GFs in Basque

I. Subject vs object functions in Basque and their morpho-syntactic properties

Basque is an ergative language, that means it is a language where subjects of transitive verbs (A) are marked for the ergative (Erg) case and subjects of intransitives (S) and direct objects of transitives (O) are both marked for the absolutive (Abs) case. Thus ergativity presents apparent difficulty for identifying morpho-syntactic properties of Subject and Object since not only subject arguments of transitive and intransitive verbs are marked differently, but, moreover, direct objects are marked in the same way as some subjects. Here are some examples (J.I.Hualde & J.Ortiz de Urbina 1993:72):

1. Ni-Æ etorri n-aiz

‘I arrived’

2. Ni-k liburu-a-Æ irakurri d-u-t

‘I read the book’

But despite ergativity Basque grammar makes use of the notions subject and object much in the same way as many other languages of the world. I’ll first briefly discuss the three major coding properties such as word order, case marking and cross-referencing that fail in identifying the grammatical functions of subject and object and then turn to some other syntactic properties that come to rescue in this question.

Basque has a free word order, i.e. the arguments can appear in any order. (It should be noted that there are some restrictions on word order. For example, no constituent can intervene between the verb and the auxiliary in non-negative sentences.)

Here is an example of the sentence in Basque with all the possible alternations in the word order (J.I.Hualde & J.Ortiz de Urbina 1993:73):

3. ni-k Jon-i liburu-a-Æ ema-n d-ieza-io-ke-t

‘I can give the book to John’

The sentence above can appear with any of the listed below alternations in the word order:

1. nik liburua Joni eman diezaioket
2. liburua Joni nik eman diezaioket
3. Joni nik liburua eman diezaioket
4. Joni eman diezaioket liburua nik

Therefore, as we see, it is impossible to distinguish the grammatical function of subject/ object by the word order.

1. Case marking

There are three grammatical cases in Basque: Ergative, Absolutive and Dative that are indicated by a case suffix. Subjects of transitive verbs are always marked by the ergative case marker –k; subjects of intransitives and objects of transitives are marked by the absolutive case marker -Æ (zero); and the indirect objects are marked by the dative case marker -(r)i (we will consider them later). This is shown in examples (1-3) repeated her for convenience:

4. Ni-Æ etorri n-aiz

I-A arrive 1sg.Abs-aux.

‘I arrived’

5. Ni-k liburu-a-Æ irakurri d-u-t

I-Erg book-det-Abs read 3sg.Abs-aux.-1sg.Erg

‘I read the book’

6. Ni-k Jon-i liburu-a-Æ ema-n d-ieza-io-ke-t

I-Erg John-Dat book-det-Abs give-asp 3sg.Abs-aux-3sg.Dat-mod-1sg.Erg

‘I can give the book to John’

Hence, we see that case marking cannot serve the purpose of coding the morpho-syntactic properties of the subject and object grammatical functions in Basque.

2. Cross-referencing

Verbal inflection in Basque obligatorily agrees with all three case-marked arguments (ergative, absolutive and dative) that are introduced in the paradigm below (J.Ortiz de Urbina 1989:8):

All Basque verbs include an absolutive agreement marker. So the choice about the grammatical function of the nominal cross-referenced with the verb by absolutive marker can be made by presence or absence of the ergative verb inflection cross-referenced with the ergative nominal. If there is a nominal with ergative case then the absolutive corresponds to the object of the verb, if not, then absolutive marks the intransitive subject. The presence of dative doesn’t distinguish any coding property for subject/direct object functions.

So we are dealing with 4 logical combinations of Ergative, Absolutive and Dative:

Though in spite of the fact that we can derive indirect coding of the subject vs object grammatical functions the cross-reference system provides good evidence that the grammatical functions S and O pattern together opposed to the grammatical function S.

We have considered the major coding properties that fail in determining the grammatical functions of subject vs object. Nevertheless there is syntactic evidence that Basque as the majority of the world languages treats A and S as a class of subject with the behavior different from the object function.

The same way as other ergative languages, there is a well-known case of split in the person agreement system cross-referencing noun phrases on the verb.

