Relativization in Basque

David Wimmer

3/22/13

LING 304: Research Report #3

Basque: A language isolate spoken in Northeastern Spain and Southwestern France

1. In his grammar of Basque, Hualde posits gapping as the strategy that the language uses to construct its relative clauses. In this method, Basque exhibits neither the relativized NP nor any corresponding pronominal form present within the relative clause, as shown in (1):

1. [Pello -k ekarri due -n] dirua galdu dut.

Peter -ERG bring AUX -COMP money.DET lose AUX.1SG

'I lost the money Peter brought.' (764)

As is typical of SOV languages, the relative clause precedes the head NP. The absolutive is relativized on here, which is reflected here by the " gap and corresponds to "money." While the gapping in this strategy is complete, verb agreement can in some cases provide indirect traces of the relativized NP:

2. merkatuan erosi dituzu -n sagarrak

market.LOC buy AUX.PL -COMP apple.DET.PL

'The apples that you bought in the market.'

3. sagarrak erosi duzu -n merkatua

apple.DET.PL buy AUX -COMP market.DET

'The market where you bought the apples.' (765)

In (2), the relativized NP is absolutive, and thus the auxiliary verb in the relative clause agrees in number with it through the infix (-it-). In (3), however, the relativized NP is a locative oblique, with which the verb shows no such number agreement. These agreement markings could be argued to contribute to the greater ease in accessibility shown both by absolutives and ergatives, which also show the same agreement pattern.

Hualde posits an invariant suffix as a complementizer that attaches itself to the inflected verb within the relative clause, which in this case is the auxiliary. He claims that this complementizer "relativizes the relative clause," thus serving as a marker that said verb and its corresponding argument are bounded as such (764). No evidence is presented by the author to support this claim; however, it must be noted that the complementizer does appear within every relative clause example given in this grammar (814). While this is insufficient evidence in and of itself to posit a category such as "REL" or "COMP," it is quite likely that this suffix is critical to the formation of relative clauses (and perhaps other syntactic structures, too). Further work is needed to determine exactly how it functions.

2. Relativization in Basque is defined by the language's ergative-absolutive nature. Hualde astutely recognizes this, and only deals with grammatical relations in these terms For instance, he treats absolutive properly as an S & P category and does not posit an erroneous "subject" category. Thus, in a strict sense, Hualde never contends that "subject" is the most accessible GR. If we take absolutives to be the most "subject-like" category, however, Hualde still at the very least modifies Keenan's assertion, to use the author's own words (775). He argues that absolutives and ergatives are both equally accessible in Basque, thus creating a hybrid "subject-direct object" category on the accessibility hierarchy. Hualde supports his claim by arguing that both ergatives and absolutives can always be relativized without any restriction. Datives and obliques, on the other hand, can require additional case-marking or be ungrammatical to certain speakers or, in some cases, to all speakers. For instance, the oblique relativized on in (6) must also agree with a relational suffix, "-ko," placed on the auxiliary verb, otherwise it would be ungrammatical.

I find this argument, and the relative clause evidence he uses to support it, convincing. I see no reason why absolutives should be treated as more accessible as ergatives other than to simply conform to the accessibility hierarchy itself. I would further add, as I mentioned above, that ergatives and absolutives have more transparent case-marking within the relative clause, thus likely making their construction less cognitively taxing. In the below questions I will go over specific evidence regarding GRs and accessibility.

3. The gapping relativization strategy can apply to absolutives (1-2), ergatives (5), datives (4), obliques (3,6). Genitives, as in (7), cannot strictly be relativized; however, certain constructions can be semantically understood as genitive within relative clauses though they employ different grammatical relations. Additionally, Hualde notes that in certain cases, such as the oblique shown in (6), certain speakers show reluctance to productively produce such constructions.

4. burni hotsa darion Bizkaiko kizkeraz

iron sound pour.3ABS.COMP Biscay.REL dialect.INSTR

'in the dialect of Biscay, which pours out an iron sound' (776)

(Note that in this dialect, Hualde posits that the verb 'darion' is ditransitive, with 'hotsa' and 'Bizkaiko' serving as its two objects).

5. hor dauzkagu Bonaparteren lanari ekin zioten langile

there have Bonaparte.GEN work.DAT tackle AUX.3E.PL.COMP worker

'There we have the workers who took on the work of Bonaparte.' (775)

6. usoak iragaiten direneko haroa zen

doves pass.IMPF AUX.COMP.REL season be.3SG.PST

'It was the season when pigeons arrive.' (782)

7. *etxea argazkhian hartu dudan gizona da

house photograph take AUX.COMP.GEN man be.3SG

'It is the man whose house I took a photo of.' (780)

8. hainbetse aldiz kontra aritu nintzen arerioek lagundu ninduten

so.many time.INSTR against act AUX.COMP enemies.ERG help AUX

'He whose adversaries fought against me so many times helped me.' (780)

4. I believe that Basque relativization conforms to the accessibility hierarchy, aside from the point Hualde raises about the equal accessibility of ergatives and absolutives. As examples 1-6 demonstrate, Basque can relativize items as low on the hierarchy as obliques, and continues the process through ergative/absolutive. Genitives cannot be relativized strictly, as shown in (7), but rather must be understood through context, as in (8). I could find no evidence in Hualde's grammar, explicitly stated or otherwise, that objects of comparison exist as a separate grammatical relation.

5. The same evidence also shows that Basque relativization applies to a contiguous portion of the accessibility hierarchy. Obliques are lowest on the list of Basque items relativized, followed by datives, ergatives and absolutives. Genitives cannot be directly relativized and objects of comparison do not appear to exist in as a separate grammatical relation.

6. The form of predicates within relative clauses differs from independent finite sentences. In nominal predication, such as (9), Basque uses one of two copulas, "egon" or "izan" to form the construction. In predicates within relative clauses, such as (10), the normal forms of the verb stem "da" are used.

9. Mikel urduri izan

Michael nervous COP

'Michael is nervous.'

10. naizen gizonak nekez onets dezake horrelakorik

be.1SG.COMP man.ERG hardly accept AUX.POT such.thing.PRT

'The man I am can hardly accept things like that.'

I'm puzzled, actually, why this difference wouldn't lead to the construction in (10) to be disqualified as a predicate nominal. Perhaps (10) is in fact a copula, but no glosses in the relativization section indicate that it is such (they're all glossed auxiliary). Regardless, further work is needed to clarify this discrepancy.

Gloss Abbreviations

# - Cannot be interpreted as the specified meaning

1 - 1^st Person

2 - 2^nd Person

3 - 3^rd Person

ABS - Absolutive

ABL - Ablative

ALL - Allative

AUX - Auxiliary

COMP - Complementizer

COP - Copula

DAT - Dative

DET - Determiner

EMP - Emphatic

ERG - Ergative

FUT - Future

HAB - Habitual

IMPF - Imperfect

INSTR - Instrumental

LOC - Locative

REL - Relational

PL - Plural

PRF - Perfective

PRT - Preterite

PST - Past

POT - Potential

SG - Singular

References

Hualde, JosIgnacio, and Jon Ortiz de Urbina. 2003. A Grammar of Basque. Berlin and New York: Mouton de Gruyter.