MIT RLE Speech Communication Group Seminar
Carol Fowler
Haskins Laboratories"Imitation in Speech" Infants imitate both facial gestures and speech. Adults do too. The ability to imitate raises a variety of questions that have implications for viable theories of speech perception and production. These are questions about the nature of percepts yielded by the sensory modalities such that cross modal cross talk is possible and so that percepts can guide action planning. In speech, the questions extend to those about how talkers can communicate with listeners. Recently, theorists in different domains have converged on the understanding that cross talk between percepts in different modalities ("supramodal representations"; Meltzoff and colleagues), between perception and action ("common coding"; Hommel and colleagues) and between talkers and listeners ("parity"; Liberman and colleagues) require the same kind of explanation. I will discuss that explanation in the context of my recent research on imitation of speech. Infants and adults have more than the ability to imitate, however. They are also disposed to imitate one another. I will discuss this in the context of my recent research on speech as a between-person "coordination device."
GSAS Workshop on Comparative Syntax and Linguistic Theory at Harvard
A reception will follow immediately upstairs in the Department of Linguistics.
Francesca Del Gobbo
Harvard/UC-Irvine"Appositives"
Eunice Kennedy Shriver Center - Psychology Colloquium Series
Refreshments served 10 minutes prior to talk
Catherine Snow
Harvard Graduate School of Education"Speaking in Words: The Importance of Vocabulary Knowledge in Academic Achievement"
UMass, Amherst: Linguistics Colloquium
Chris Golston
California State University, Fresno"Segmentalism"
MIT - Phonology Circle
Friday, November 7, 2001, 2:00-3:15, Room E39-335
Donca Steriade, MIT Continuation of Analysis of Rhyming Structures in Romanian This is a study of half rimes (or semi-rimes, SR) which makes the assumption that a poet's observed preference for one SR type over another reflects his judgment that the certain SR's involve more similar strings and thus a rhyming pair closer to perfect identity. Relative similarity will be reflected in correspondence rankings: if Corr1 >> Corr2 (where Corr1 and Corr2 are distinct correspondence constraints evaluating the identity between the rhyme domains of two lines) then a pair of lines violating Corr1 will be, all else equal, less harmonic and thus dispreferred to a pair violating Corr2.
The interest in SR patterns and relative similarity judgments in general is motivated by the hypothesis that the learner of a phonological system operates with certain a priori assumptions regarding the relative ranking of correspondence constraints. In the absence of evidence to the contrary, the learner's assumption is, according to this hypothesis, that a more discriminable contrast (say d vs. t in intervocalic position) is protected by a higher ranked correspondence condition (say Ident voice/V_V, I-O) than a less discriminable contrast (say d vs. t in interconsonantal position, corresponding to Ident voice/C_C, I-O). The talk begins with a review of the reasons to entertain this hypothesis about a priori correspondence rankings and then provides the SR evidence that bears on it.
The talk differs a bit from the one I gave last December here. There is now the beginning of an explicit analysis of SR's that characterizes two basic facts: first, poets speaking the same language draw on a shared hierarchy of relative similarity between strings. Second, different poets draw the line between tolerable and intolerable dissimilarity at different points on this shared hierarchy: the formal question is how to express the fact that in any given corpus, certain SR's are absolutely excluded. The empirical generalizations are somewhat richer now: the trends observed first in translations have be verified against native poetic texts and a riming dictionary. Finally, we now have evidence shedding some light on what causes certain pairs of strings to qualify as highly similar.
CAS/CNS Fall 2001 Colloquium Series
Wm. Tecumseh Fitch, Lecturer
Department of Psychology, Harvard University"Animal vocal production and the evolution of speech"
MIT LING-LUNCH
Thursday, December 6, 2001 at 12:00, MIT E39-335 (conference room)Ora Matushansky
MIT"Invitation to beheading, or why head-movement doesn't exist "
MIT RLE Speech Communication Group Seminar
Sheila Blumstein
Brown University"Effects of Lexical Competition and Phonetic Degradation on Lexical Processing: Evidence from Normal Subjects and Aphasic Patients" This research explores how listeners map the properties of sound on to the lexicon (the mental dictionary) and investigates the neural basis of such processing. A number of stages of processing have been identified in the selection of a word candidate. These stages include the extraction of generalized auditory patterns from the acoustic waveform, the conversion of this spectral representation onto a more abstract phonetic feature/phonological, the mapping of this phonological representation onto lexical form (i.e. a word in the lexicon), and the ultimate selection of the word candidate from other potential lexical competitors that share phonetic/phonological properties. Spoken language is often produced in a noisy medium in which the listener must extract the properties of speech to ultimately understand what is being said. Moreover, there is a great deal of variability inherent in the speech production process itself. Among a number of sources of variability, there is variability inherent in the implementation of the phonetic categories of speech. For example, there is a range of voice-onset time values associated with the production of voicing in stop consonants. Another factor that influences access to a word candidate is lexical competition. When a lexical item is presented, it not only activates the lexical representation of that item but also partially activates those words that are phonologically related to it. In the current research, we report on a series of experiments with both normal subjects and aphasic patients exploring the effects of acoustic-phonetic structure on lexical activation and the extent of that influence as a function of the presence/absence of lexical competition. Specifically, we investigated the extent to which acoustic-phonetic changes of an auditorily presented prime stimulus affect the magnitude of semantic priming to a real word target in a lexical decision task. Results from normal subjects are consistent with the view that activation of the lexicon is graded, and that phonetic categories have an internal structure to them organized around prototypical or exemplar members of a phonetic category. Moreover, results suggest that an acoustically modified stimulus not only activates it lexical representation but also partially activates its phonetically contrasting lexical competitor. Results from aphasic patients suggest that they have deficits in the dynamics of lexical activation. Results from Broca's aphasics are consistent with the view that they have a reduction in lexical activation. Thus, they show priming results similar to normal subjects when an acoustically modified prime stimulus does not have a lexical competitor. However, they fail to show priming when the acoustically modified prime stimuli have a lexical competitor. In contrast, results from Wernicke's aphasics suggest that they have an overactivation of the lexical system or a failure to inhibit lexical competitors. As a consequence, Wernicke's aphasics show a similar magnitude of semantic priming irrespective of phonetic manipulation or the presence of lexical competition.
MIT RLE Speech Communication Group Seminar
Jesse Snedeker
Harvard UniversityTitle TBA Not available at this time
Harvard FAS Workshop on Indo-European Historical Linguistics and Poetics
Place: Linguistics Department Seminar Room, Boylston Hall, Fong Auditorium
All are welcome, and refreshments will be served.
P. O. Skjaervo CAN URINE KILL? or
SMASHING URINE: INDO-IRANIAN MYTH AND POETICS
Not available at this time
UMass, Amherst: Linguistics Colloquium
David Dowty
Title TBA
MIT Linguistics Colloquium
Chris Barker
University of California, San Diego"In-situ quantification: beyond Flexible Types" Some relevant manuscripts available at <http://semanticsarchive.net>:
Barker:
- Continuations and the nature of quantification
- Introducing continuations
- Integrity: a syntactic constraint on quantifier scoping
Chung-chieh Shan:
- A continuation semantics of interrogatives that accounts for Baker's ambiguity
- Monads for natural language semantics
MIT - Phonology Circle
Friday, November 16, 2001, 2:00-3:15, Room E39-335
Donca Steriade, MIT with Jie Zhang
"Semi-rimes and perceptual maps: a second try"" This is a study of half rimes (or semi-rimes, SR) which makes the assumption that a poet's observed preference for one SR type over another reflects his judgment that the certain SR's involve more similar strings and thus a rhyming pair closer to perfect identity. Relative similarity will be reflected in correspondence rankings: if Corr1 >> Corr2 (where Corr1 and Corr2 are distinct correspondence constraints evaluating the identity between the rhyme domains of two lines) then a pair of lines violating Corr1 will be, all else equal, less harmonic and thus dispreferred to a pair violating Corr2.
The interest in SR patterns and relative similarity judgments in general is motivated by the hypothesis that the learner of a phonological system operates with certain a priori assumptions regarding the relative ranking of correspondence constraints. In the absence of evidence to the contrary, the learner's assumption is, according to this hypothesis, that a more discriminable contrast (say d vs. t in intervocalic position) is protected by a higher ranked correspondence condition (say Ident voice/V_V, I-O) than a less discriminable contrast (say d vs. t in interconsonantal position, corresponding to Ident voice/C_C, I-O). The talk begins with a review of the reasons to entertain this hypothesis about a priori correspondence rankings and then provides the SR evidence that bears on it.
The talk differs a bit from the one I gave last December here. There is now the beginning of an explicit analysis of SR's that characterizes two basic facts: first, poets speaking the same language draw on a shared hierarchy of relative similarity between strings. Second, different poets draw the line between tolerable and intolerable dissimilarity at different points on this shared hierarchy: the formal question is how to express the fact that in any given corpus, certain SR's are absolutely excluded. The empirical generalizations are somewhat richer now: the trends observed first in translations have be verified against native poetic texts and a riming dictionary. Finally, we now have evidence shedding some light on what causes certain pairs of strings to qualify as highly similar.
MIT LING-LUNCH
Thursday, November 15, 2001 at 12:00, MIT E39-335 (conference room)Tania Ionin
MITUtpal Lahiri
UC Irvine"Left out in the cold: Frozen scope and discourse function in Russian" "Embedded Interrogatives Revisited" Abstracts Lahiri:
In this talk, I examine the properties of the interrogative complements of verbs like "know" and propose that they are interpreted via their most informative answers. I examine a range of data including the quantificationally variable readings of embedded readings; strong, weak and weaker than weak exhaustiveness (and the factors that are responsible for such variation); the behavior of plural interrogatives (and their similarity to plural expressions generally).