In the past tense, first and second person ergative noun phrases are cross-marked in the verb by the absolutive agreement markers when the absolutive is third person. The third person is then unmarked or marked by zero which normally corresponds to the third person singular ergative. We are presenting the split (7) and non split (8) combinations below (J. Ortiz de Urbina 1989:11):

7. Ni-k polizi-ei harri-ak bota n-izk-i-e-n

‘I threw stones at the police’

8. Gu-Æ polizi-ek harrapatu g-int-u-z-te-n

‘The police caught us’

Summing up, where the split occurs ergative nominals are cross-referenced with absolutive agreement prefixes, and absolutives are left unmarked. The resulting agreement pattern restructures the four verb classes (Abs, Abs-Dat, Erg-Abs, Erg-Abs-Dat) reducing them into two:

That means that we have a clear-cut coding distinction between A and S on the one hand, and O on the other hand.

1. Participial complements - indirect wh-questions

Another argument for the subject/object distinction could be based on indirect wh-questions, which can appear as perfect participial complements of certain verbs as in the following examples (J. Ortiz de Urbina 1989:16):

9. Ez dakit [aldizkari-ak nor-i ema-n]

neg know magazine-3pl.Abs who-Dat give-prf

‘I don’t know who to give the magazines to.’

10. Zu-k ez dakizu [no-ra joa-n]

you-Erg neg know where-to go-prf

‘You do not know where to go.’

The embedded wh-questions show that the ellipsis in the perfect participial complement has the elliptic subject that can correspond only to the subject of the matrix verb and no case marked nominal may appear as subject of the participle as it is shown in the ungrammatical sentence below (J. Ortiz de Urbina 1989:17):

11. *Jon-ek ba daki ni-k zer egi-n

John-Erg know I-Erg what do-prf

‘John knows what I/me to do’

In addition, we should notice that the ellipsis in the complement can be either A (as in 9) or S (as in 10) but it can never be O.

Hence, indirect wh-questions serve as an argument revealing that Basque treats A/S as the same functional category with properties different from these of O.

2. Coordinate structures

A common phenomenon occurring in coordinated clauses with the coordinators edo ‘or’ and eta ‘and’ involves the elimination of all inflection in one of the conjunctions; its verb will appear in its participial form, while the verb in the other conjunct maintains its full shape including the auxiliary, if it is tensed. Some examples are shown below ((J. Ortiz de Urbina 1989:22):

12. Sartzen lagundu nion eta helbide-a eman gidari-ari

‘I help him to get on and I gave the driver the address’.

13. Miloi asko ba-nitu edo Hong-Kong-en bizi ...

‘If I had many millions or if I lived in Hong-Kong ...’

The data reveal that the only restriction in the coordinate structures is that both conjuncts share the subject argument, without specifying if it is absolutive or ergative as shown in 12 and 13: the missing subject in 12 would be absolutive in the first conjunct and ergative in the second conjunct, and the opposite holds true for 13.

Therefore, coordinate structures can serve another argument for distinction between the grammatical function of ‘subject’ as opposed to the function of ‘object’.

We find that the absolutive reciprocal anaphor elkar ‘each other’ in Basque is possible only if it happens to function as O (as shown in 14), but not if it happens to be as S (as shown in ungrammatical sentences 15) (J. Ortiz de Urbina 1989:21):

14. Lagun-ek elkar ikusi z-u-te-n

‘The friends saw each other’.

15.*Elkar joan zen/ziren lagun-ekin

* ‘Each other went with the friends’

The restriction is not, then, on absolutive anaphors, since O and S, both absolutive, behave differently; in fact it is on subjects, since intransitive subjects (Ss) share their behaviour with ergative transitive subjects (As) as shown in the ungrammatical sentence 16 below (J. Ortiz de Urbina 1989:21):

16. *Elkarr-ek ikusi zituzten lagun-ak

*Each other saw the friends.

Therefore by showing that the absolutive anaphors are possible only if they correspond to the object of a clause where subject anaphors are always impossible we state the important generalization that the grammatical function of object has its properties distinct from the properties of the subject.

The investigation of the morpho-syntactic properties of the object functions in Basque reveals that this language distinguishes between direct object (DO) and indirect object (IO). We have considered the syntactic properties of direct object in part I, so now we are turning to the morpho-syntactic properties of indirect objects (IO). First we should provide the properties of indirect object that would distinguish it from second object (O2) function.

There are two properties that distinguish the grammatical function of indirect object in Basque: thematic role of IO and the ability to appear alone without the direct object. Let’s consider these properties in detail.

