C'ali: antipassivization, "passivization", more on pivot alignment

Some weeks ago I posted some comments about the C'ali split-S
pivot, and the question was raised how it deals with that pivot.
C'ali has grammaticalized considerably fewer grammatical categories
into the verbal head, and so to express some categories it uses
periphrastic constructions.  One voice construction it has
grammaticalized is the antipassive, which demotes the patient
of transitive verbs to create an agentive intransitive. This is
probably better described as a basically derivational, not an
inflectional, process, since the presence of an antipassive
morpheme does not necessarily imply the presence of a corresponding
transitive verb. There are three primary such derivational morphemes:
/-uo/, /-(e)hui/ and /-sta/:

(1)
  suota-     'break X'
  suotatani  'I will break it'
  suotauota  'I am going about breaking (things)'

(2)
  tal-       'look at X'
  taltani    'I am looking at it'
  talehuita  'I am looking about (at things)'

(3)
  eskha-     'lay waste'
  eskhatani  'I am laying it waste'
  eskhastata 'I am going about laying (things) waste'

In addition to these three major affixes, there are about 25 other
affixes, most of which are not productive and instantiated by only a
handful of lexemes at most.  There are certain semantic tendencies of
some classes; /uo/, for example, often derives imperfective agentive
intransitives from perfective transitives. That this is not always
the case is clear from the root _eskha-_, which though perfective
takes a different antipassive suffix. Thus, antipassives must simply
be learnt along side their transitive counterparts.

"PASSIVES"

"Passives", on the other hand, are periphrastic constructions.  Such
constructions technically take the form of a matrix verb with a
subordinate clause fully inflected for person, number and gender of
both agent and patient. I use scare-quotes because this construction
is a little ambiguous, having optional omission of the agent but verbal
cross-reference with that agent. Discourse-functionally, however, they
have many of the same features of passives, including topicalization,
pivot alignment, and referencing politeness, so it is useful to treat
them along with antipassivization.

The choice of which matrix verb depends the semantic features of
the passivized verb:  causative-statives take a matrix verb _lotw-_*
"to be". *Except* in passive constructions, this verb always takes class
two (agentive intransitive) verbal morphology. All other fields
of verbs take _c'anes-_ "undergo, suffer", a class three (patientive
intransitive) verb. Importantly, in the standard language verbal
agreement and case assignment in the matrix clause follows the
prescriptions of the auxilliary verb:

(4) kxait-si  saxmë-n  lotu-xela   me-q'æ?-tle-n-xela
    face-PAT4 man-DAT3 be-3SgP.III 3SgD.I-be.painted-CAUS-3SgA.I-3SgP.III
    "The man's face is being painted"
    (lit. "The man's face is that one paints to him the face";
    the dative marker on the verb

(5) saxmë-tsi c'anes-këi    tal-essa-këi           olmaqa
    man-PAT3  suffer-3SgP.I look.at-3SgA.II-3SgP.I woman-AGT1
    "The man is being looked at by the woman"
    (lit. "The man suffers that the woman looks at him")

Another unusual feature about passivization is that passivized arguments
in the matrix clause come *before*, not after, the matrix verb, contrary
to the general VSO pattern of the language as a whole. This may be an
example of grammaticalization of the structural requirements of one of
the passive's uses, topicalization.  In some dialects, this kind of
construction may be a full passive, since the agent marking on the lower
verb may disagree with the gender, number and person features of the
agent:

(6) saxmë-tsi c'anes-këi    tal-an-këi           olma-qa
    man-PAT3  suffer-3SgP.I look.at-3SgA.I-3SgP.I woman-AGT1
    "The man is being looked at by the woman"
    (lit. "The man suffers that one looks at him by the woman")

(Here, the lower verb is marked for 3rd singular agent of the first
gender, not the second as required by _olma_) In these cases, this
disagreeing marker is always -(a)n-, homophonous with the 3rd singular
gender I agent, which may mean it has simply become a passive marker
for these speakers.

SPLIT-S PIVOT

As was mentioned recently on the list, languages with a Split-S pivot
need to be able not just to adjust to a mismatch between the thematic
marking of transitive plus intransitive sentences, but also sentences
where each verb is intransitive, differing only in their thematic marking,
agent or patient.  Like antipassives, this seems to be a derivational
phenomenon.  Each intransitive has a derived counterpart which may be
used to correct a mismatch.  For example, the mismatch in (7) may be fixed
by choosing the agentivized intransitive /tV-?os/:

(7)  *t|-anat-essa         olma-qa    som ?os-këi [X]
     thither-run-AGT.II=DP woman-AGT1 and die-3SgPAT.II
     *"The woman ran off and [she] died"

(8)  t|-anat-essa          olma-qa    som to-?os-essa [X]
     thither-run-AGT.II=DP woman-AGT1 and AGT-die-3SgAGT.II
     "The woman ran off and [she] died off"

Note that the agentive inflectional suffix changes also. Which
intransitive verb is more basic, the agentive or patientive,
depends entirely on the verb and pragmatic factors.  In the case
of "die", for good reason the patientive verb is more basic, but
this has nothing to do with the system as such.  Some other common
prefixes and suffixes that alter the thematic role of the verb
are as follows:

Agentive:
/t[eno-/, /qau-/, /phæsti-/, /-nyöre/, /-k'es/, /qwë-/, /khumna-/

Patientive:
/tš'æi-/, /c'elly-/, /xan-/, /mu?ex-/, /ath'i-/, /tleo-/, /-e?i/

These prefixes and suffixes have a variety of subtlties and are
frequently used to distinguish between different lexical items
even of the same thematic marking.

Anways, that's about all for now.  I hope to post on tense and aspect
sometime soon.

 =========================================================================
Thomas Wier            "I find it useful to meet my subjects personally,
Dept. of Linguistics    because our secret police don't get it right
University of Chicago   half the time." -- octogenarian Sheikh Zayed of
1010 E. 59th Street     Abu Dhabi, to a French reporter.
Chicago, IL 60637