Ionin:
English sentences like (1) that contain two quantifiers typically exhibit scope ambiguity. The corresponding sentences in Russian, on the other hand, allow only scope that corresponds to the surface configuration (cf.: 2). In this paper, I argue that, despite appearances, Russian has covert QR as well as reconstruction; I show that both are directly related to discourse function.
Russian exhibits scope rigidity in both standard SVO order (2a), and scrambled OVS order (2b), in intonationally neutral sentences. Any account of these facts must answer two questions:
Q1: Does covert QR exist in Russian, and (if so) what is its distribution?
Q2: Does scope reconstruction exist in Russian, and (if so) what is its distribution?
Answer to Q1: Contrary to appearances, covert QR exists in Russian, as evidenced by a quantifier’s ability to scope over a higher adverb (3). However, a quantifier is unable to scope over another quantifier (2) or over an intensional verb (4). Adopting the framework of Heim and Kratzer (1998) / Fox (2000), I argue that covert QR in Russian is motivated only by type considerations. Assuming that QPs are of type <et, t>, a QP must move at least as far as the spec of vP in order to be interpretable (Heim and Kratzer 1998). The sentences in (2) would then have the LFs in (5). I argue that in Russian, in contrast to English, QR to a higher position &emdash; not motivated by type considerations &emdash; is impossible. This difference is tied to the existence in Russian, but not in English, of overt quantifier movement through scrambling (2b): operations that can be done overtly are not done covertly. This explanation raises the question of why QR for type considerations is not overt in Russian &emdash; i.e., why (under neutral intonation), cases such as (6) are outlawed. I suggest that this is due to the nature of scrambling in Russian: I will show that in intonationally neutral sentences, leftward scrambling is always motivated by topicalization (King 1995), and that there is typically only one preverbal internal topic position.
Answer to Q2: I suggest that lack of scope reconstruction in Russian is also tied to discourse function. King (1995) argues that all leftward DP-movement in intonationally neutral sentences, including subject raising (5a), is motivated by the discourse property of topicalization. I suggest that full reconstruction of the topicalized DP is impossible, since the reconstructed element would no longer function as the topic. By Economy, vacuous application of topicalization is not allowed. I formulate the constraint in (i):
i) Preservation of Discourse Function: The topicalizing operator cannot undergo reconstruction at LF
This constraint predicts that partial reconstruction of the leftmost DP should be possible, as long as the topicalizing operator stays in its surface position. I will show that this prediction is carried out in two domains: pair-list readings and anaphor binding.
Sauerland (2000) argues that the pair-list reading of a how-many question can be derived through partial reconstruction. In his analysis,the [n-many NP] part of the phrase undergoes reconstruction, while [how(n)] stays in its surface position. On the other hand, Sauerland argues that which-phrases cannot undergo partial reconstruction, since that would create a WCO effect. Given the constraint in (i), we then expect to find pair-list readings available for Russian how-many phrases (through partial reconstruction), but not for Russian which-phrases. This prediction is confirmed, as (7) and (8) demonstrate.
Evidence for availability of partial reconstruction is also found in the domain of anaphor binding. Reconstruction of a scrambled DP for anaphor binding is allowed in Russian (9a). However, when reconstruction for binding results in reconstruction for scope, the sentence becomes ungrammatical under neutral intonation (9b). I suggest that (9a) and (9b) have the logical structures in (10a) and (10b), respectively, with the topicalizing element being separated from the anaphoric part of the phrase, which reconstructs. (10b) describes the pragmatically impossible situation of a single dog (the topic of 9b) belonging to every boy &emdash; hence the ungrammaticality of (9b). Full reconstruction of the scrambled DP in (9b) is ruled out by (i).
Given this account of scope rigidity in intonationally neutral sentences, I will also show that sentences in which two constituents receive contrastive focus (and stress) allow for scope ambiguity (11). Unlike topics, contrastively focussed DPs do not have to be in preverbal position (cf.: Junghanns and Zybatow, 1997). Adapting Krifka’s (1998) analysis of focus in German to the Russian data, I will suggest that focus creates LF-movement possibilities that are otherwise disallowed, such as optional QR and full reconstruction for scope. Thus (11a) can have either one of the LFs in (12), and (11b) &emdash; in (13).
- one boy kissed every girl (a>every), (every>a)
- a) odin mal’chik poceloval kazhduju devochku (neutral intonation)
one boy-NOM kissed every girl-ACC (one>every) *(every>one)
b) odnu devochku poceloval kazhdyj mal’chik t (neutral intonation)
one girl-ACC kissed every boy-NOM (one>every) *(every>one)- Masha chasto obedaet s kazhdym mal’chikom (neutral intonation)
Mary often dines with every boy-INSTR ?(often>every), (every>often)- Masha otkazalas’ uvolit’ bol’she des’jati rabochix (neutral intonation)
Mary refused fire-INF more ten-GEN workers
(refuse>more than ten): Mary refused to fire more than ten workers (but she is willing to fire up to ten)
*(more than ten>refuse): There are more than ten workers whom Mary refused to fire- a) [IP odin mal’chik1 [vP kazhduju devochku2 t1 [VP poceloval t2]]]
b) [IP odnu devochku2 [vP t2’ kazhdyj mal’chik1 [VP poceloval t2]]]- ??/*odin mal’chik kazhduju devochku poceloval (neutral intonation)
one boy-NOM every girl-ACC kissed- (ja xochu znat’) skol’kix devochek poceloval kazhdyj mal’chik (neutral intonation)
(I want know-INF) how-many girls-ACC kissed every boy-NOM
"(I want to know) how many girls every boy kissed"
?(how many>every): possible answer "Every boy kissed 2 girls"
(every>how many): pair-list answer "John kissed 2 girls, Bill kissed 3 girls, and Sam kissed 5 girls"- (ja xochu znat’) kakuju devochku poceloval kazhdyj mal’chik (neutral intonation)
(I want know-INF) which girl-ACC kissed every boy-NOM
"(I want to know) which girl every boy kissed"
(which>every): possible answer "Every boy kissed Mary"
*(every>which): *pair-list answer "John kissed Mary, Bill kissed Sue, and Sam kissed Alice"- a) [svoju1 sobaku]2 videla Masha1 t2 (neutral intonation)
[self’s1 dog-ACC]2 saw Mary1-NOM t2
"Mary saw her own dog"
b) ??/* [svoju1 sobaku]2 videl [kazhdyj mal’chik]1 t2 (neutral intonation)
[self’s1 dog-ACC]2 saw [every boy-NOM]1 t2
"Every boy saw his own dog"- a) $ x such that (x is a dog, and x belongs to Mary, and Mary saw x)
b) $ x such that (x is a dog, and " y (if y is a boy then x belongs to y and y saw x))- a) ODIN mal’chik poceloval KAZHDUJU devochku (double focus)
one-FOC boy-NOM kissed every-FOC girl-ACC (one>every), (every>one)
b) ODNU devochku poceloval KAZHDYJ mal’chik (double focus)
one-FOC girl-ACC kissed every-FOC boy-NOM (one>every), (every>one)- a) [IP odin mal’chik1 [IP kazhduju devochku2 [vP t2’ t1 [VP poceloval t2]]]]
b) [IP ____1 [IP kazhduju devochku2 [vP t2’ odin mal’chik1 [VP poceloval t2]]]]- a) [IP odnu devochku2 [IP kazhdyj mal’chik1 [vP t2’ t1 [VP poceloval t2]]]]
b) [IP ____2 [IP kazhdyj mal’chik1 [vP odnu devochku2 t1 [VP poceloval t2]]]]References:
Fox, D. 2000. Economy and Semantic Interpretation. Cambridge, MA: MIT Press.Heim, I., and A. Kratzer. 1998. Semantics in Generative Grammar. Blackwell Publishers.
Junghanns, U. and G. Zybatow. 1997. Syntax and information structure of Russian clauses. Proceedings of FASL: The Cornell Meeting, 1995, 289-319.
King, T.H. 1995. Configuring Topic and Focus in Russian. Stanford, CA: CSLI.
Krifka, M. 1998. Scope inversion under the rise-fall contour in German. Linguistic Inquiry 28, 662-688.
Sauerland, U. 2000. "How many"-questions and pair-list situations. Snippets 1, 17-18.
THE CAMBRIDGE PROSODY GROUP
(Note: We'll meet in 36-511 MIT and possibly move to another room..)
Nanette Veilleux, Duane Watson, Lisa Lavoie
Reports from prosody-related conferences We will also discuss some possibilities for subsequent meetings. As always, feel free to forward this message to others who might be interested.
MIT Center for Bilingual/Bicultural Studies
The Language of Power/The Power of Language Claude Hagège
"English as a Global Language: Real or Imagined Threat?" Not available at this time.
MIT RLE Speech Communication Group Seminar
Ken Stevens and S.Jay Keyser
MIT"Enhancement and Gestural Overlap in Speech Production" We describe a model of speech production that converts a discrete linguistic representation of an utterance into time-varying gestures executed by the various articulators and then to an acoustic output. The linguistic representation contains a planning stage consisting of a sequence of phonological segments, together with prosodic information. In this talk we examine the interaction of enhancement (to increase perceptual saliency) and gestural overlap (to increase efficiency) with this representation to produce an acoustic output.
