo >) †com/coma comme
(& interrogative ‘how’) nola … -(e)n como
-(e)nez
Copulative conjunction (-la) eta n.a. n.a. n.a.
In this section, we consider the basic connectors used in internal causal clauses
in Basque. As mentioned, internal causal clauses are always rhematic and in
grammaticalization processes in causal subordination 321
the content domain. As such, they can be used to provide a direct answer to
the question ‘why?’. With this function, modern English uses the conjunction
because, Spanish has porque (por ‘for’+ que ‘that’) and French parce que (par
‘for’+ ce ‘this’ + que ‘which’). In contemporary standard Basque, the main sub-
ordinator used in this context is the suffix -(e)lako. Morphologically, -(e)lako
can be analyzed as bimorphemic, with the relational suffix -ko attached to the
subordinating suffix -la ‘that’ (cf., e.g., ikusi dugu-la ‘that we have seen it’, ikusi
dugu-la-ko ‘because we have seen it’). Some examples illustrating the use of
-(e)lako are given in (2) (adapted from EGLU 2005: 192) and (3) (from a journal-
istic interview). We provide Spanish translations as well for comparison:2
Unlike their Romance and English counterparts, causal clauses with -(e)lako
often appear before the main clause as in (4) (adapted from Artiagoitia 2013:
715):
2 The following abbreviations are used in the glosses: abl ‘ablative’, aux ‘auxiliary’, comp ‘com-
plementizer’, det ‘determiner’, erg ‘ergative’, fut ‘future’, gen ‘genitive’, imp ‘imperative’,
impf ‘imperfective’, inst ‘instrumental’, loc ‘locative’, neg ‘negation’, part ‘partitive’, pl ‘plu-
ral’, rel ‘relative’, subor ‘subordinator’.
322 hualde and pérez saldanya
The reason for this difference between the languages is that in Basque focal-
ized elements are generally placed in preverbal position (Artiagoitia 2003: 714,
Ortiz de Urbina 2003: 454–455).
In addition to its use as a causal connector, -(e)lako can also be used to link
an inflected clausal complement to a following head noun (a function that can
also be fulfilled by the suffix -(e)n). This is in fact the only use in which this
verbal suffix is found in early Eastern authors (again, Spanish translations are
provided for comparison):
The bridging context between these two rather distinct usages of -(e)lako might
be found in potentially ambiguous examples like (7), where the embedded
clause can be understood as a complement of the noun poz ‘happiness’, but
also as a causal complement of the main clause:3
3 In modern Eastern usage, there is a formal differentiation between -(e)lakotz, which is used
as causal suffix, and -(e)lako, used with complements of nouns and adjectives.
grammaticalization processes in causal subordination 323
4 Modal adverbs also take the suffix -(e)la (e.g. honela ‘thus, this way’, bestela ‘otherwise’). When
these adverbs function as a complement of a noun we find the suffix sequence -la-ko (e.g.
honelako galderak ‘this type of questions’, bestelako iritziak ‘different oppinions’, also with an
empty noun head: bestelakoa da ‘it is different’).
324 hualde and pérez saldanya
the ground is wet) or speech-act causals (Go now, because it is about to rain),
where other connectors are employed instead (see sections 4 and 5). Its (pos-
sible) origin in inflected complements of nouns and adjectives may account
for the restriction of -(e)lako to the internal causal context. In both cases, the
subordinate clause is an internal adjunct, and if -(e)lako causals derives from
-(e)lako noun complements, the internal character of the former may be a con-
servative property of the latter. Nevertheless, nowadays -(e)lako is sometimes
also used in external causal clauses, as in (9) and (10), both from the same
novel:
(9) Endredo bat zen enkargu hura, idazten ari nintzen nobelatik ateraraziko
ninduelako.
‘That request was a complication, because it would take me out of the
novel I was writing.’ (J.L. Zabala, Ospa, 2017)
Sp Ese encargo era un enredo, porque me sacaría de la novela que estaba
escribiendo.
Leaving aside causal subordinators that have their origin in modal and time
conjunctions (see sections 6 and 7), we find two additional main subordinators
in external causal clauses in Basque, whether content, epistemic or speech-act:
the verbal prefix bait- (attached to the inflected verb, sometimes with elision
of t), as in (11), and enclitic -eta (usually enclitic to the inflected verb, but not
exclusively, and frequently phonologically reduced to ta), as in (12).
grammaticalization processes in causal subordination 325
(11) Nik atzo hik baiño bide gehiago iragan nian inguratu bainuen
I yest. you but way more pass aux surround since.aux
mundu guztia.
world all
‘Yesterday I walked more than you did, because I went around the whole
world.’ (Axular, Gero: 21, 1642)
As in the two examples just mentioned, both subordinators are generally used
in rhematic clauses, and are thus normally found in the right periphery of the
sentence. Nevertheless, causals with bait- may also be thematic, located at the
left periphery, as in (13).
