comunicações sobre línguas amazônicas apresentadas no congresso da LSA e SSILA

[diinibii]

Caros colegas,

Lev Michael, doutorando na Universidade de Texas, fez algumas 
anotações sobre comunicações recentes relacionadas às línguas 
amazônicas, no congresso da LSA e SSILA.  Ele concordou em 
compartilhá-las com os membros da Etnolingüística.

Dear Friends,


What follows are brief summaries of talks given on Amazonian 
languages at the
2006 Albuquerque LSA & SSILA meetings. I have written it for 
colleagues who
were unable to make it to the meetings.  I have construed `Amazonia' 
quite
broadly to mean the Amazon Basin proper and (roughly) adjacent 
neotropical
lowland(ish) areas. Apologies in advance to the presenters for any 
inadvertent
misrepresentations on my part. I have included the email addresses 
for the
presenters when available.


Please feel free to circulate this message to anyone who you think 
might be
interested.


Lev Michael
UT Austin


*********


Anderson, Cynthia, Christine Beier, IWen Lai, and Lev Michael. SVO 
versus SOV
constituent order in Iquito (Zaparoan): A post-syntactic explanation.
cianderson at mail.utexas.edu, cmbeier at mail.utexas.edu, 
iwenlai at mail.utexas.edu,
lmichael at mail.utexas.edu


The authors began by presenting evidence that suggest there are two 
constituent
orders in Iquito: SVO in realis mood clauses and SOV order in 
irrealis mood
clauses. However, they then showed that in irrealis clauses, a 
variety of
element other than object can appear in the position between the 
subject and
verb, including adverbs, negation, determiners, and post-positions. 
It is
ungrammatical for any of these elements to appear in this position 
in irrealis
clauses. SOV order, they argued, is simply a special case of a more 
general
irrealis order where a variety of elements can intervene between the 
subject
and verb.


They then showed that the element that intervenes between the 
subject and verb
in irrealis clauses occupies the immediately post-verbal position in 
realis
clauses. They then argued that irrealis order is derived by the 
dislocation of
post-verbal material to the immediately pre-verbal position. 
Evidence they
presented for this analysis included the fact that irrealis order 
results in
discontinuous constituents, with the parts of the constituent 
straddling the
verb, and the fact that realis clauses are unmarked, and hence, 
presumably,
basic.


They then argued that the irrealis dislocation process could not be 
considered a
syntactic process, in the strict sense of the term. Arguing that 
syntactic
processes operate on both syntactic natural classes of elements and 
on
syntactic constituents, they pointed out that the elements 
dislocated to the
pre-verbal irrealis position do not form any syntactic natural 
class, and
moreover, some of the material dislocated to that position is not a 
proper
syntactic constituent (e.g. postpositions with fragments of their 
complements).
>>From this, they concluded that the irrealis dislocation process 
must be a
postsyntactic one.


The authors then proposed an analysis for the irrealis dislocation 
process. They
first observed that irrealis order is in a one-to-one relationship 
with irrealis
mood, and that there are no surface-segmentable morphemes that 
correspond to
irrealis mood. In other words, irrealis order is the sole surface 
realization
of irrealis mood, and it does not correlate with any other 
grammatical feature.
On this basis, they proposed that irrealis mood in Iquito is marked 
at the
syntactic level by a pre-verbal morpheme that lacks phonological 
material. They
further proposed that irrealis dislocation is a consequence of a 
requirement for
morphemes to have an overt phonological realization, and that 
irrealis
dislocation meets this requirement post-syntactically by moving 
phonological
material, irrespective any syntactic characteristics of that 
material, to the
position of the irrealis morpheme.


***


Brauschuber, Brianna. Degenerate feet and minimum word requirements 
in Iquitu.
brauschuber at mail.utexas.edu


Brauschuber showed that the Iquito stress system is mostly quite
straightforward, but that it also shows some typologically 
interesting
phenomena . Iquito assigns right to left moraic trochees (i.e. 
prosodic feet
are either two light syllables with left-most stress or a single 
heavy
syllable), and in the general case, forbids degenerate feet (in this 
system,
monomoraic feet). Primary stress falls on the rightmost stressed 
syllable.
There is also a minimum word size of two moras.