UMass, Amherst: Linguistics Colloquium
Alan Prince and Bruce Tesar
"Learning Phonotactic Distributions"
MIT Linguistics Colloquium
Beth Levin
Stanford University"What Alternates in the Dative Alternation?" (This talk presents work carried out with Malka Rappaport Hovav, The Hebrew University of Jerusalem)
Most analyses of the dative alternation fall into two broad types. In the first type, each dative verb is associated with a single semantic representation, giving rise to a basic variant, with the other variant being syntactically derived from it. Such analyses tend to be syntactically motivated, with syntactic principles accounting for most of the properties of the alternation. In contrast, in the second type, the two variants are associated with distinct meanings, each paired with the appropriate syntactic form. The alternation arises because the recipient argument qualifies for linking to object on one meaning and the theme does on the other. Such analyses tend to be motivated by lexical semantic considerations. We propose dative verbs have a single semantic representation as in the one-meaning approach, but as in the two-meaning approach the two variants are not syntactically related.
The two-meaning approach is often motivated by apparent differences in the interpretation of the variants. However, the relevant facts are better explained on the assumption that each dative verb is associated with a single meaning, with apparent lexical semantic effects arising from Gricean implicatures and information structure considerations. More generally, we attribute the choice of variant in English to information structure and heaviness considerations, following a number of researchers. As support, we show the same considerations explain the distribution of idioms in the two variants.
We propose the source of the alternation is not the availability of two meanings for a dative verb, but rather the availability of two semantic analyses for a recipient: goal (via the Localist Hypothesis) and possessor. The recipient-as-goal analysis gives rise to the to-variant (give a book to Pat) since in English "to" is the marker of goals, including recipient goals. The recipient-as-possessor analysis gives rise to the double object variant (give Pat a book); first objects are the English expression of possessors in transitive clauses, some of which are recipients. The dative alternation, then, reflects a recipient alternating between its expression as goal and its expression as possessor. We suggest languages have a core grammatical relation dedicated to the expression of possessors: in some a dative case-marked NP, but in others, including English, the first object in a double object construction. It is unsurprising, then, that there are repeated observations that possessors in the double object variant don't show the full range of direct object properties. In fact, the distribution of these properties across recipient and theme is consistent with our analysis. Thus, our analysis provides better coverage of English data in the context of a unified account of crosslinguistic facts.
MIT LING-LUNCH
Thursday, November 8, 2001 at 12:00, MIT E39-335 (conference room)Ted Gibson
MIT"Resource influences and contextual influences on sentence complexity" Why is sentence (1) so much harder to understand than sentence (2)?
(1) The student who the professor who the scientist collaborated with had advised copied the article.
(2) The scientist collaborated with the professor who had advised the student who copied the article.
The two sentences have close to the same meaning, yet (1) is far more complicated than (2).
In this presentation, I will present evidence gathered in my lab over the last few years for four potentially independent factors in sentence complexity at play in sentences like (1) and (2):
1. Integration distance between syntactic dependents. The processing cost of integrating a new word w is shown to be proportional to the distance between w and the syntactic head to which w is being integrated. The syntactic dependents in (1) are generally much further apart than they are in (2), making (1) more complex.
2. Syntactic storage in terms of the number partially processed syntactic dependents. Our evidence suggests that complexity increases as the number of predicted syntactic dependents increases. This factor also predicts greater complexity for (1) relative to (2).
3. Information flow. Background information (such as the information in a restrictive relative clause) is comprehended more easily early in a sentence. In contrast to the other two factors, this factor makes the relative clause in (1) easier to understand than the relative clause in (2).
4. Discourse complexity. The more complicated the mental model is for the sentence being comprehended, the more difficulty people have in reading that sentence. Restrictive relative clauses like those in (1) and (2) implicate a contrast set. In a null context, it is shown that restrictive RCs are therefore more complex than non-restrictive RCs, which do not have the same implication of a contrast set.
Evidence for these four factors will be provided in the form of reading times and comprehension questionnaires across a variety of English materials.
Collaborators: Dan Grodner, Tessa Warren, Duane Watson, Timothy Desmet., Evan Chen, Florian Wolf, Kara Ko, James Thomas, Maria Babyonyshev, Ken Nakatani.
MIT SEMINAR SERIES: CBCL Brains & Machines Seminar
Dr. Stephen Smale
UC-Berkeley, Dept. of Mathematics"Learning and the Evolution of Language"
Stephen Smale is one of the leading mathematicians of our time. He was awarded a Fields Medal at the International Congress at Moscow in 1966. One of Smale's impressive results was his work on the generalized Poincaré conjecture. Another area in which Smale has made a very substantial contribution is in Morse theory, which he has applied to multiple integral problems. In fact, Smale attacked the generalized Poincaré conjecture using Morse theory. A further discovery of Smale's related to strange attractors.
Smale's career changed direction in the late 1960s when he moved into applications. He modeled physical processes by dynamical systems, opening new lines of inquiry. The n-body problem and electric circuit theory were among the applications that Smale framed in the language of dynamical systems. For much of the 70's, Steve focused on economics, injecting topology and dynamics into the study of general economic equilibria. Smale's recent work has been on theoretical computer science. With L. Blum and M. Shub, he has developed a model of computation, which includes both the Turing machine approach and the numerical methods of numerical analysis. Even more recently, he has started to work in the area of statistical learning. A long paper co-authored with F. Cucker entitled, "On the Mathematical Foundations of Learning," will appear soon in the Bulletin of the American Mathematical Society.
Smale has received many honors for his work. In addition to the Fields Medal, he was awarded the Veblen Prize for Geometry by the American Mathematical Society in 1966 "for his contributions to various aspects of differential topology." In 1996, Smale received the National Medal of Science for four decades of pioneering work on basic research questions, which have led to major advances in pure and applied mathematics. A biography of Steve Smale, The Mathematician Who Broke the Dimension Barrier, was written by S. Batterson and published by the Mathematical Association of America in 2000.
MIT RLE Speech Communication Group Seminar
Rupal Patel
Columbia UniversityTitle TBA Not available at this time
Plenary Speaker: Daniel A. Dinnsen, Indiana University
MIT - Phonology Circle
Friday, November 2, 2001, 2:00-3:15, Room E39-335
Jonathan Barnes
UC Berkeley"Moraicity and speech timing in Turkish and elsewhere" This paper is a reevaluation of the moraic isochrony, or mora-counting hypothesis underlying such recent investigations of the relationship between phonological representations and speech timing as Broselow, et al. 1997, Hubbard 1994, Maddieson 1993. The essence of this hypothesis is that an abstract unit of syllable weight, i.e. the mora, is responsible for determining the assignment of phonetic segmental durations within the syllable. Specifically, each mora in a string is associated with a more or less fixed amount of phonetic duration (relativized of course to speech tempo, inherent segment duration, segmental environment, etc.),which becomes the sole property of the segments to which that mora is linked. A single segment linked to a mora will fill that mora's entire durational span, but if more than one segment is linked to the same mora, that duration will be shared among them. Other things being equal, then, a consonant or vowel monopolizing a single mora is predicted to be longer than a comparable segment sharing its mora with another consonant or vowel.Assuming this, we are able to account for a variety of phenomena such as Closed Syllable Shortening and Compensatory Lengthening.
I will present the results of an experiment conducted on segment durations in Turkish, a language in which vowels in closed syllables are, somewhat disturbingly, systematically longer than comparable vowels in open syllables. I will argue that the mora-timing hypothesis is insufficient to deal with the complexity of the facts of Turkish, and further makes a number of incorrect predictions about syllable and word duration in Turkish and other languages. Instead, I propose an isochronic constraint mandating uniform durations of syllable rhymes, rather than of their individual subcomponents, which both accounts for the typologically aberrant vowel durations in Turkish, as well as for other data previously adduced insupport of moraic isochrony. Additionally, having shown the Timing Tier tobe ill-equipped to account for, at the very least, timing, I will suggest we might in fact begin to reevaluate the mora-as-autosegment more generally, in connection with the other phenomena it was originally proposed to explain.
MIT Linguistics Colloquium
Bill Ladusaw
University of California, Santa Cruz"Restriction and Saturation" [Reporting joint work with Sanda Chung, University of California, Santa Cruz] [This talk reports joint work with Sandra Chung.].
MIT LING-LUNCH
Thursday, November 1, 2001 at 12:00, MIT E39-335 (conference room)Martha McGinnis
U. CalgaryJonathan David Bobaljik
McGill U."On the systematic aspect of idioms" "What's in a Paradigm?" (McGinnis)
It has traditionally been assumed that the meaning of some or all phrasal idioms is non-compositional. However, I will argue that the aspectual meaning of idioms is completely systematic: there are no special aspectual restrictions on idioms, and moreover, the aspectual properties of an idiom are compositional, combining the aspectual properties of its syntactic constituents in the usual way. This observation supports the theory of Distributed Morphology (DM; Halle & Marantz 1994).
All aspectual classes contain idiomatic VPs. Any aspectual classification of non-idiomatic VPs also applies to idiomatic VPs. In this sense, idiomatic VPs are aspectually systematic. More intriguing is the evidence that the aspectual properties of idiomatic VPs are, at least in part, syntactically derived. The theories of
Representational Modularity (RM: Jackendoff 1997) and DM make different predictions for the interpretation of idioms. RM maintains that idioms are syntactically complex, but differ arbitrarily from non-idioms in the mapping to semantic interpretation. In DM, however, certain ('structural') components of meaning are assembled in the syntax, so the syntactic derivation of idioms is predicted to have semantic consequences.
Marantz (1997) suggests that one such consequence is aspectual. He argues that kick the bucket cannot mean 'die' because it has the punctual aspect of a transitive VP with a definite complement. If this analysis is correct, it predicts that even if a VP has a non-compositional idiosyncratic meaning, it will have a compositional structural meaning. Specifically, it will have the same aspectual properties as any VP with the same syntactic properties. As I will show, this prediction is confirmed. The meaning of idioms is not arbitrarily related to the syntax, as RM maintains, but is instead partly derived from it.