(15) Utzten ditut plazer haren aitzingibelak, anarteraiñoko, eta ondoreko egit-
ekoak, eta atsekabeak, handiak baitira.
‘I leave aside the pros and cons of that pleasure, the preceding ones, and
those that follow, which are big (or: because they are big).’ (Axular, Gero:
251, 1643)
Sp Dejo ahora a un lado las circunstancias de dicho placer, los que-
braderos de cabeza y los disgustos que le preceden y le siguen, que son
grandes. (L. Villasante’s translation)
It is likely that the grammaticalization process that has led to bait- being used
as a causal connector is the same that is responsible for a general subordina-
tor being able to introduce external causal clauses, as is the case with Spanish
que ‘that’. For instance, in a Spanish example such as (16a), literally, ‘I am leav-
ing, that is late’ an inference of causality allows the general subordinator que
‘that’ to be interpreted as ‘because’. We may envision a similar process in the
grammaticalization of Bq bait- as a causal connector, as in (16b).
This explanation, however, does not account for why bait- has become special-
ized as a causal subordinator in modern Basque (whereas Sp que has not had
this development). It also leaves unexplained the fact that bait- can be used in
thematic causals, whereas Sp que is restricted to rhematic ones, especially in
speech-act causals (Pérez Saldanya 2014: 3510–3514).
grammaticalization processes in causal subordination 327
This causal construction, which does not have a structural parallel in Romance,
appears to have arisen from coordinated structures with some specific proper-
ties, as explained in EGLU (2005: 201) which provides the example in (18) to
illustrate the process:
From the earliest texts in Basque we find causal clauses headed by conjunctions
such as zeren (or zerren), zeren eta and zergatik. In these subordinate clauses,
the inflected verb may take the prefix bait-, as in (19) and (20), the suffix -(e)n
or even -(e)lako. It may also be unmarked as a subordinated verb, as in (21).
5 The conjunction edo ‘or’ is also used as a postposition; specifically it is postposed to measure
phrases to convey the meaning ‘or so’, e.g. lauretan edo ‘at 4:00 or so’. In addition, both dis-
junctive conjuntions edo and ala are used as tags in questions, like German oder, cf. Eng or
what, Sp o qué.
grammaticalization processes in causal subordination 329
(20) Zeren Jeinkoa egun oroz ongi ari baitzaigu guk ere hala
because God day all.inst well act comp.aux we.erg too thus
behar harzaz untsa orhitu.
must 3.inst well remember
‘Since God rewards as every day, we too should remember Him in grati-
tude.’ (Etxepare, Primitiae: I.17, 1545)
Zeren and the nowadays stigmatized form zergatik are polyfunctional causal
conjunctions, in the sense they may introduce internal clauses, as in (19) above,
and external clauses, either thematic, as in (20), and rhematic, as in (21). On the
other hand, zeren eta always introduces external clauses, mostly with epistemic
value, as in (22).
As has been already mentioned, zerga(i)tik is also the interrogative word ‘why’,
so that there is an obvious parallelism with Sp por qué ‘why’ and porque
‘because’. The structure of this form is transparently zer ‘what’+ gai ‘matter’+
tik ‘through, from’. As for zer(r)en, it is the genitive form of zer ‘what’, which in
earlier times was also used as benefactive; that is, it may be literally translated
as ‘for what’. The use of zeren as ‘why’ is also found in early texts, as in (23)–
(24):
(25) question word > conjunction in internal causal clauses > polyfunctional
causal conjunction > conjunction in external causal clauses
Within the diachronic path in (25), zeren appears to be in the third stage, as it
clearly has a polyfunctional use, as mentioned above. In this respect it is com-
parable in its range of functions to Spanish porque and to car in French, Occitan
and Old Catalan (and English because). We may distinguish a number of causal
contexts in which Bq zeren is employed.
a) First of all, it is used in internal causal clauses, with a meaning similar to that
of -(e)lako, as in (19) above, where the subordinate clause expresses the reason
for the subject fleeing, and in (26), where it expresses the reason for Peter being
grieved:6
(26) Triste zedin Pierris zeren erran baitziezon herenean, On dariztak niri?
‘Peter was grieved because he said unto him the third time, Lovest thou
me?’ (Leizarraga, Testamentu: John, 21:17, 1571)
6 All English translations from the Bible are from the King James version, which is roughly con-
temporary with Leizarraga’s Basque translation.
grammaticalization processes in causal subordination 331
(27) Zergatik ene lengoajea eztuzue aditzen? zeren ezin enzun baitirokezue
ene hitza.