The interesting phenomena is found in certain three and two syllable 
words, in
that they permit degenerate feet in the leftmost position. These 
are: two
syllable words of the form (L)(H) and three syllable forms of the 
form (L)(LL)
and (L)(HL). These are the only degenerate feet permitted in Iquito. 
Rauschuber
notes that the effect of these degenerate feet is to produce words 
with two
feet, and that degenerate feet are not found in words that already 
possess two
feet without recourse to degenerate feet (e.g. L(H)(H) or L(LL)(H)). 
Rauschuber
argues that this suggests that there is a requirement in Iquito that 
prosodic
words consist of two feet, which she captures in a BinWd constraint 
(in analogy
with BinFt).


One major empirical issue raised by this line of reasoning is the 
absence of
forms like (L)(L) and (H)(L), which actually surface as (LL) and 
(HL),
respectively. Rauschuber argues that forms like (L)(L) and (H)(L) are
unacceptable because they would require primary stress to fall on a 
light
syllable. In the spirit of the peak prominence constraints, which 
result in
stress to preferring to fall on `prominent' constituents (e.g. 
prosodic heads,
or sonorous segments), Rauschuber proposes that primary stress 
rigorously
avoids degenerate feet in Iquito, forcing a violation of WdBin.


Rauschuber presented an OT analysis which predicted the empirical 
patterns
observed. From a typological perspective, what is most interesting 
about the
Iquito stress system is that it only permits degenerate feet in 
words of very
limited sizes, and analytically, that this is a consequence not of 
the ranking
of Parse-syl>>FtBin.


***


Danielson, Swintha. 2006. Person cross-reference clitics in Baure 
(Arawak).
Swintha.Danielson at mpi.nl *and* swintha at hotmail.com


Danielson began by noting that in most of the literature on Arawak 
languages
person cross-reference markers are referred to as affixes, but that 
the
evidence indicates that Baure person markers are not affixes, but 
instead
clitics.


Baure person markers can appear on verbs as either A or S 
proclitics, or O
enclitics. They also occur as proclitics possessive markers on 
nouns, and as
enclitics on non-verbal predicates. In addition, they appear as 
enclitics on
interrogative particles.


Danielson evaluated arguments both in favor of Baure person markers 
being
clitics and those against them being clitics. The arguments 
Danielson presented
in favor of Baure person markers being clitics included: 1) The are 
unstressed
and always attach to a host, 2) they attach to a wide variety of 
word classes,
and the position of the clitic depends on the word class of the 
host. Baure
affixes, on the other hand, have fixed positions, with little 
variation in the
word class of the stem. 3) person markers may either enclitize to 
the preceding
word, or procliticize to the following one. 4) person markers are 
external to
inflectional morphology. 5) There is a pausal form /ha/, which does 
not
intervene between a stem and an affix, but which can intervene 
between a
person-marker and the verb or noun.


The arguments that Danielson presented against the Baure person 
markers being
clitics included: 1) person proclitics are obligatory, and hence 
seem more like
agreement marking (but enclitics are not, and alternate with full 
NPs) . 2)
person markers and affixes share certain phonological processes, 
such as /o/
deletion in syllable-final and word-final positions. 3) There are 
other
phonological processes that show strong phonological interaction 
between the
host and person marker.


Danielson concludes that the arguments in favor of evaluating these 
person
markers as clitics is stronger than those for evaluating them as 
affixes.
Danielson proposed that it is necessary to view clitics and affixes 
as part of
a continuum between affixes and free particles.


***


Dickinson, Connie, Simeon Floyd and Jenny Seeg. Evidentiality and 
mirativity in
Cha'palaa and Tsafiki.


This talk presented a discussion of evidentiality, epistemic 
modality, and
mirativity in Cha'palaa and Tsafiki, two Barbacoan languages of 
lowland
Ecuador. There were two principal themes: 1) that in these two 
languages,
evidentiality, epistemic modality, and mirativity form distinct 
morphosyntactic
paradigms, and 2) that in these two languages, evidential marking 
also codes
`participation' (about which more below).


Both languages make four evidential distinctions: direct evidential, 
inferred
evidential, deduced from general knowledge, and reportative. In 
Cha'palaa, the
reportative is not fully grammaticalized (unlike Tsafiki), but it is
discursively obligatory. Evidential markers are in free distribution 
with mood
markers, of which there are six (in both languages, I infer): 
declarative,
interrogative, dubitative, speculative, declarative emphatic, and 
imperative. 
The sole exception to the free distribution generalization is with 
the
imperative mood marker, which apparently does not co-occur with 
evidentials.
Note that evidentials do co-occur with interrogative mood marking, 
in which
case the speaker is understood to be asking about the interlocutor's 
source of
information.