References
Halle, Morris, and Alec Marantz. 1994. Distributed Morphology and the pieces of inflection. In The view from Building 20: Essays in linguistics in honor of Sylvain Bromberger, ed. Kenneth Hale and Samuel Jay Keyser, 111-76. Cambridge, Mass.: MIT Press.
Jackendoff, Ray. 1997. The architecture of the language faculty. Cambridge, Mass.: MIT Press.
Marantz, Alec. 1997. No escape from syntax: Don't try morphological analysis in the privacy of your own Lexicon. In University of Pennsylvania Working Papers in Linguistics 4:2: Proceedings of the 21st Annual Penn Linguistics Colloquium, 201-25. Department of Linguistics, University of Pennsylvania, Philadelphia
///////////////////////////////
(Bobaljik)
The Question: Plank 1991 begins with the observation that "[t]he earliest extant grammatical texts are paradigms." (p.161) The long linguistic and philological tradition have established a wealth of knowledge about the properties of paradigms, notably regarding the issue of syncretism, but one fundamental question has not been definitively answered, namely (1):
(1) Does knowledge of language (grammar) include knowledge (memorization) of paradigms themselves or just of the pieces that constitute paradigms and rules for generating them? The issue: That is, given that a set of features, and a set of "morpheme realization rules" or "vocabulary items" as in (2) is sufficient information to generate the nominative declensional paradigm for regular nouns (and pronouns, short adjectives and past participles) in Russian, is there any reason to posit knowledge of paradigmatic structure beyond (2)?
(2) a. ACTIVE FEATURES: b. VOCABULARY ITEMS:
Halle 1997:428
3 genders -/i/ ¤ plural
2 numbers -/a/ ¤ feminine
-/o/ ¤ neuter
-/O/ ¤ elsewhere ("O" = yer)
A purported argument. Williams 1994:26, e.g., criticises theories such as (2) observing that "the pattern of syncretism is a quite abstract structure, standing above particular words, particular rules, particular suppletive relationships." The notation (2b) treats it as an accidental property of these particular morphemes that gender is not expressed in the plural, but this is clearly a quite general fact of Russian morphology, holding across declension class, case, part of speech (N, V, A), and regardless of the particular desinences expressing case, number and gender. The existence of such systematic syncretisms (i.e., as opposed to accidental homophony) is taken as an argument that (2) is insufficient and thus for the grammatical relevance of paradigms over and above the pieces that constitute them.
The incompleteness of this argument. The existence of such syncretisms establishes that (2) is insufficient, but it does not establish that paradigms are necessary. Such syncretisms are captured within frameworks such as Distributed Morphology via (e.g.,) impoverishment rules (see especially Bonet 1991); such rules delete features from the Morphosyntactic Representation prior to the application of Morphological Realization (or Vocabulary Insertion) Rules as in (2b). Impoverishment rules capture language-specific syncretisms and must be learned, but they are of no greater formal complexity than the devices employed within theories advocating paradigms. As the existence of syncretism follows directly from neither approach, and is statable in either, syncretism alone in principle does not distinguish between the theories.
Are the theories distinct? Williams 1994 offers an argument that can distinguish the classes of theories in principle. He argues that there are universal constraints on the patterns of syncretism that can exist in a given language. In particular, he proposes that every language must have an "instantiated basic (sub-)paradigm," such that all other (sub-)paradigms mark a subset of the distinctions of it. Thus, it follows from his theory that if two cells "A" and "B" are syncretic in some paradigm, and cells B and C are syncretic in a another, then there will be a third paradigm in which A,B,C (i.e., all three are distinguished). This requirement is not statable in frameworks such as Distributed Morphology; if substantiated as a universal, it would constitute a convincing argument for paradigmatic knowledge, i.e., a "yes" answer to (1).
Against basic paradigms. In the Russian nominal and adjectival declension, accusative (A) is systematically identical to either genitive (B) or nominative (C) (conditioned by animacy) in all plurals and masculine and neuter singulars. This syncretism is systematic and not merely a fact about individual morphemes. Williams's basic paradigm hypothesis predicts (correctly) the existence of another paradigm in Russian in which accusative, genitive and nominative are distinct; this happens to be the feminine singular. However, exactly in the feminine singular, the dative and prepositional cases (distinct elsewhere) are systematically syncretic. In essence, since all six cases are distinguished in Russian, but no single paradigm overtly distinguishes all six, the basic paradigm hypothesis cannot be maintained. Hence, this potential argument in favour of the existence of paradigms fails. Conclusion. We do not claim to have answered (1) definitively in this paper. We hope merely to have shed further light on the issue by (a) clarifying formally what is at stake, i.e., what can be taken as an argument for paradigms as a theoretical primitive, and (b) showing that on one point where the paradigmatic theory can be made to be more restrictive, such restrictions are found to be counter-evidenced by even moderately inflected languages such as Russian.
Bonet, Eulalia (1991). Morphology after syntax. Ph.D. Dissertation, MIT, Cambridge, Mass.
Halle, Morris (1997). Distributed Morphology: Impoverishment and Fission. Benjamin Bruening, Yoonjung Kang & Martha McGinnis (eds.). PF: Papers at the Interface: MIT Working Papers in Linguistics 30, 425-449.
Plank, Frans (1991). Rasmus Rask's dilemna. In Frans Plank (ed.). Paradigms: The economy of inflection. Mouton de Gruyter, Berlin, 161-196.
Williams, Edwin (1994). Remarks on lexical knowledge. Lingua 92, 7-34.
Memorial for Ken Hale
MIT RLE Speech Communication Group Seminar
Ioana Chitoran
Dartmouth College"Recoverability Constraints on Gestural Overlap:
Articulatory Evidence from Georgian Stop Sequences"Recent investigations of gestural patterning have found that consonant gestures exhibit less temporal overlap in a syllable/word onset than in a coda or across syllables. One possible account for this difference is that substantial overlap of obstruent gestures may threaten their perceptual recoverability, particularly in utterance-initial position, where no acoustic information is provided during the formation of the consonant gestures (no VC transitions). Additionally, recoverability considerations may account for results of Byrd (1992, 1996), Zsiga (1994), Surprenant and Goldstein (1998) who show that front-to-back order of place of articulation in stop-stop sequences (labial-coronal, coronal-dorsal, labial-dorsal) allows more overlap than the opposite order. Presumably, the C1 release would produce no acoustic manifestation if C2 were already formed just in case C2 constricts the vocal tract anterior to C1. If C2 were posterior to C1, then at least some acoustic information would be generated on release of C1 (even without a substantial release burst). Similar recoverability requirements are proposed here to account for consonant sequencing phenomena in Georgian, which violates the sonority sequencing generalization. This talk reports on a magnetometer study of gestural overlap in Georgian, a South Caucasian language which allows a variety of stop-stop sequences in onsets, and elsewhere. Two native speakers of Georgian served as subjects. We tested for the positional effect by examining C1C2 stop sequences as a function of position in the word. We also tested the hypothesis that front-to-back ordering of stops allows more overlap than back-to-front ordering. We therefore examined sequences with different orders of place of articulation of C1 and C2. This hypothesis has previously been tested in an acoustic study in Georgian (Chitoran 1999) by measuring the inter-burst interval between C1 and C2. The interval was found to be significantly shorter in front-to-back than in back-to-front sequences, suggesting a higher degree of overlap in the former. The recoverability-based predictions were borne out in the magnetometer study. Effects of both position in the word and of order of place of articulation were obtained. A significant difference was found in the timing of the onset of the C2 gesture relative to C1. C2 was found to overlap a greater proportion of C1's movement duration word-internally and across words than word-initially. More overlap was also found in sequences with a front-to-back order of articulation than in back-to-front sequences. Georgian syllable onsets violate the sonority sequencing generalization but are apparently sensitive to gestural recoverability requirements as reflected in overlap patterns. Thus, if sonority sequencing has the function of allowing gestures to overlap while still allowing recoverability (cf. Mattingly 1981), this function can apparently be filled in different ways.
UMass, Amherst: Linguistics Colloquium
Cathi Best
"The Perceptual Assimilation Model and Research on Nonnative Speech Perception"
MIT Linguistics Colloquium
Hilda Koopman
University of California, Los AngelesTitle TBA
MIT LING-LUNCH
Thursday, October 25, 2001 at 12:00, MIT E39-335 (conference room)Milan Rezac
Dept. of Linguistics, University of TorontoTitle TBA
MIT RLE Speech Communication Group Seminar
Don Eddington
Massachusetts Eye and Ear Infirmary"Electrode Interaction and Speech Perception Using Lateral-Wall and Medial-Wall Electrode Systems" The observation that percepts elicited by the stimulation of one intracochlear electrode can be influenced by the stimulation of a second electrode (electrode interaction) has long been recognized as one factor that may influence the speech reception of multichannel cochlear implant users (e.g., Eddington, Dobelle et al. 1978; Shannon 1983). In this presentation, we compare measurements of simultaneous electrode interaction and speech reception made in Clarion implantees using standard lateral-wall electrodes to Clarion implantees with electrodes designed to hug the medial wall of the scala tympani.
An interaction index (II) characterizing the degree to which a below-threshold stimulus delivered to a masking electrode influences the threshold of a test electrode was computed as: II = (THRTmM-THRTpM)/(2*ML), where THRTmM is the threshold measured for the test electrode in the presence of an out-of-phase, below-threshold masker, THRTpM is the threshold of the test in the presence of an in-phase, below-threshold masker, and ML is the subthreshold masker level.
The IIs measured for different test/masker electrode combinations in each subject were averaged to give a mean II for each subject and these were then compared across electrode type/configuration and correlated with the subjects9 single-syllable word scores.