‘Why do ye not understand my speech? even because ye cannot hear my
word’ (Leizarraga, Testamentu: John 8: 43, 1571)
(28) Eta haur erraiten du hunela, ez zeren dantzatzea bera hain gauza gaixtoa
den, baiña zeren dantzatik anhitz okhasino behar eztenik sortzen den.
‘And he says this this way, not because dancing is by itself such an evil
thing, but because from the dance many inappropriate occasions arise.’
(Axular, Gero: 268, 1643)
(31) Errege Daviti eman zerautzatenean Saulen harmak, baitziren ezin hobea-
goak eta halakoak, guztiarekin ere, probetxu baiño kalte gehiago egiten
zioten, trabatzen zuten. Zeren nola anarteraiñokoan bethi arzain ibili
baitzen, ezpaitzen oraiño harmetan usatua.
‘When they gave King David Saul’s weapons, even though they were the
best, they did him more harm than good; they bothered him. Because,
since up to that moment he had always been a shepherd, he was not yet
used to weapons’ (Axular, Gero: 61, 1643)
c) It may also occur in external causals in the speech act domain, as in the
following examples. In (32) the causal clause justifies a command, in (33) it
justifies a question and in (34) it motivates an exclamation:
(32) Boz eta alegera zaitezte, zeren zuen saria handi baita zeruetan
‘Rejoice, and be exceeding glad: for great is your reward in heaven’ (Leizar-
raga, Testamentu: Matthew 5:12, 1571)
332 hualde and pérez saldanya
(33) zenbat gehiago eztire bihotzez estekatuak, josiak beren tokiko mintza-
iari, zeren erlisionea kasik orok haurdanik mintzai hartaz ikhasia dute,
guk emen euskaraz bezala?
‘How many more are not tied and attached in their hearts to their native
language, for almost all of them were taught religion in childhood in that
language, just like we here in Basque?’ (H.U.Gonz 193, ap. Hualde & Ortiz
de Urbina 2003:730)
(34) Ai gure urrikalkizunak, galdu gara, egin du gureak. Zeren akhabatu hur-
ran da eguna, beheititu da iguzkia, eta itzala ere luzatuago da arrat-
saldean.
‘Poor us, we are lost, our luck has finished. Because the day has almost
ended, the sun has lowered and even the shade is longer in the evening.’
(Axular, Gero: 130, 1643)
(35) Gero iguzkia goratu eta, erre izan dirade, eta zeren ezpaitzuten errorik,
eiarthu izan dirade.
‘And when the sun was up, they were scorched; and because they had no
root, they withered away.’ (Leizarraga, Testamentu: Matthew: 13:6, 1571)
(36) Zeren amak diren koleratsu, halatan dira umeak ere faltatsu.
‘Because the mothers are quick to anger, the children are also defective.’
(Axular, Gero: 200, 1643)
e) Finally, zeren may introduce clauses that express both cause and finality,
which is also possible in Spanish with porque:
(37) Zeren unsa orhit ziten nizaz, ama eztia, / gogo honez erranen dut zuri
Abe Maria.
‘So that you may remember me, sweet mother, with good intention I will
say Hail Mary to you’ (Etxepare, Primitiae: 93, I.445, 1545)
In addition to this use as a temporal connector, gero may also have a causal or
conditional interpretation:
The word gero itself may take the instrumental suffix (gero-z) or both the
instrumental and the ablative (gero-z-tik). We may note that a similar process of
morphological reinforcement is found in the Romance languages in its transla-
tion equivalents. Thus, in Spanish when pues (< post) acquired a causal value,
a new form después (des ‘from’+ pues) developed as a time adverb. In addi-
tion, there is a morphologically more complex nonstandard form endespués.
In Gascon as well we find similar reinforced forms: despuix (des + puix), apuix
(a +puix), d’après (de + après).
By morphological reanalysis, the sequence -(e)z gero may appear as ezkero
reflecting a pronunciation with devoicing of /g/ after a sibilant. There is also a
historical variant azkero, as in (44), where the initial vowel is from reanalysis
of the determiner sufix -a ‘the’. The reanalyzed form ezkero is alive in western
dialects, but it is no longer used in writing.
All the forms that we have mentioned may have either temporal or non-
temporal interpretation, but there is a tendency for inflected clauses with
-(e)nez gero to be used as causal, for participial clauses with -(e)z gero to have
conditional value and for -(e)nez geroztik to be used to express time, especially
in the South (EGLU 2005: 191). To some extent, this specialization is parallel to
that between pues and después in Spanish.
