argument structure of k'u 'give'

Thu Nov 18 22:23:25 UTC 2004

[John]
> It's certainly typical of Dhegiha to mark one kind of object or the

> other, not both, and I had always thought the same was true also of

> Dakota, barring a few marginal examples in either language.

> Rephrasing this, the pattern I thought I was seeing in Siouan

> languages is that their verbs code one object.  Non-dative verbs -

> "simple transitives" - code a patient and no uncoded dative object

> nominal can be mentioned in the clause.  Some kinds of more remote

> objects can be introduced with locative prefixes.  Dative verbs code

> a dative and an uncoded patient object nominal can be included in the

> clause.  Most dative verbs involve some specific marker of

> dativeness, which is usually manifested to some degree in a

>  modification of the form of explicit patient pronominals.

Today's Lakota field session was the Lakota argument structure lover's paradise (or madhouse, depending on whether one judges it by its output or by the energy input required to keep one's mental focus in the face of too many person markers showing up in too many unexpected positions).

First, a note on wicha-ma-k'u-pi, which I had previously glossed as 'I was given to them' or 'they gave me to them'. This example does, in fact, not prove the possibility of double object coding in transitives as clearly as I thought. My Pine Ridge speaker confirmed its grammaticality, but spontaneously provided the translation 'I was given away (in marriage)'. And I believe that this translation fits the context in which the example occurs (cf. my previous email) better than my original translation, because the extra-linguistic referent of the benefactive wicha- is not mentioned at all in the context in which the form wicha-ma-k'u-pi occurs. Which points to the possibility that wicha-, in this case, is not a referential third person marker, but rather, a marker for impersonal, i.e. the animate counterpart of wa- 'non-specific patient'. I have numerous examples suggesting that wicha- can be used that way. I know that some people on the list might feel inclined to interpret
 non-referential/impersonal wicha- as a valence-reducing device, rather than as a full-fledged person marker, and on this interpretation, we lose wicha- as a second object marker in wicha-ma-k'u-pi.

More importantly, in trying to elicit a Linda-style example with wicha-ma-k'u-pi, I was told that only

s^uNh^pala ki       hena   ma-k'u-pi

puppy          DEF  those  1SG.PAT-give-PL

'they gave me those puppies'

is grammatical, not, however

* s^uNh^pala ki       hena    wicha-ma-k'u-pi

puppy             DEF  those  3PL.PAT-1SG.PAT-give-PL

'they gave me those puppies'

So right now, it doesn't look like a strong case can be made for double objects with k'u, after all. I also tried *ni-ma-k'u-pi and *ma-ni-k'upi for 'I was given to you in marriage/you were given to me�', but these combinations are not acceptable. But if the wicha-slot in the original example cannot be filled by anyhting other than wicha-, chances are that this wicha- is not a real person marker here.

[Marianne]

> the benefactive prefix is a benefactive applicative, a derivational

> prefix that alters the argument structure of the verb stem rather

> than representing an argument directly. 'Give' typically doesn't

> appear with benefactive applicatives, because it already has a

> beneficiary in its argument structure.

Yes, that would take us closer to answering the question of why k'u 'give' seems to be the only verb that codes an (Enlgish) benefactive by means of patient affixes in Lakota. A potentially related explanation (albeit one that may be taken to merely reflect my personal obsession with Zipfian frequency approaches) is that syntactic combinations that occur frequently in discourse will be structurally less complex/less marked than less frequent ones. Since expression of the beneficiary with k'u is enormously salient in discourse, i.e. extremely frequent, with k'u, the beneficiary might be expressed by a form that is structurally less complex than "normal" benefactive forms, which contain ki-. That is, the ki-"extension" is lacking with k'u.

So far, so good. Returning to the question about double object affixes in Lakota, verbs of transportation -- e.g. a'u 'bring', ahi 'take to' -- turn out to be a real treasure-trove when it comes to checking on cooccurrences of patient and benefactive markers within a single verb form. At least one relevant example was in Jan's recent post already. But there's more:

s^uNh^pala  ki        hena    wicha-mic-a'u

puppy           DEF   those   3PL.PAT-1SG.BEN-bring

'he brings the puppies for me'

s^uNh^pala  ki        hena    wicha-nic-a'u

puppy           DEF   those   3PL.PAT-2SG.BEN-bring

'he brings the puppies for you'

s^uNkakhaN  ki      hena    wicha-mic-ahi

horse              DEF  those  3PL.PAT-1SG.BEN-take to

'he takes the horses there for me'

s^uNkakhaN  ki      hena    wicha-nic-ahi

horse            DEF  those  3PL.PAT-2SG.BEN-take to

'he takes the horses there for you'

Note that in all these cases, skipping wicha- also yields acceptable examples. A phonological complication in dealing with these forms lies in the fact that apparently, the final i of the benefactive markers mici-/nici- is assimilated to the adjacent verb-initial vowel a, i.e. it disappears.

So given these examples, the one-affixal-object-only hypothesis can't be defended, I guess. But in fact, this hypothesis captures a strong statistical tednency because examples like the ones above are extremely infrequent in natural discourse. One factor that is quite likely to obscure the possibility of having double object affixes is the fact that at least many third person objects remain cryptic because they are coded by zero markers -- unless we're dealing with third person animate plural wicha-.

Another question worth asking is if the semantic combination benefactive + patient is the only one that produces double object affix constellations in a single verb. The answer is, no, the referents of postpositional phrases may also appear as verbal affixes together with a benefactive affix. Consider:

ob     wicha-nici-yiN-kte

with  3PL.PAT-2SG.BEN-go-FUT

'he'll go with them for you'

ekta   wicha-nici-yiN-kte

to      3PL.PAT-2SG.BEN-go-FUT

'he'll go to them for you'

Regina

---------------------------------
Do you Yahoo!?
 Discover all that�s new in My Yahoo!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/siouan/attachments/20041118/cc290dc1/attachment.html>