The results show a significant difference between the IIs computed for the bipolar and monopolar configurations using lateral-wall electrodes. Except for two of the 11 subjects tested with lateral-wall electrodes, the relationship between their mean II and their single-syllable word reception was highly correlated (r = 0.91). IIs for monopolarly-coupled medial-wall electrodes were significantly lower than for monopolarly-coupled lateral-wall electrodes. Preliminary speech-reception measures made in two CII/HiFocus (medial-wall electrode) subjects showed substantial improvements in speech reception as the number of analysis channels was increased from 8 to 16.
Support for this work was provided by the W.M.Keck Foundation and Advanced Bionics Corporation.
Eddington, D. K., W. H. Dobelle, et al. (1978). "Auditory prosthesis research with multiple channel intracochlear stimulation in man." Ann Otol Rhinol Laryngol Suppl 87(6 Pt 2): 1-39.
Shannon, R. V. (1983). "Multichannel electrical stimulation of the auditory nerve in man: II Channel interaction." Hear Res 11: 157-189.
Third annual
Speech and Hearing Foundation of Massachusetts Lecturesponsored by Emerson College's School of Communication
Department of Communication Sciences & Disorders
Rachel Mayberry
McGill University"The Role of Early Experience in Sculpting Language Systems: Insights from ASL"
Admission is free. Handicap accessible.
For more information: Voice: 617-824-8730; TTY: 617-824-8307
MIT - Phonology Circle
Friday, October 19, 2001, 3:00, Room E39-335
Jie Zhang
Harvard University"The Effects of Duration and Sonority on Contour Tone Distribution-Typological Survey and Formal Analysis" This talk addresses two general questions in phonology: (a) Are positional prominence effects contrast-specific? (b) For a specific phonological contrast, is its positional prominence behavior tuned to language-specific phonetic patterns?
These questions are investigated through the behavior of contour tones. Unlike many other phonological features, the production and perception of contour tones crucially depend on the duration and sonority of the rime. This provides a testing ground for whether the positional prominence behavior of contour tones is tied to its specific articulatory and perceptual needs. Furthermore, there exist multiple phonological factors that affect the duration of the sonorous portion of the rime, and the magnitudes of these effects differ on a language-specific basis. This provides a testing ground for whether the contour tone behavior of a language is tuned to its specific phonetic patterns.
In a survey of 187 languages, the distribution of contour tones is found to correlate closely with the duration and sonority of the rime. Syllables with longer sonorous rime duration, e.g., those that are long-vowelled, sonorant-closed, stressed, prosodic-final, or in a shorter word, are more likely to carry contour tones. This supports the contrast specificity of positional prominence, since the distribution of contour tones is decidedly different from that of many other phonological features, and it is tuned to the specific articulatory and perceptual needs of contour tones.
In phonetic studies of languages with the same multiple factors that induce rime lengthening, it is found that contour tones always favor the factor with the greatest lengthening, even though different languages have different factors that induce the greatest lengthening. This is evidence for the relevance of language-specific phonetics in positional prominence.
Formal apparatus couched in Optimality Theory is proposed to account for the effects of duration and sonority on contour tone distribution. The apparatus necessarily encodes many phonetic details. But it predicts only general patterns that are attested in the survey, and it can account for both the 'phonological' effects of tone and length neutralization and the 'phonetic', albeit language-specific, effects of partial contour reduction and rime lengthening.
MIT LING-LUNCH
Thursday, October 18, 2001 at 12:00, MIT E39-335 (conference room)Cedric Boeckx
University of Illinois, Urbana-ChampaignAndrew Nevins
MIT"The Nature of Resumption" "Counterfactuality without Past Tense" Boeckx Abstract:
In this talk I examine the nature of resumption in several, typologically unrelated languages and develop an analysis that accounts for the robust (syntactic, semantic, morphological) generalizations that are found cross-linguistically. The proposal I put forth goes against the standard base-generalization or minimal-copy-pronunciation accounts of resumptive pronouns. In a nutshell, I am arguing that the resumptive pronoun and its antecedent form a constituent at an early stage of the derivation and are separated by movement processes. The consequences of this theory for islandhood will also be examined.
Nevins Abstract:
Iatridou (2000) has argued that many languages employ past tense morphology to express counterfactuals (CFs). According to this proposal, languages employ the past tense because it functions as an exclusion operator; in the temporal domain, the set of topic times excludes the time of utterance. In the modal domain, then, past tense morphology allows the set of topic worlds to exclude the actual world, thus establishing the implicature of counterfactuality. Iatridou analyzes the counterfactuality contributed by past tense morphology as an implicature, rather than an assertion, following arguments in the philosophical literature (Anderson (1951), Stalnaker (1975))that counterfactuality is cancelable (1). The goal of this paper is twofold: first, to examine how counterfactuality is expressed in a language without past tense morphology, and second, to explore whether, at the morphology-pragmatics interface, the choice of particular morphological means to expresses counterfactuality affects the set of resulting implicatures.
In exploring how counterfactuality is expressed in languages that lack past tense morphology, we consider the case of Mandarin Chinese, in which a specialized complementizer, yaobushi, signals counterfactuality in the protasis of a conditional, and can be used for past (2), present (3), and Future-Less-Vivid (FLV) (4) CFs. Importantly, yaobushi can never be used in a non-CF environment (5), in which a ruguo-type conditional is used (Cheng & Huang, 1996)). We point out two important morphosemantic differences between Chinese yaobushi CFs and those in which past tense expresses CF. (A) In contrast to the Iatridovian systems in which verbal-inflectional material (tense, aspect) is used to express CF, in Chinese, any aspectual morphology that co-occurs with yaobushi makes the expected semantic contribution (and is not co-opted for CF/modal purposes). (B) Contrary to almost all CF cases (but cf. Krause (2001)) in which both protasis and apodosis of conditional have morphological CF expression, the apodosis of the CF need not have any counterfactual marking to be interpreted as such; suprisingly, however, while interpreted as counterfactual, the apodosis is not in the scope of the yaobushi complementizer.
An important fact about those CFs not expressed through past-tense morphology is that they are not cancelable in the sense described above (6). We suggest that cancelability as it has been demonstrated for past-tense type CFs may be a result of the past tense operator itself, which is cancelable in temporal domains (7). The past tense operator establishes that the topic worlds exclude the actual world; however, since the topic worlds are a subset of the P-worlds, it is only an implicature, and thus remains cancelable. In examining one particular case of specialized counterfactual complementizers, we note an important semantic difference between yaobushi and other CF constructions. In effect, while past-tense CFs establish the implicature that topic worlds exclude the actual world, yaobushi-type CFs establish the assertion that in all of the $\neg$P worlds, Q holds. In a counterfactual conditional of the form yaobushi P --> Q, the assertion is not, as one might expect, that the eventuality P did not occur, but rather, that P did occur. In other words, for a past CF, yaobushi P --> Q carries the implication ``If it had not been the case that P had happened, then Q would have happened''. In the absence of past tense, then, Mandarin Chinese employs negation to express counterfactuality; yaobushi is morphologically compositional, comprised of yao, with a conditional ``if'' meaning, and bu, a negation operator. In considering other languages that signal counterfactuality via specialized morphemes (Tagalog, Hebrew, Burmese) we consider the possibility that the noncancelability of yaobushi CF implicatures may be independent from the fact that they contain negation. We show then, that what years of philosophical investigation has concluded -- that counterfactuality is an implicature -- is the result of the morphological expression of CFs in Indo-European languages, and not a hallmark of counterfactual constructions at large.
(1) If the patient had the measles, he would have exactly the symptoms he has now. We conclude, therefore, that the patient has the measles.
(2) Yaobushi ta he le neige duoyao, ta jiu bu hui si le
YAOBUSHI he drank LE that poison, he then not would die LE
``If he hadn't drank that poison, he wouldn't have died'' (PastCF)(3) Yaobushi Lisi you qian, tade nupengyou jiu bu hui dasuan gen ta jiehun
YAOBUSHI Lisi have money, his girlfriend then not would plan with him marry
``If Lisi didn't had money, his girlfriend wouldn't plan to marry him.'' (PresCF)(4) Yaobushi ni qu Jiazhou, women keyi mingtian wanshang zai zheli war.
YAOBUSHI you go California, we could tomorrow evening at here have-fun.
``If you weren't going to California, we could have fun here tomorrow night.'' (FLV)(5) *Yaobushi ni gen wo lai kan dianying, wo jiu hui yige ren qu.
YAOBUSHI you with me come watch movie, I then will one person go.
``If you don't come with me to watch the movie, I will go alone''.(6) *Yaobushi ta mei you fengzhen, tade pifu shang hui you bao. Qishi, yinwei tade pifu shang xianzai you zhei yang de bao, ta haoxiang you fengzhen.
YAOBUSHI she didn't have measles, her skin top would have bumps. Actually, since her skin top now has those kind of bumps, she appears have measles.
*``If it were the case that she had measles, she would have bumps on her skin. Actually, since she does have bumps on her skin now, she appears to have the measles.''
(7) When I saw her, she was in the library. In fact, she may still be in the library.
Boston University
undergraduate
Linguistics Association
OPEN HOUSEfrom 1:00 to 2:00 PM on Wednesday, October 17, 2001
Place: Geddes Video Studio, CAS 533
(right near the Geddes Language Lab)
All BU undergraduates with an interest in linguistics are welcome !!! Come meet other linguistics students and plan for future activities.