The conditional and causal uses of gero are already found in the first texts
from the 15th–16th centuries. It is thus not possible to trace the grammati-
calization path of this element. Nevertheless, similar evolutions, from time
expression to causal connector, have been well-studied for other languages.
The standard view is that the causal meaning is the result of conventionaliza-
tion of an inference: post hoc, ergo propter hoc (Geis & Zwicky 1971; Levinson
1995; Traugott & Dasher 2002: 16–17, 80; Pérez Saldanya 2014: 2536–2540, among
others). By hypothesis, the causal value would arise in potentially ambigu-
ous examples like (45), where precedence in time may also be interpreted as
cause:
The causal inference can only emerge in contexts like that in (45), in which the
temporal clause is external to the predicate of the main clause and appears at
the left margin of the sentence, in the thematic position (Gutiérrez Ordóñez
2000: 91; Goethals 2002: 111; Pérez Saldanya 2014: 2536–2540; Pérez Saldanya &
Hualde 2017).
The characteristics of the bridging context where the grammaticalization
process is triggered tend to persist (Heine 2002: 84). Thus, Sp pues and ya que
are only found in causal clauses external to the main predicate and this is also
true of causal clauses with gero in Basque (cf. also causal clauses with since in
English). For this reason, they cannot be used to answer the question ‘why’ in
any of these languages.
Moreover, at the initial stage in the grammaticalization process, causal con-
junctions of temporal origin are used in causal clauses providing given or pre-
supposed information, as in (46).
336 hualde and pérez saldanya
(49) Bere garaiko Itsasondo badoa, baina jende gehiago ere etorriko da, etxeak
eraikitzen ari direnez gero.
‘The Itsaondo of that time is going, but more people will come, because
they are building more houses.’ (Berria, 2004-10-02)
grammaticalization processes in causal subordination 337
(50) Baña aizen ezkeroz burutik maingutu artaraño, non nai dek nerekin
jokatu
‘But since you are so short of wit, you want to play with me everywhere.’
(Iturriaga, Ipuinak, 1842)
Sp ‘Pero como eres tan corto de mente, en todas partes quieres apostar
conmigo.’
(51) desengañoa arturik zegoen ezkeroz eskatu zeban mesedez laga zegiotela
edaten baso bat ur
‘As he was taken by disappointment, he asked them to please let him drink
a glass of water.’ (Iztueta, Gipuzkoako probintziaren kondaira, 1847)
Sp ‘como estaba desengañado, pidió que por favor le dejaran beber un vaso
de agua’
(52) Baina Lotek hainbeste erregutu zienez gero, berarekin joan ziren etxera.
‘But, since Lot begged them so much, they went home with him.’
(Elizen Arteko Biblia: Hasiera [Genesis] 19:3, 2008)
Sp ‘Pero como Lot les suplicó tanto, se fueron con él a casa.’
Basque causal clauses with gero have moved farther along the grammatical-
ization path than Fr puisque, which is used only in thematic clauses in the
speech-act domain or expressing modality (Ducrot 1983; Zufferey 2012), but less
than Sp pues and ya que, which have completely lost all temporal usages (Pérez
Saldanya 2014).
Two factors may have contributed to the extension of usages of gero in con-
tent thematic clauses. One of them may have been the development of adver-
bial clauses with -enez with thematic content value (e.g. ikusi dugunez ‘as we
have seen’, section 7.1), which may have created an equivalence between -enez
and -enez gero in this type of causal clause (Artiagoitia 2003: 717–719). The other
factor could be the desire to avoid the use of clause-initial nola ‘how, as’ (sec-
tion 7.2), since clause-initial subordinators have sometimes been regarded as
calques from Romance syntax.
338 hualde and pérez saldanya
The following 19th century example cited in Hualde & Ortiz de Urbina (2013:
834) shows the use of nola as a causal conjunction ‘since, as’ and as a compar-
ative (‘like, as’):
As can be seen in the Spanish translations of the examples above, the use of
nola as a subordinator is very much like in Spanish, where interrogative cómo
‘how’ has also developed an unstressed variant como with these modal and
causal values:
However, whereas Spanish unstressed como can also introduce purely modal
clauses, in Basque a different construction is used in this case, with bezala ‘like,
as’, moduan ‘in the manner’ or legez (Bizkaian, from lege ‘law’ in the instrumen-
tal case) ‘like, as’, postposed to a verb bearing the subordination marker -(e)n:
c. *Nola esaten dizudan egin behar duzu / *Egin behar duzu nola esaten
dizudan
Nevertheless, in a relatively small western area the construction with legez ‘as,
like’ has expanded its use to cover the same space as Spanish como, including
subordinate clauses with both a modal and a causal interpretation (Zuazo 2017:
61). This seems to be a very recent development.