• According to the definition provided by A. Andrew the second object is a grammatical relation in double object constructions that is usually expressed by the thematic role of Theme. However in Basque the grammatical function expressed by the indirect object usually has the thematic role of Recepient or Goal but not Theme that is usually expressed by direct object.
• Based on the definition of O2 we conclude that O2 is restricted to appear only with direct object but it cannot appear alone for example with intransitive verbs. Unlike O2 the indirect oject can appear with intransitive verbs.

These two properties are illustrated in examples 17- 20 below:

(J.I.Hualde & J.Ortiz de Urbina 1993: 27)

17. ni-k Jon-i liburu-a-Æ ema-n d-ieza-io-ke-t

‘I can give the book to John’

(J.I.Hualde & J.Ortiz de Urbina 1993: 73)

18. Hi-Æ ni-ri etorri h-atzai-t

‘You have come to me’

(J.Ortiz de Urbina 1989:9)

19. Joan-go g-in-tza-izk-i-e-n

‘We would go to them’.

20. Hurbil n-a-ki-o-ke

‘I can approach him’.

As we see from the data the indirect object can be either in the form of a case marked nominal cross referenced with the inflection on the verb (as in 17 and 18) or the grammatical function (in case of pronouns) can be simply expressed by the verb inflection (as in 19 and 20).

But in any case the grammatical function has the thematic role of Recepient (17) or Goal (18-20) and it can appear alone with intransitive verbs (as in 18-20). Therefore, we prove that we are dealing with indirect object grammatical function not second object.

There are some other interesting tests revealing the coding properties of IOs that we are going to consider below.

2. Case marking and cross referencing

As we showed in part I indirect objects in Basque are marked by the dative case suffix –(r)i attaching to the arguments performing this grammatical function. In addition, verbal inflection in Basque obligatorily agrees with the indirect object argument the same way as it does with other core grammatical functions (A, S and O). We have shown in the paradigm in part I point 3 that dative verb inflection is suffixed to the form of the verb cross-referencing the dative-marked arguments. The two properties of IOs (case marking and cross referencing) can be illustrated in the following example (J.I.Hualde & J.Ortiz de Urbina 1993:73):

17. ni-k Jon-i liburu-a-Æ ema-n d-ieza-io-ke-t

I-Erg John-Dat book-det-Abs give-asp 3sg.Abs-aux-3sg.Dat-mod-1sg.Erg

‘I can give the book to John’

So as we see the case marking and cross referencing on the one hand distinguish indirect object from direct object function and on the other hand distinguish its properties from the nominals having an oblique function.

3. Relativization

In Basque subjects, objects and indirect objects can be relativized by deletion while nominals having an oblique function don’t have this property. This can be illustrated in the following examples (Comrie 1981: 134-135):

18. Gizon-a-k emakume-a-ri liburu-a-Æ ema-n d-i-o-Æ

‘The man has given the book to the woman’

19. [Gizon-a-k liburu-a-Æ ema-n d-i-o-Æ -n ] emakume-a-ri

‘The man to whom the woman has given the book’

So as we see from the data Basque has no movement in the formation of relative clauses, but simply deletes the relativized element in the relative clause adjoining a special relativizing marker to the verb of the clause with the deleted element. Though the range of nominals that can be relativized in Basque is highly restricted, in particular only core functions (subject, DO and IO) have this syntactic property and so relativization could serve as a test for distinguishing Indirect objects from Obliques.

1, 2, 3 – person marker

sg. - singular number

pl. - plural number

Ap - absol. pluralizer

prs - present tense

pst - past tense

prf - perfective

pot - potential

fut - future

asp - aspect

aux. - auxiliary

mod. - modal auxiliary

neg. - negative particle

comp - complementizer

rel.m. - relative marker

Erg - ergative

Abs - absolutive

Dat - dative

dat - pre-dative marker

recip. - reciprocal

1. J. I. Hualde & J.Ortiz de Urbina (eds.) "Generative Studies in Basque Linguistics". Current Issues in Linguistic Theory, 105;1993.
2. J.Ortiz de Urbina "Parameters in the Grammar of Basque". Studies in Generative Grammar, 33; 1989.
3. M, Saltarelli "Basque"; Croom Helm, 1988.
4. B.Comrie "Language Universals and Linguistic Typology". Basil Blackwell, 1981.

Back to languages table of contents

Back to topics

Revised February 19, 1999