There will also be a showing of a 23-minute videotape:"Dialects of Canadian English"
GSAS Workshop on Comparative Syntax and Linguistic Theory at Harvard
Place: Linguistics Department Lounge, Boylston Hall, Room 303, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Reception immediately following the talk
Gulsat Aygen "Case licensing in Turkish, Kazakh and Tuvan subordinate clauses" Based on data from Turkic languages (Turkish, Kazakh, Tuvan ), I will argue against the widely accepted claim (Hwang 1997, Kornfilt 1985) that the case on the subject is licensed either by the [+T] feature on T (in English type languages) or by the phi features on Tense or Agreement (in Turkish type languages). I will propose that In Turkic, case on the subject is licensed by neither the [+Tense] features nor the phi features on T, but by a case feature on C. I will provide evidence from the Nominative-Genitive Case alternation observed on the subject of complement and adjunct clauses in Turkic, both of which bear identical [T] and phi features, and account for the alternation by the difference in the internal and external syntax of dependent clauses and consequently, the nature of C in dependent and independent clauses.
Data in (1&2) illustrate Turkish dependent clauses that have identical surface form, except for the case on the subject. Interestingly, (1) is a complement clause and its subject bears Genitive Case; the one in (2) is an adjunct clause and its subject bears Nominative Case. The verbal predicate in both clauses is identical in form, and it bears a controversial morpheme -DIK , and the nominal agreement morpheme. Turkish has been argued to belong to the group of languages where the phi features on Tense or Agreement license the case on the subject: in this approach verbal agreement licenses nominative and nominal agreement licenses genitive case (Kornfilt 1985, George and Kornfilt 1981, Hwang 1997). Such an approach predicts genitive case in structures like (1&2), but the prediction is not attested.
In terms of the internal syntax of these clauses, I will argue that complement clauses are internally Relative Clauses in Turkish, whereas adjunct clauses are CPs with a lexically field C position. The word zaman/time is a head noun extracted from within the CP in (1), whereas it is a Complementizer at C in (2). Evidence from coordination test in (5) indicates that complement clauses in Turkish are Relative Clauses since they can be coordinated by other Relative Clauses, whereas adjunct clauses cannot (6). Further evidence from the Postverbal scrambling (PVS) tests in (3&4) indicate that V is at C in complement clauses like (1), whereas it is lower than C in adjunct clauses like (2). V at C prohibits adjunction to CP in (3), whereas when V is not at C (presumably at T) due to a lexically filled Comp in (2), adjunction to CP is allowed (Aygen 2000a,b).
The internal syntax of clauses identical in surface form is determined by their external syntax : Complements , selected by a lexical head are CPs, the head of which licenses Genitive on the subject. Adjuncts, not being selected, have C's that license nominative case just like root clauses. Root clauses in Turkic are like adjuncts in terms of not allowing V-to-T-to-C because they do not allow PVS, either, and bearing Nominative Case on their subject.
The prediction of this approach is that adjunct clauses such as PPs that select their complement should bear genitive case, whereas those which do not, such as adverbial clauses should bear nominative. This prediction is attested in a pair of adjuncts in Turkish, identical in form yet different in internal syntax. (7) and (8) are both adjunct clauses with differing internal syntax. (8) allows a head noun to be inserted, indicating that it has the structure of a RC (10), whereas (7) does not (9). The internal structure of (7) is [ PP [ NP [ S-Gen V ] ] Prep ] whereas that of (8) is [ CP S-Nom V Comp ]. The CP is a selected by the Prep, hence the genitive case on the subject in (7) licensed by a selected C; whereas, (8) is not selected; hence the nominative licensed by the unselected C.
To conclude, I will argue that languages are parametrized in terms of what licenses the case on subject, and that the external syntax of clauses, that is being selected or not, is the defining phenomenon in licensing case on the subject in Turkic languages, whereas it is the [T] feature in English. I will also argue that Portuguese, Korean and Berber, which are grouped with Turkish and claimed to be languages where phi features license nominative c ase, should in fact be analyzed in terms of the external syntax of clauses, that is whether they are selected or not.
Complement Clause
(1)Ben- [Ali-nin cam-I kir-dig-i zaman]i
bil-iyor-du-m
I-Nom -GEN glass-acc break-DIK-agr time-Acc
know-prog-past-1sg
'I knew the time when Ali broke the glass'
Adjunct Clause
(2) Ben- [Ali- cam-I kir-dig-i zaman ]
gerceg-i bil-iyor-du-m.
I-Nom -NOM glass-acc break-DIK-agr time truth-acc
know-prog-past-1sg
'I knew the truth when Ali broke the glass'
P(ost)V(erbal)S(crambling)Test
(3)*Ben- [Ali-nin ti kir-dig-i zaman]i cam-ii
bil-iyor-du-m
I-Nom -GEN break-DIK-agr time-Acc glass-acc
know-prog-past-1sg
'I knew the time when Ali broke the glass'
(4) Ben- [Ali- ti kir-dig-i zaman ] cam-ii
gerceg-i bil-iyor-du-m.
I-Nom -NOM break-DIK-agr time glass-acc truth-acc
know-prog-past-1sg
'I knew the truth when Ali broke the glass'
Coordination Test
(5) Ben- [Ali-nin git-tig-i zaman]i ve Hasan-in bin-dig-I ucag-i
I-Nom -GEN go-DIK-agr time-Acc and -GEN get on-DIK-agr
plane-acc
bil-iyor-du-m
know-prog-past-1sg
'I knew the time when Ali went and the plane that Hasan got on'
(6) *Ben- [Ali- git-tig-i zaman ] ve Hasan-in
bin-dig-I yer
I-Nom -NOM go-DIK-agr time and -GEN get
on-DIK-agr place
Ingilizce bil-iyor-du-m.
English know-prog-past-1sg
'*I knew English when Ali went and the plane that Hasan got on'
Contrastive set of adjuncts:
(7)[[Hasan - anla-dig-in]a gre]
herkes anla-r.
-NOM understand-DIK-agr-dat acording.to everybody
understand-present
'Since Hasan understood, everybody will'
(8) [[Hasan-in anla-dig-in]a gre]
kisi gel-ecek-mis.
-GEN understand-DIK-agr-Dat according to three person
come-fut-reported
'According to what Hasan understood three people are going to
come'
(9)*[[Hasan - anla-dig-i] sey-e gre]
herkes anla-r.
-NOM understand-DIK-agr thing-dat acording.to
everybody understand-present
(10) [[Hasan-in anla-dig-i] sey-e gre]
kisi gel-ecek-mis.
-GEN understand-DIK-agr thing-Dat according to
three person come-fut-reported
'According to what/the thing Hasan understood, three people are
going to come'
References
Aygen, Gulsat. 2000a. Extractability and the nominative case feature on tense. (to appear in) Proceedings of the ICTL 2000 (International Conference in Turkish Linguistics). Bogazici University Aug 2000.
Aygen, G. 2000b. V-to-T-to-C: Extractable Subjects and EPP. (to appear in) Proceedings of WECOL 2000, Fresno University, California, October 2000.
George, L. and Jaklin Kornfilt. 1981. Finiteness and Boundedness in Turkish. in Binding and Filtering, ed. Frank Heny, p 105-128. London: Croomhellm Ltd.
Hwang, Kyu-Hong. 1997. Nominative and Default Case Checking in Minimalist Syntax. PhD Diss. University of Washington.
Kornfilt, Jaklin. 1985. Case Marking, Agreement and Empty Categories in Turkish. PhD Diss. Harvard University.
Kural, Murat. 1993. V-TO(-I-TO)-C in Turkish. UCLA Occasional Papers in Linguistics,v.11.ed.Filippo Beghelli &Murat Kural.
MIT Center for Bilingual/Bicultural Studies
The Language of Power/The Power of Language Claude Hagège Claude Hagège holds the Chair of Linguistic Theory at the College de France and is Director at Ecole Pratique des Hautes Etudes. He was the recipient of the Gold Medal from the Centre National de la Recherche Scientifique in 1995. Hagege who has degrees in Hebrew, Arabic, Chinese and Russian, is the author of fifteen books, many of which have been translated into English, including L'Homme de Paroles, L'Enfant aux Deux Langues, Le Souffle de la Langue, and Le Français: Histoire d'un Combat. Recently, he has published Halte à la Mort des Langues (2000) which focuses on the subject of endangered languages.
"Endangered Languages: Birth, Death and Resurrection" Not available at this time.
MIT LING-LUNCH
Thursday, October 11, 2001 at 12:00, MIT E39-335 (conference room)Bridge Copley
Massachusetts Institute of TechnologyLiina Pylkkänen
Massachusetts Institute of Technology"A linguistic argument for
indeterministic futures""Root-selecting, Verb-selecting, and VoiceP-selecting Causatives" A linguistic argument for indeterministic futures
It is possible to treat future operators as either deterministic (i.e., linear) or indeterministic (i.e., modal). The deterministic versions are a great deal less complicated than the indeterministic versions, but I will argue in this talk for the latter. The evidence comes from English _be going to_ (and similar futures in other languages), which I argue has the structure of a future operator under a progressive aspect morpheme. To ensure that this structure has the appropriate meaning, an indeterministic future operator is required. I present further evidence that the future operator in _be going to_ is the same as that in _will_.
Pylkkänen Abstract
Root-selecting, Verb-selecting and VoiceP-selecting Causatives
1. PROBLEM. While much current research assumes that all causative formation takes place in the syntax (Hale and Keyser, Miyagawa, Harley, Doron, Pylkkanen), there is still no serious response to the Fodorian objection that the caused events of lexical causatives are generally unavailable for adverbial modification (1). Thus it appears that modification facts remain a strong argument for a lexicalist position.