Notice that the use of nola as both an interrogative ‘how’ and as a causal
subordinator ‘as, since’ is parallel to that of zeren ‘why; because’. It can thus be
explained in a similar fashion, as a possible syntactic calque from Romance.7
7 Modal expressions may also give rise to causal conjunctions, cf. Eng how come? In Sp we find
cómo es que, cómo así and in Bq nolatan ‘why, how come’, with nola ‘how’ inflected in the
locative case.
grammaticalization processes in causal subordination 341
This construction is historically very recent, dating back only to the 18th cen-
tury. It appears to have resulted from the omission of gero in the older construc-
tion -enez gero (§6). The first examples of -enez without gero have modal value
and involve expressions such as dirudienez, antza denez ‘as it seems, appar-
ently’, dioenez ‘as s/he says’ and the like.
Clauses with -enez have a very similar distribution and interpretation as
Spanish subordinate modal and causal sentences headed by unstressed como
‘as, like; since’. Thus, the Spanish translation of the subordinate clause in the
three examples given in (61)–(63) would be, respectively, Como hemos sabido
‘As we have found out’, Como no tenemos una administración de todo el País
Vasco ‘Since we do not have administration of the whole Basque Country’ and
Como no tenemos gimnasio ‘Since we do not have a gym’. This thus seems to be
a case of convergence with Spanish in recent Basque usage, even if the actual
morphosyntactic mechanisms are quite different.
8 Conclusions
The Basque subordinators bait- and -(e)ta have a certain functional paral-
lelism with the Spanish conjunction que (and Old Spanish ca), since all of them
can appear in external rhematic causal clauses. Nevertheless, Bq bait- may also
be found in external thematic clauses.
In both Basque and Spanish (and other Romance languages) we find causal
subordinators of temporal origin: Bq gero, Sp pues (que), ya que. These subor-
dinators, in a first stage, are mostly used in thematic causals in the speech-act
domain. Later, they tend to broaden their usage to other external causal mean-
ings. This evolution has been different in Basque in Spanish. In Basque the
broadening of usage has been towards content thematic clauses. In Spanish,
instead, it has been towards rhematic clauses, in particular, those in the con-
tent domain.
Finally, in both languages we also find causal subordinators of modal origin,
employed in content thematic causal clauses: Bq -nez, nola and Sp como.
Acknowledgements
The first author acknowledges the research funding received from the Spanish
Ministry of Economy and Competitiveness (FFI2014–51878-P).
References
Alibèrt, Lois. 1935–1937. Gramatica occitana segons los parlars lengadocians. Montpel-
hièr: CEO.
Artiagoitia, Xabier. 2003. “Adjunct subordination”. In Jose Ignacio Hualde & Jon Ortiz
de Urbina (eds.). 2003. 710–762.
Azkue, Resurrección María de. 1923–1925. Morfología Vasca. Published in fascicles in
Euskera. [Repr. 1969, Bilbao: La Gran Enciclopedia Vasca.]
Baños, José Miguel. 2011. “Causal clauses”. In Philip Baldi & Pierlougi Cuzzolin (eds.),
New perspectives on the historical Latin syntax, 4: Complex sentences, grammatical-
ization, typology. Mouton de Gruyter: New York-Amsterdam. 195–234.
Baños, José Miguel. 2014. Las oraciones causales en latín: su evolución diacrónica.
Madrid: Escolar y Mayo editores.
Ducrot, Oswal. 1983. “Puisque: essai de description polyphonique”. Revue Romane 24.
166–185.
EGLU = Euskaltzaindia. 2005. Euskal gramatika lehen urratsak: Lehen urratsak-VI, (Men-
deko perpausak-2, baldintzazkoak, denborazkoak, helburuzkoak, kausazkoak, kontse-
ziozkoak eta moduzkoak). [Available online: www.euskaltzaindia.eus].
344 hualde and pérez saldanya
Ortiz de Urbina, Jon 2003. “Word order”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.). 2003. 448–459.
Oyharçabal, Bernard. 2003. “Relatives”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.). 2003. 762–823.
Palay, Simin. 1974. Dictionnaire du béarnais et du gascon modernes (Bassin Aquitain):
embrassant les dialectes du Béarn, de la Bigorre, du Gers, des Landes, et de la Gascogne
maritime et garonnaise. Paris: Centre National de la Recherche Scientifique.
