Summary of what went on in the Argument Structure Seminar on February 15, 2005

Michal clarified some issues regarding the morpheme order in causatives, which I had represented as:

Nivkh: inner V-u, outer V-g-u, doubling V-u-g-u
Hindi: inner V-aa, outer V-v-aa
Kitharaka: inner V-i, outer V-ith-i
Amharic: inner a-V, outer a-s-V
Malagasy: inner an-V, outer an-f-V, doubling an-f-an-V

Let's descriptively call the little middle bit a "connector" for the moment. I had been assuming that the connector was g in Nivkh, v in Hindi, s in Amharic, f in Malagasy, and ith in Kitharaka.

Michal pointed out that Kitharaka has some properties that distinguish it from what we know about

if you concentrate on the doubling patterns, and on the fact that the outer causative is never separated from the connector, then you can represent Nivkh and Malagasy like this:

Malagasy: an-V, anf-an-V
Nivkh: V-u, V-u-gu

Now, compare Kitharaka:

Kitharaka: V-i, V-ith-i

If we number the "big" part 1, the "small" part 2, and the verb V, then we see this pattern:

Malagasy: 2-V, 1-2-V
Nivkh: V-2, V-2-1
Kitharaka: V-2, V-1-2

This is similar to the pattern made famous from Greenberg's generalization 20; in noun phrases you find

Dem-Adj-Num-N
N-Dem-Adj-Num
N-Num-Adj-Dem

As Michal pointed out, Cinque has suggested that this is more general: to the left of a head, you see the base order, to the right of a head, you see the base order, its reverse, or some combination of the two.

So the suggestion was that Malagasy represents the base order, Nivkh the reverse or roll-up order, and Kitharaka involves V crossing the base-ordered affixes 1-2. As we saw at the beginning of last semester, Hyman has evidence from consonant mutation that could be taken to support this: even when other morphemes intervene, the morpheme numbered "2" above triggers consonant mutation in the stem, in some Bantu languages. This could suggest that it was adjacent at some level of representation:

1-2-stem
1-mutated.stem-2
mutated.stem-1-2

Interestingly, morpheme 2 also triggers mutation in the stem in Nivkh, but morpheme 1 doesn't, also consistent with Michal's interpretation of the facts.

After that, Gillian presented the Hindi/Urdu pattern in detail, recapping and building on her presentations from last semester. Unfortunately, I don’t have time here to do justice to the entire analysis, which was carefully constructed in accord with her latest developments in First Phase Syntax; in brief, both the /-aa/ and the /-v/ suffixes are verbal heads, in that they are components of a decomposed verb; however they each have a strict place in the fseq, the hierarchical functional sequence.

The /-aa/ is associated with initiation (formerly "v"), and the /-v/ with process (formerly "V"). The presence of /-v/ in the structure leaves the verbal root only one possible place to be realized, namely below it. Rules governing subevent combination ensure the difference between direct and indirect causation. An important part of the story is that the argument structure ordinarily associated with a verb is present even if the verb doesn't combine with unfilled heads for realizing those arguments: this phenomenon is called "underassociation," as the argument slots that the verb bears (in the form of features init and proc, corresponding roughly to Agent and Patient) are prevented from associating with the syntactic nodes (init and proc, corresponding roughly to v and V). This leads to the arguments being either implicit or expressed as adjuncts, as we have seen is common for causees.

As I anticipated, this doesn't do justice to her presentation, but fortunately she has a manuscript which she will be ready to share with you all soon so you can read the analysis in her own words.

What I would like to do next time is to present an updated version of the two-domains analysis that I presented at the Argument Structure workshop; it will be updated to take into consideration Patrycja’s and Michal’s and Gillian’s observations from the last few sessions, and restructured to focus on the generalizations that hold up. Again, I apologize for the delay; I had hoped to present today before getting on the plane for Rome, but there was a local disruption in the spacetime fabric late last week and the ensuing warp sent ripples forward, resulting in a shrinking of the time available to me to prepare. To offset the inevitable recriminations and to dampen the din of the hand-wringing I attach a document for your perusal which may inspire you to think about this seminar as one that will lead to a series of papers (picture your name on one of them).

Argument Structure Newsflash: Hagit Borer's long-awaited Exo-Skeletal volumes are said to have finally been issued by Oxford, with the date 2005. I have spoken with somebody who knows somebody who has seen one.

na Slavny Boj, na Smjertny Boj!
Peter