2. PROPOSAL. This paper argues that the above problem is solved by assuming a radical version of the syntactic approach where functional heads not only introduce the external argument (Kratzer 1994) but also define the syntactic category of otherwise category-free roots (Marantz 1997). In such a theory there are three sites where causative, or any verbal morphology, can attach: at the root (2a), after verbal category has been determined, (2b), or after the external argument has been introduced (2c). We argue that these three merging sites constitute the core classification of causative constructions crosslinguistically.
To translate Marantz's notion of category-freeness into a Kratzerian type-driven event-semantics we assume that lacking category means lacking specification as to what type of an entity a predicate ranges over: verbs are predicates of events <s,t>, nouns and adjectives are predicates of (non-event) individuals, <e,t>, and, consequently, roots are predicates of unspecified entities, <n, t>.
3. PREDICTED CORRELATIONS OF PROPERTIES. The present proposal predicts a tight correlation between adverbial modification facts and possibilities for intervening morphology between the root and CAUSE.
3.1. Root-selecting causatives (Japanese lexical causatives, English, Pulaar). A root-selecting causative head takes a root as its argument and states that (i) there is an event which the root predicate is true of and that (ii) this event is caused by another event. Thus CAUSE-P is a function from the causing event to truthvalues and the event variable of the caused event is syntactically inaccessible. The modification facts follow: there is no place in the syntax where we have a function from the caused event to truthvalues. Merging an event modifier at the root doesn't work either since event modifiers modify predicates of events, not of unspecified entities.
Root-selecting causative heads cannot combine with a constituent whose category has already been determined. Thus no verbalizing morphology can appear between the root and CAUSE. While Japanese sase-causatives are often ambiguous between a "lexical" and a "productive" analysis, adversity interpretations are only available for lexical causatives (Oehrle and Nishio, Miyagawa). However, for an adversity interpretation to be available, root and 'sase' must be adjacent, as in (3a,b). If other verbalizing morphology intervenes, such as the desiderative 'tai', (3c), or the '-e-' of the associated with the intransitive alternant of the 'kogeru/kogesu' pair, (3d), adversity interpretations are impossible, as predicted.
3.2. Verb-selecting causatives (Bemba, Finnish). A verb-selecting causative head selects for predicates of events and consequently modification of the caused event is possible (4a, b). This modification cannot, however, be agentive since verb-selecting causatives cannot embed an external argument (5a, b). Verbalizing morphology is possible between the root and CAUSE (6a, b). But this morphology cannot be high applicative morphology (Pylkkanen 2000) since high applied arguments are instances of external arguments (cf. McGinnis 2001) (6c).
3.3. VoiceP-selecting causatives (Venda, Luganda, Japanese "productive"). With VoiceP-selecting causatives, all types of modification of the caused event are possible, including agent oriented (7a,b). Similarly, all types of verbal morphology between the root and CAUSE are possible, including external argument-introducing, such as high APPL (8a,b).
4. CONCLUSION. This paper captures previously undescribed correlations between adverbial modification and possibilities for intervening morphology between root and CAUSE while maintaining a unified semantics for causative constructions cross-linguistically. The proposal predicts, and argues for, the existence of verb-selecting causatives, which add to the existing typology of causative constructions (i.e. lexical vs. syntactic) and provide an argument for defining syntactic category in the syntax: in such a framework the distinction between root and verb-selection comes for free. Finally, the present proposal makes strong predictions about causative-applicatives in Bantu (Hyman 2000): the possibility to causativize high applicatives is expected to be rare since there is only one type of causative head that can merge after a high applied argument has been added.
(1) Hahaoya-wa kodomo-o nikai-de nek-ase-ta.
'Mother put the child to sleep upstairs'
(Mother's act must take place upstairs) (Shibatani 2000,11:18a)(L = lambda, E = existential quantifier)
(2a) Root-selecting causative: [CAUSE [Root]]
CAUSE: Lf<n,t>.Le. (Ee') f(e') & CAUSE(e,e')
(2b) Verb-selecting causative: [CAUSE [v [Root]]]
v: Lf<n,t>.Le. f(e)
CAUSE: Lf<s,t>.Le. (Ee') f(e') & CAUSE(e,e')
(2c) VoiceP-selecting causative: [CAUSE [Voice[v [Root]]]]
CAUSE: Lf<s,t>.Le. (Ee') f(e') & CAUSE(e,e')
(3a) Taroo-ga hahaoya-o sin-ase-ta. (3b)Taroo-wa niku-o kog-asi-ta.
'Taro caused his mother to die' 'Taro scorched the meat'
'Taro's mother died on him' 'The meat got scorched to Taro's detriment'(3c) Taroo-ga hahaoya-o sini-taku-sase-ta. (3d)Taroo-wa niku-o koge-sase-ta.
Taro-NOM mother-ACC die-want-CAUSE-PAST 'Taro scorched the meat'
'T made his mother want to die' *'The meat got scorched to T's detriment'
*'Taro was adversely affected by his mother wanting to die'(4) Nonagentive modification:
a. Opettaja laula-tti kuoro-a kauniisti
teacher sing-CAUS choir-PAR beautifully
'The teacher made the choir sing beautifully'
(teacher's action does not need to be beautiful) (Finnish)b. Naa-butwiish-ya Mwape ulubilo.
I.past-run-CAUSE Mwape fast
OK: 'I made Mwape RUN FAST' (Bemba, Givon 1976)(5) Agentive modification:
a. Ulla rakenn-utti Mati-lla uude-n toimistopoyda-n innokkaasti.
Ulla.NOM build-CAUS Matti-ADE new-ACC officetable-ACC enthusiastically(i) 'Ulla ENTHUSIASTICALLY HAD Matti build a new officedesk'
(ii) *'Ulla had Matti ENTHUSIASTICALLY BUILD a new officedesk' (Finnish)
b. Naa-butwiishya umuana ukwiitemenwa
(i) 'I WILLINGLY MADE the boy run'
(ii) '*I made the boy RUN WILLINGLY' (Bemba)
(6) a. seiso- 'stand' seiso-skel-utta 'cause to stand around' (Finnish)
b. Naa-mon-an-ya Mwape na Mutumba
I.past-see-REC-CAUS Mwape and Mutumba
'I made Mwape and Mutumba see each other' (Bemba)c. *Naa-tem-en-eshya Mwape Mutumba iciimuti
I-cut-BEN-CAUS M. M. stick
'I made Mwape cut Mutumba a stick' (Bemba)(7) a. Muuhambadzi o-reng-is-a Katonga modoro nga dzangalelo
salesman SC-buy-CAUSE-FV Katonga car with enthusiasm
'The salesman made Katonga BUY THE CAR EAGERLY' (Venda)b. Omusomesa ya-wandi-s-a Katonga ne obu nyikivu
teacher 3sg.PAST-write-CAUSE-FV Katonga with the dedication
'The teacher made Katonga WRITE WITH DEDICATION' (Luganda)(8) a. Katonga o-n-tandul-el-is-a tsimu ya mukegulu
Katonga SC-1sg-survey-APPL-CAUSE-FV old.woman the field
'Katonga made me survey the field for the old woman' (Venda)b. tambula 'walk' tambu-za 'make walk'
tambul-ir-a 'walk for' tambul-i-z-a 'make walk for' (Luganda)
MIT LING-LUNCH
Thursday, October 4, 2001 at 12:00, MIT E39-335 (conference room)Elena Guerzoni
Massachusetts Institute of Technology"EVEN NPI QUESTIONS" THE PROBLEM: As is known, standard NPIs like 'any' and minimizers NPIs like 'lift a finger' are grammatical in questions. The two categories of NPIs, however, affect the interpretation of the question differently as shown in (1) (an aspect of the problem addressed in Ladusaw '79, but see also Borkin '71).
(1) a. Did John lift a finger to help?
only Negative Rhetoricalb. Did anybody call?
also Information SeekingThis paper focuses on this contrast. My proposal can be seen as an elaboration of Ladusaw's original insight, which attributes the difference between (1a) and (1b) to a general pragmatic principle that links the way a question is asked to the speaker's expectations concerning the answer.
-THE PROPOSAL: Y/n questions denote sets containing a proposition and its negation. However I show that in all contexts when minimizers occur in a y/n question the set of possible answers relative to the context is the singleton containing the negative answer - a situation that yields the effect of a rhetorical question. The reason why only minimizers trigger this effect is that they contain a (silent) 'even' (like Hindi NPIs (Lahiri'98) and Dutch ook maar-NPIs (Rullmann'96)), while 'any' does not (Schmerling '71, Heim'84).
-DISCUSSION: Minimizers like 'lift a finger' denote the low endpoint of a scale. When the two elements involved in their structure ('even' and an expression referring to a lower scale-endpoint) occur in a question only a negative rhetorical reading is indeed available (Wilkinson '96), as shown in (2). Thus, an explanation of the rhetorical reading of (2) will automatically extend to (1a).
My account is based on an analysis of y/n questions as elliptical alternative questions where ellipsis affects the second (negative) disjunct (Bennett'77). I begin by noting that a y/n question containing 'even' has two different LFs depending on the scope of 'even' relative to negation in the ellipsis site (Wilkinson'96); an ambiguity that affects only its presuppositions, given that 'even' lacks truth-conditional import. Thus, in principle, (2) is ambiguous between the two readings A and B: A.LF1:[even you slvd [t.e.pr][f]]or[not[even you slvd [t.e.pr][f]]]
NOT>EVEN
B.LF2:[even you slvd [t.e..pr.][f]]or[even[not you slvd [t.e.pr.][f]]]
EVEN>NOT
I show, however, that the ambiguity is actually resolved in favor of reading B. This happens because both possible answers to A have an unsatisfiable presupposition, but one answer to B has a presupposition that is satisfiable. Let's now see how this comes about.