Pérez Saldanya, Manuel. 2014. “Oraciones causales”. In Concepción Company (dir.), Sin-
taxis histórica de la lengua española. Tercera parte: Preposiciones, adverbios y conjun-
ciones. Relaciones interoracionales, vol. 3. Ciudad de México: Universidad Nacional
de México / Fondo de Cultura Económica. 3447–3610.
Pérez Saldanya, Manuel and José Ignacio Hualde. 2017. “From theme to rheme: The evo-
lution of causal conjunctions of temporal origin in Catalan”, Studies in Hispanic and
Lusophone Linguistics 10. 319–348.
Pineda, Anna & Manuel Pérez Saldanya. 2017. “Les construccions causals i finals”. In
Josep Martines & Manuel Pérez Saldanya (eds) Gramàtica del català antic. In prepa-
ration.
Santos Río, Luis. 1982. “Reflexiones sobre la expresión de la causa en español”, Studia
Philológica Salmanticensia 6. 231–277.
Sweetser, Eve. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects
of Semantic Structure. Cambridge: Cambridge University Press.
Traugott, Elisabeth C. & Richard B. Dasher. 2002. Regularity in Semantic Change. Cam-
bridge: Cambridge University Press.
Villasante, Luis. 1981a. “Las oraciones causales en Axular (I): La construcción de zeren
con partícula que modifica el verbo”. Fontes linguae vasconum: Studia et documenta
37. 9–18.
Villasante, Luis. 1981b. “Las oraciones causales en Axular (II): Causales subordinadas y
Causales coordinadas”. Fontes linguae vasconum: Studia et documenta 38. 9–22
Villasante, Luis. 1982a. Las oraciones causales en Axular (III): Oraciones causales fron-
terizas. Fontes linguae vasconum: Studia et documenta 39. 9–20.
Villasante, Luis. 1982b. “Las oraciones causales en Axular (IV): Oraciones causales con-
struidas a base de bait solo”. Fontes linguae vasconum: Studia et documenta 40. 359–
386.
Zuazo, Koldo. 2017. Mendebaleko euskara. Donostia: Elkar.
Zufferey, Sandrine. 2012. “‘Car, parce que, puisque’ revisited: Three empirical studies on
French causal connectives”. Journal of Pragmatics 44. 138–153.
Name Index
EGLU 321, 323, 327–328, 335 Grodzinsky, Yosef 4, 222, 230, 235
Eguren, Luis 86 Gross, Maurice 192–193
Elordieta, Arantzazu 42–44, 46, 48, 75, 87, Gutiérrez Ordóñez, Salvador 335
248
Elordieta, Gorka 44, 46, 75, 248 Haase, Martin 66
Embick, David 7, 82–83, 87–88, 90–92, 95, Haddican, Bill 48
99, 107, 111–113 Hale, Kenneth 139–140, 145, 176, 179, 198,
Epelde, Irantzu 212 203–204, 206
Escandell Vidal, Victoria 44–46 Harley, Heidi 140, 148–149, 168, 205, 292–
Etxeberria, Urtzi 22 293
Etxepare, Ricardo 42, 46, 145, 151, 194, Harre, Catherine 73
196, 203–204, 207–208, 212, 255, 261, Haspelmath, Martin 33
299 Haugen, Jason 204, 206
Euskaltzaindia 26–27, 37, 48, 50, 287 Heine, Bernd 335
Eythórsson, Thórhallur 139, 155 Hernanz, Maria Lluïsa 264
Hidalgo, Bittor 40, 42
Fábregas, Antonio 90–91, 106, 244, 251, 259– Hofmann, Johann Baptist 330
260 Hook, Peter E. 179, 181
Fernández, Beatriz 4, 8, 14, 24–25, 82, 145– Hopper, Paul J. 179, 181
146, 148, 198, 204, 221–223, 225, 228, 232, Horn, Laurence R. 4, 8, 221–223
234–235, 243–244, 246–247, 249, 251, Hualde, José Ignacio 2–3, 6, 9, 14, 44, 46, 51,
255–257, 261–262, 266–267, 269, 280, 75, 87, 92, 110, 114, 116, 248, 332, 335–
300, 305, 310 336, 339
Fernández González, José Ramón 319 Huidobro, Susana 8, 221, 223, 225–229, 231–
Fernández Lagunilla, Marina 221 233, 239
Fernández Ordóñez, Inés 243, 252–253, 255,
260, 312–313 Iatridou, Sabine 94, 112, 114