Both languages possess a mirative marking system that marks a binary
conjunct/disjunct distinction. A sentence is mirative-marked only if 
the source
of information for the proposition is a primary participant in the 
event
(`locutor situations'). In both languages, the conjunct/disjunct 
marking occurs
in declarative sentences with first person subject. Parallel 
sentences with
second and third person subjects do not take mirative marking. In 
interrogative
sentences, those with a first person subject take disjunct marking, 
and those
with a second person subject take conjunct marking. Third person 
remains
unmarked.


[At this point the presenter zipped through the rest of the 
discussion of
mirative marking for reasons of time. I am embarrassed to say that I 
had a very
hard time following this, so I will spare you most of my (probably 
garbled)
understanding. The basic idea, however, appears to be that conjunct 
marking
indicates congruence of the knowledge contained in the predicate 
relative to
the knowledge base of the subject of the verb, whereas disjunct 
marking
indicates incongruence.]


In any event, evidentials and mirative markers can co-occur on the 
verb.


The remainder of the talk concerned pragmatically marked uses of 
evidentials,
such as the use of the indirect evidential although the speaker 
directly
experienced an event, which indexes the fact that speaker was not
psychologically prepared for what they witnessed. Another use of 
this type
involves the use of the direct evidential when the person did not 
directly
experience the event in question. This is apparently permissible in 
certain
circumstances when the speaker is a member of the same household as 
the person
carrying out the action indicated by the direct evidential-marked 
verb. In
addition, mirative markers can be used in non-canonical ways to 
distance
oneself from the events being described, or for irony. Apparently 
mirative
markers can also be used to indicate deference by indicating 
distance with
respect to some domain of knowledge in comparison with the 
interlocutor


***


Guillaume, Antoine. More on the typology of inverse systems: the 
Reyesano suffix
-ta


Guillaume's talk concerned the relationship between a morpheme -ta, 
and person
marking on the verb. Superficially, the correlation between person 
marking and
the presence or absence of this morpheme makes it appear that 
Reyesano has
something like an inverse system (as found in the Algonquian 
languages).


Reyesano has a single verbal prefix slot for marking person, but 
this position
does not consistently correspond to an argument with a particular 
grammatical
relationship (i.e. it can mark S, A, or P). Instead, the argument 
that is
marked is decided by the following person hierarchy: 2>1>3.


With intransitive verbs, -ta appears with plural subjects; with 
transitive
verbs, -ta appears under particular circumstances when one argument 
is a SAP
(i.e. 1 or 2) and the other argument is 3: when the prefix is the 
subject, -ta
does not appear; when the prefix is the object, -ta does appear. So 
this
vaguely resembles an inverse system.


In the local configuration (both arguments are SAPs), the hierarchy 
tells us
that 2 is always coded, and the GRs have to be inferred from 
context. -ta does
not appear on the verb.


With two 3 arguments, -ta is obligatory.


Guillaume then explained that there is no proximate/obviate marking 
on NPs, and
there are no animacy hierarchy effects relevant to -ta affixation.


This led Guillaume to comment that it looks like that there are 
three different
-tas: plural marker (intransitive verbs), inverse (SAP and 3rd; 
transitive
verbs), Obligatory (3 and 3; transitive verbs). He notes that with 
transitives,
however,  it only occurs with 3 subjects


Guillaume concludes that instead of an inverse system as such, there 
are two
-tas, a plural marker with intransitive verbs, and a third 3 subject 
in
transitive ones. This is a person hierarchy that produces an 
inverse `effect',
but is not an inverse `system',


***


Lai, I-Wen. The realization of sentential negation in Iquito: its 
dependence on
clause type and mood. iwenlai at mail.utexas.edu


Lai's talk examined the structural characteristics of clausal 
negation in Iquito
clauses. Lai identifies two structurally-distinct clausal negation 
strategies.
The first consists of the negation `caa' in the immediately pre-
subject
position, whereas the second consists of caa in the post-verbal 
position,
together with a negation concord morpheme -ji, affixed immediately 
after the
verb root, and before inflectional morphology. Lai refers to the 
former type of
negation as `caa negation', and the latter, `ji-caa negation'.