Fernández Soriano, Olga 265 Ibarra, Orreaga 248
Fleischman, Suzanne 60 Irurtzun, Aritz 46–47
Folli, Rafaella 140, 168, 292–293 Izvorski, Roumyana 112
Franco, Jon 8, 221, 223, 225–229, 231–233, Jackendoff, Ray 178–179, 189, 198
239, 243, 253 Jaeggli, Osvaldo 4, 221, 228, 230
Jaxontov, Sergej 93
García Fernández, Luis 91 Jespersen, Otto 178
García Murga, Fernando 182, 189, 198, 200, Jiménez Fernández, Ángel 44, 46
203–204 Jouitteau, Mélanie 4, 222, 228, 230, 233
Gawron, Jean Marc 106, 108
Gehrke, Berit 5, 7, 83, 84, 112, 120, 128– Kayne, Richard S. 18, 24, 139, 155
131 Kearns, Kate 178–179
Geis, Michael 335 Keller, Frank 139, 155
Geuder, Wilhelm 178, 181 Kennedy, Christopher 97
Giry-Schneider, Jacqueline 185, 193, 201 Keyser, Jay S. 140, 145, 176, 179, 198, 203–204,
Goenaga, Patxi 33, 87 206
Goethals, Patrick 317, 335 Kiss, Katalin É. 46
Goldberg, Adele 194 Koizumi, Masatoshi 262
Gómez, Rikardo 59, 78 Koontz-Garboden, Andrew 106
Grafmiller, Jason 49 Kortman, Bernard 319, 330, 338
Greenberg, Joseph 17–18 Krajewska, Dorota 5, 73, 83–84, 86, 112, 114–
Grimshaw, Jane 178–190, 193–194, 198 117, 133, 325
name index 349
Kratzer, Angelika 7, 82–84, 88, 97, 105, 107, Mohanan, Tara 178
111–113, 119, 124–125, 141, 143 Mounole, Céline 2, 6, 15, 63–66, 68, 71–73,
75, 93, 243, 246–247, 249–250
Labelle, Marie 165, 168
Lafitte, Pierre 40, 47, 204 Nedjalkov, Vladimir 93
Lafon, René 2, 71
Lahiri, Aditi 178, 180 Odria, Ane 4, 8, 82, 222–223, 229, 232, 234,
Laka, Itziar 43, 145–146, 148, 204, 213, 278, 240, 243–244, 246–247, 249, 251, 261–
299 263, 266–267, 269, 310, 312–313
Lakarra, Joseba 72, 75 Ordóñez, Francisco 41
Landa, Alazne 8, 243, 252–255, 312 Ormazabal, Javier 251, 261, 303
Landa, Josu 261–262 OEH 20, 34, 257, 276–280, 287
Larson, Richard K. 176 Ortega, Iván 44–45, 47
Lazard, Gilbert 3 Ortiz de Urbina, Jon 3–5, 6, 8, 40, 42, 47, 73,
Ledgeway, Adam 60 76, 83, 86, 92, 99, 126, 139, 142, 204, 208,
Legrende, Géraldine 168 223, 232, 244, 255–257, 261–262, 269,
Lehman, Winfred P. 17 291, 300, 304–305, 322, 332, 339
Leonetti, Manuel 44–46, 243, 251, 259 Oyharçabal, Bernard [also Beñat Oihartza-
Levin, Beth 4, 7, 95, 105, 139, 141–143, 146, bal] 4, 29, 31, 87, 145–148, 197, 198, 204,
152, 160, 179, 194, 204, 278, 299 206, 207, 209, 210–213, 227, 261–262,
Levinson, Stephen 335 285, 287, 291, 293, 305, 325
Lin, Tzong Hong 176, 178
López, Luis 204 Palay, Simin 319
Luján, Marta 5, 83 Pancheva, Roumyana 112, 114
Parsons, Terence 113
MacKenzie, Ian 139, 155 Pastor, Alberto 38
Maienborn, Claudia 106 Perez Gaztelu, Elixabete 28
Maldonado, Ricardo 221, 223, 225, 228, 233, Pérez Saldanya, Manuel 3, 9, 317, 319, 326,
237 330, 332, 335–338
Manterola, Julen 65 Pensado, Carmen 243, 251, 258
Marantz, Alec 5, 7, 82, 189, 191 Perlmutter, David 7, 139, 143, 230, 234, 283,
Marco, Cristina 5, 7, 83–84, 128–130 291, 301
Marín, Rafael 5, 83, 90–91, 106 Pesetsky, David 205
Martinez, Arantzazu 178, 182, 185, 188, 194, Piera, Carlos 200
196–200, 203–206, 208, 213–214 Pineda, Anna 7, 152, 154, 194, 244, 258, 260–
Masullo, Pascual 198, 262, 264 261, 330
Mateu, Jaume 140, 148–149, 157–158, 168, Pinkster, Harm 93
183 Potts, Chris 222, 235
Mathieu, Éric 204 Pylkkänen, Liina 4, 187, 194, 223, 232, 297,
Matushansky, Ora 204 305–306
McClure, William 152
McFadden, Thomas 119, 187, 262 Quer, Josep 44
McIntyre, Andrew 95, 124–125
McNally, Louise 97 Rafel, Joan 192, 193
Merlan, Francesca 139 Ramchand, Gillian 91, 105, 140, 148, 178,