Caa negation is found in principal declarative  and in yes-no 
questions. Ji-caa
negation is found in Wh-questions, relative clauses and in embedded 
clauses in
cleft constructions. In addition, the latter constructions 
demonstrate a number
of variants with respect to the realization of negation when the 
clause is in
irrealis mood.


As discussed in Anderson, Beier, Lai, and Michael (see above), 
irrealis mood is
marked in Iquito by dislocating post-verbal material to the 
preverbal position,
or by copying it to preverbal position, which results in the 
identical material
in postverbal and preverbal positions. In irrealis clauses with ji-
caa negation
this dislocated or copied material can be caa, frequently resulting 
in caa
appearing in both pre-verbal and post-verbal positions, or simply in 
the
preverbal position. It is also grammatical, in irrealis clauses with 
transitive
verbs, for the object to be dislocated instead of the negation.


Lai offers a formal analysis, based on GB work in negation. This 
work posits two
negation positions to account for the typological variation in 
negation: a
TP-selecting negation head and a VP-selecting head. The TP-selecting 
head
analysis is fairly straightforward: the verb raises to TP to acquire
inflection, and caa appears to the left. The VP-selecting head 
analysis is a
little more complicated. Lai take -ji to occupy the head of the VP-
selecting
Neg, and caa to occupy [Spec, NegP]. The verb raises to T to acquire
inflection, and in raising through NegP acquires -ji as an affix. 
Caa, being in
[Spec, NegP], is found to the right of the verb after the verb 
raised to T.


Lai argues that the TP-selecting Neg is default, and that the VP-
selecting Neg
is triggered by feature in C (Note the cooccurence of ji-caa 
negation with
clause-types in which CP is filled).


>>From a typological perspective, Iquito is interesting because 
although previous
work has argued for the necessity of TP- and VP-selecting heads to 
account for
the cross-linguistic variation in negation, Iquito appears to be the 
first
documented case of a language that exhibits both TP- and VP-
selecting Neg heads
in a single language.


***


Seifart, Frank. An unusual typological mode for handling 
countability in Miraña
(north west Amazon). frank.seifart at berlin.de


Miraña possesses a large set of shape-denoting noun classifiers 
(~70) that also
surface pervasively as agreement markers on numbers, adjectives, and 
verbs. The
topic of the talk concerns the fact that most `bare' NPs, i.e. NPs 
without
classifiers, cannot take number morphology (dual or plural). For NPs 
to be able
to take number morphology, the must be `unitized' or individuated, 
by affixing a
classifier. After the classifier has been affixed, the NP can take 
number
morphology. Moreover, NPs can only be unitized if the classifier is 
affixed to
the NP; that is, the unitization cannot be achieved by only affixing 
the
classifier to an independent (non-affixal) number term.


There are some additional interesting wrinkles. There are some nouns 
that also
function as classifiers (e.g. hook). These noun obligatorily appear 
with a
number term to which the same phonological shape (now functioning as 
a
classifier) must be affixed, when the noun is non-singular in 
reference.
Seifart refers to this class of nouns as `repeater nouns'.


Most inanimate nouns are non-count when they appear without a 
classifier, and
they are unitized by classifiers (i.e. they are not repeater nouns).


The situation with animate nouns is more complicated. Most small 
animals are
unitized with a classifier, but curiously, the choice of classifier 
does not
appear to be at all related to any discernable geometrical 
characteristic of
the animal in question. Larger, culturally important animals are 
repeater
nouns.


Human referents are usually unitized with classifiers, but some are 
optionally
unitized. Kin terms are `directly countable', i.e. do not require 
classifiers,
and are not repeater nouns.


Seifart then presented a typological discussion which highlighted 
the unusual
nature of Miraña unitization. First, in numeral classifier 
languages, plural
marking does not appear with unitizing elements in the same phrase, 
although it
does in Miraña and in languages with singulative marking. This 
points to Miraña
unitization being more like singulative marking. Similarly, numeral 
classifiers
are more closely attached to the numeral than to the NP, but 
singulative marking
and Miraña classifiers are affixed to the noun.


On the other hand both numeral classifiers and Miraña classifiers 
occur with a
large set of nouns, which is not the case with singulative marking. 
Similarly,
the choice of classifier is dependent on nominal semantics (shape), 
which is
characteristic of numeral classifier systems, and not of singulative 
systems.


This makes the Miraña unitization system an interesting cross between
singulative and numeral classifier systems.


[Comment: Frank's dissertation, on which his talk was based, won 
this year's
Mary Haas Award. ]


***


Tonhauser, Judith. Paraguayan Guaraní as a tenseless language.
juditht at stanford.edu


Tonhauser began by noting that Paraguayan Guaraní (PG) has 
previously been
described as having tense morphology


Tonhauser showed that past and present temporal reference is 
achieved as a
pragmatic inference using unmarked predicates, and is based on the 
telicity of
the verb, with telic predicates generally yielding perfective/past
interpretations, and atelic predicates generally yielding 
imperfective/present
interpretations.


Most of the talk was devoted to analyzing the morphology previously 
analyzed as
a future tense morpheme, /-ta/.


First, Tonhauser showed that /-ta/ yields a future interpretation in 
contexts of
prediction, expectation, intention, willingness, promising, and in 
embedded
clauses. Then she presented three arguments against /-ta/ being a 
future tense
marker.
First, -ta cannot be used to express future in negated clauses. 
Instead, a
counterfactual marker must be used. If -ta is a future tense marker, 
there is
no obvious reason it should be restricted to *positive* 
eventualities.


Second, a future tense locates eventualities in the future of a 
reference time
(typically the time of speaking, or some contextually given time).
Consequently, if the reference time is in the past, the future tense 
should
assert the realization of the eventuality. However, the use of -ta 
does not
assert realization


Third, if -ta is a future marker, its appearance should not be 
compatible with
expressions of past time reference (e.g. temporal adverbs). However, 
one finds
that in PG, -ta does appear with expressions of past time reference.


These observations lead Tonhauser to conclude that -ta is not a 
future tense
markers, and that, consequently, PG is a tenseless language.


Tonhauser then goes on to propose that -ta is a modal of non-asserted
realization. She explained that this means that -ta blocks the 
entailment of
event realization, and is, consequently, a kind of irrealis 
modality. In
effect, -ta means that the eventuality is not being realized at the 
reference
time. Consequently, future interpretations result as implicatures in 
context.


***


Tonhauser, Judith. Nominal temporal markers on relative and 
complement clauses
in Guaraní. juditht at stanford.edu


[I was unable to make it to this talk, so what follows is just the 
published
abstract]


Based on data collected in recent fieldwork, this talk examines the 
meaning of
the nominal temporal markers of Paraguayan Guaraní. The markers, -
kue and -rã,
appear on noun phrases and affect the temporal interpretation of the 
phrase,
e.g. ko abogado-kue, `this lawyer-KUE' refers to an individual who 
was a lawyer
but is not anymore. Besides (non-)possessive noun phrases, both 
temporal markers
appear on relative clauses, and -kue also occurs on complement 
clauses. I argue
that relative and complement clauses in Guaraní are nominalized, and 
that the
meaning of -rã prevents it from co-occurring with complement clauses.


***
Beier, Christine and Lev Michael. The Iquito Language Documentation 
Project:
Developing a Team-Based Methodology for Language Documentation 
[Poster].
cmbeier at mail.utexas.edu, lmichael at mail.utexas.edu


Beier and Michael presented a poster that described the team-based 
methodology
developed for the Iquito Language Documentation Project, and 
evaluated its
successes and difficulties.


For those interested in the text of the poster, please visit:
http://www.iquito.org/ildp_poster_2006.htm


***


Also at LSA:


There was a reception at the LSA for this year's Bloomfield Book 
Prize, which
was awarded to - you guessed it - a work on an Amazonian language: 
R.M.W.
Dixon. The Jarawara Language of Southern Amazonia. OUP.


[I wasn't able to go, so I don't have anything more interesting to 
report.]


***END**








----------------------------------------------
Colóquios Lingüísticos do Museu Antropológico
Participe! Para maiores informações, visite
http://geocities.com/coloquioslinguisticos
----------------------------------------------

O uso dos recursos do grupo Etnolingüística baseia-se no reconhecimento e aceitação de suas diretrizes. Para conhecê-las, visite http://geocities.com/linguasindigenas/normas  
Links do Yahoo! Grupos

<*> Para visitar o site do seu grupo na web, acesse:
    http://br.groups.yahoo.com/group/etnolinguistica/

<*> Para sair deste grupo, envie um e-mail para:
    etnolinguistica-unsubscribe at yahoogrupos.com.br

<*> O uso que você faz do Yahoo! Grupos está sujeito aos:
    http://br.yahoo.com/info/utos.html