Mester, Armin 178, 179–190, 198 180–181, 189, 192, 214
Michelena, Luis [also Koldo Mitxelena] 15, Rapp, Irene 120
19, 42, 59, 65, 257 Rappaport Hovav, Malka 7, 95, 105, 139, 141–
Mithun, Marianne 139, 182, 204, 206 143, 146, 152, 160, 179, 194
350 name index
experiencer 177, 222, 230, 232, 240, 244, event 180, 189, 198, 237, 297
261–262, 265, 284, 288, 301 bare event noun 180
external possessor 221–223, 228, 230, configuration and theta-role 150, 167
232–233, 235, 239, 240, 244, 261, 262, configurations 140, 148
264 event structure 186
goal 233–234, 244, 262, 301, 303, events and states 148
309 flavor 149, 153, 169
high and low 223 in light verb constructions 176
in bivalent unergative 269 kinds and tokens 129, 131
in unergative and ditransitive 256, 260 exclamatives 45
malefactive 225 experiencer subject 177
subject co-referential 235, 239 extraposition 34, 37
typology 239
degree words 32, 33–34, 36, 38, 98 final clauses 323, 333
degree adverbs and PP s 34 focalization 42–44, 207, 209, 322, 330
extraction 39 contrastive 45, 48
demonstratives. See determiners corrective 42, 47, 52–53
derivation 17, 28 focal stress 44, 46
destinative 62 Focus Fronting 44
determiners 21, 37, 198, 213 Mirative Focus 44–46
and predicative agreement 65 presentational 52
definiteness 38–39 Quantifier Fronting 44, 46
demonstratives 21, 198 verb focalization 48
indefinite 32, 36 Verum Focus 44, 46
disjoint reference 125
DOM 243, 299, 306, 310 Galician 155
and dative marking 246 Gascon 2, 64, 66, 69, 76–78, 319, 334
and secondary predication 266 genitive 21–24, 329
and tense 251 and future tense 62, 65
animacy 250, 257 German 63, 84, 119–120, 122–124, 129–132,
human objects 247 155, 231, 235, 237, 262, 328
in causatives 310 Gipuzkoan 34, 38
marking 244 goal 24, 256, 261
objects and goal IO 256 grammaticalization 59–60, 63–65, 68, 72,
person 247, 252 73–76, 116–117, 179–180, 317, 318–319,
position of DOM objects 246 326–327, 333, 335, 337–338
specificity 249, 257 Greek 84, 94, 99, 107, 123, 124, 131–132
stigmatization 247, 310, 312 Guarani 246
vs. dative in bivalent unergatives 255
Dutch 99, 152, 155, 291 heaviness 14, 49
Hebrew 222, 235
English 60, 63, 76, 88, 94, 96, 99, 101, 102, Hindi 178, 245
107, 119, 124, 178–180, 185, 206, 226, 229, Hungarian 21, 278
262, 283, 289, 302, 310, 318, 323, 328,
330 imperatives 69
ergative 126, 131, 142, 190 impersonal 296
ergativity incorporation 146, 202, 204, 213
and telicity 152 and phonology 210
semantic split 139, 151 inference 317, 326–327, 335, 338
354 subject index
vs. synthetic 178, 200, 203 unergative 104, 139, 203, 212, 290–291,
vs. transitive constructions 204 298
vs. Vague Action Verbs 179, 199 and telicity 143
weather 190, 197–198 and Voice phrase 141
without faire and hacer 184 built on light verbs 144
periphrases 75, 250. See also aspect case and auxiliary 141
radical of Bsq verbs 62–63, 67–68, 277, complex 145
279 causative 278
stative 140, 167, 285 semantic classes 145
stative predicative constructions 163 dialectal variation 150, 152, 160
statives and theta-role 167 initial arguments 143
synthetic 51, 71, 78 stative 162
vs. compound forms 71 verb classes 143
unaccusative 41, 103, 139, 142, 166, 290– weather verbs 146, 286
291, 295, 297–298, 308 Voice phrase 141, 149, 166
argument role 141
case and auxiliary 140 word order 14, 233, 307–309
Event and State phrases 168 neutral 40
telicity 168
verb classes 141