Presuppositions project from the LF of a y/n question to its answers (Bennett 77): the positive answer inherits the presuppositions of the positive disjunct, the negative answer those of the negative disjunct. This means that (3a), when uttered as an answer to (2) under reading B, presupposes (4). Similarly the affirmative answer (3b), presupposes (5). On the other hand, under reading A BOTH answers will presuppose (5). It is known that when the focus of 'even' denotes the lower endpoint of a scale the presupposition of 'even' can be satisfied only if 'even' takes scope over a negation as in (4) (Lahiri'98): there can be no context where (5) is true. This situation excludes reading A. Finally, an answer to B like (3b) is excluded: presuppositions' satisfaction in a context plays a crucial role in determining what counts as a possible answer to a question in that context. Specifically, only those sentences whose presuppositions are true in a context (possibly after accommodation) are good answers in that context. In the present case (3b) fails to satisfy the condition, no matter what the context is.
It follows that in all contexts, a question containing 'even'+the low endpoint of a scale (thus also a minimizer) has only one possible answer: a negative answer. If the set of answers is always a singleton, the question can never be used to seek information. That is why the question is felt to be rhetorical.
-EXTENSION TO WH-QUESTIONS As can be seen in (6a-d) the same phenomenon obtains in wh-questions. The account proposed for yes/no questions extends to these cases under one of the following assumptions 1.Wh- questions are sets of y/n questions (Higginbotham '93) or 2.Wh-phrases quantify over GQs.
-FURTHER IMPLICATIONS: My proposal yields an indirect argument in favor of Wilkinson's scope theory of even and supports an old view on y/n questions as concealed alternative questions (Bennett '77).
Data
(2) Q: Did you even solve [the easiest problem]f (only Negative Rhetorical)
(3) a. A': No, I didn't (even) solve [the easiest problem] [f].
b. A": Yes, I (even) solved [the easiest problem][f].
(4) Scalar Presupposition: (read 't.e.p'. as 'the easiest problem')
Ax [x ‚ t.e.pr. --> likelihood(that I didn't solve x) > likelihood (that I didn't solve t.e.pr.].
= 'the easiest problem is the MOST likely thing for me to have solved'
(5) #Scalar Presupposition:
Ax [x ‚ t.e.pr.--> likelihood (that I solved x) > likelihood (that I solved t.e. pr.)]
= 'the easiest problem is the LEAST likely thing for me to have solved'
(6) a. Who lifted a finger to help?
only Negative Rhetoricalb. Which of these girls would hurt a fly?
only Negative Rhetoricalc. Who has the least bit of taste?
only Negative Rhetorical... (see also data in Borkin'71)
d. Q: Which of your kids even solved [the easiest problem][f]?
only NR
A': Mary (even) solved the easiest problem.
#Scalar Presupposition:
Ax [x ‚ t.e.pr. --> likelihood (that M solved x) > likelihood (that M solved t.e.pr.)]
A'': No one (of my kids) (even) solved the easiest problem.
even [no one solved t [the easiest problem[f]]]
Scalar Presupposition:
Ax [x ‚ t.e.pr.--> likelihood (that no one solved x) > likelihood (that no one solved t.e.pr.)]
Selected References:
Bennett, M. (1997). A Response to Karttunen. In L&P 1. ,279-300.
Borkin, Ann (1971). Polarity Items in Questions. CLS
Heim, I (1984). A Note on Negative Polarity and DE-ness. In C. Jones and P. Sells (eds.) Nels 14.
Han,C.& Siegel,L(1996). Syntactic and Semantic Conditions on NPI Licensing in Questions. WCCFL 15.
Ladusaw (1979). Polarity Sensitivity as Inherent Scope Relation. Ph.D. dissertation, Austin.
Lahiri, U. (1998).Focus and Negative Polarity in Hindi. in Natural Language Semantics 6.1., 57-123.
Karttunen, F. & Karttunen, L. (1977). Even Questions. Nels 7, GLSA, UMASS Amherst.
Progovac, L (1993). Negative Polarity and Grammatical Representation. L&P 10
Rullmann, H. (1996). Two Types of Negative Polarity Items. In NELS 26. 335-350.
Schmerling, S. (1971). A Note on Negative Polarity. Papers in Linguistics 4.1.,
Sugizaki, H. (2001). Another Chinese-Type of Question. Ms UCONN.
Wilkinson, K. (1996). The scope of Even. Natural Language Semantics 4.2., 193-215.
MIT RLE Speech Communication Group Seminar
Lisa Lavoie
MIT
"Subphonemic Consonant Variation" Close spectral analysis reveals that consonants vary more than even trained listeners typically detect during impressionistic listening. Consonants respond phonetically to a range of conditioning environments: with hyperarticulation under stress (De Jong 1995), reinforcement in prominent prosodic positions (Fougeron 1998), reduction in colloquial speech (Brown 1990, Kohler 1990), and reduction in highly predictable words (Jurafsky et al. 1998).
The purpose of this study is to argue, based on my own research, for the theoretical relevance of careful phonetic investigation of consonant variation. This is not intentional categorical variation, but stochastic variation as Pierrehumbert (1994) describes. I demonstrate, based on acoustic studies of 20 American English and 17 Mexican Spanish consonants in read speech, that consonants display both random and environment-specific variation in degree of constriction. Stops do not always have complete seals nor do they always have release bursts. Fricatives do not always have concentrated noise. Though some consonants (sibilants, Spanish voiceless stops) are very stable, over 10% of the tokens of English /k, d, g, dZ, th, v, dh, z, zh/ and Spanish /tS, f, x, |, B, D, V, j/ are realized with an unexpected, that is, not underlying, manner of articulation.
In unstressed or medial positions, consonants may receive more sonorous realizations; in stressed or initial positions, they tend toward less sonorous realizations. Depending on environment, in English we may find voiceless stops realized as continuants; voiced stops as approximants; voiceless fricatives as stops; and voiced fricatives as stops, approximants, or glides. In Spanish, we may find voiceless fricatives realized as approximants; and voiced obstruents as stops, approximants, or glides, though seldom as true fricatives.
I argue that, as with vowels (Manuel 1990), the consonant inventory constrains possible variation, voiceless velars being a prime example. Spanish /k/ is a true stop 93% of the time but English /k/ is only 79% of the time. Lacking /x/, English is free to realize /k/ as a continuant, while still maintaining the percept of a stop. Spanish /x/ prevents /k/ from infringing on /x/'s phonetic space. Moreover, /x/ is never realized as a stop, but instead as an approximant.
Studies of consonant variation contribute both to phonetic and phonological theory, providing data to test a range of hypotheses. English and Spanish /k/ illustrate the role of inventory in variation. Consonant variation is a potential diagnostic of speech rate or style. We can determine if consonants and vowels reduce in tandem, testing the linkage between consonantal and vocalic tiers. Subphonemic consonant variation stands to offer rich theoretical rewards.
MIT LING-LUNCH
Friday, September 28, 2001 at 3:00, MIT E39-335 (conference room)John Frampton
Department of Mathematics, Northeastern University"Reduplication in Ndebele: Copying, Truncation, and Their Interaction" Zoll and Inkelas (2000) analyze unintensive reduplication in the Bantu language Ndebele. Ndebele produces an impressive display of optionality, with some underlying forms having 5 different alternate surface forms. The problem is to satisfactorily account for this variation. Zoll and Inkelas present the analysis as a showcase example of their analysis of reduplication as morpheme doubling and a refutation of Prince and McCarthy's Correspondence Theory of Reduplication. I will present a much more straightforward analysis, using a development of Raimy's theory of reduplication that I call Distributed Reduplication (DR). Looked at correctly, Ndebele unintensive reduplication is quite similar to much-studied reduplicative processes in Mokilese and Ilocano. DR gives straightforward analyses of these processes as well.
The starting point for DR is a proposal about the way that copying and truncation are carried out and the way that the two operations interact. The string is first marked appropriately and subsequently the copying/truncation is carried out. Marking is carried out by inserting special junctures into the timing tier. Interestingly, DNA uses similar techniques in replication. Promoter proteins (acting like chemical junctures) bind to specific sites on DNA (acting like structure dependent juncture insertion rules) to mark the position for the initiation of replication. Stop codons mark the terminus. Other proteins are used to mark the initiation and terminus of splicing (truncation). Less spectacularly, but more widely known, repeat marks are inserted into musical scores to indicate future duplication (at the time of performance).
The important point for phonological reduplication is that marking the stretches of the timing tier for intended duplication and truncation is a separate operation from actual copying and truncation. Other operations can therefore intervene --- operations aimed at achieving a particular prosodic form, for example.
In addition to Ndebele, Mokilese, and Ilocano, DR will be illustrated for a range of other well-and-not-so-well-known examples.
GSAS Workshop on Comparative Syntax and Linguistic Theory at Harvard
Place: Linguistics Department Lounge, Boylston Hall, Fong Auditorium, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Reception immediately following the talk
and Victor Manfredi, Boston University"West African Adverbs" From the start of Minimalism, adverb placement has been the principal diagnostic of verb raising in Romance and Germanic (Pollock 1989, Jonas & Bobaljik 1993). In languages of the Benue-Kwa group of Niger-Congo (= Greenberg's Kwa plus Benue-Congo), adverbial syntax has by contrast played a minor role in syntactic research, due to a well-known typological fact: the absence of a large morphosyntactic category dedicated to adverbial modification. This fact also correlates with the small number of dedicated adnominal modifiers (nominal quantifiers as well as attributive adjectives) in the same set of languages, a property which has caused much grief to compilers of West African wordlists using Romance and Germanic metalanguages. Instead of categorial adverbs, Benue-Kwa employs a mix of complex predicators (e.g. serial verb constructions), subject-depictive verbal auxiliaries and adjunct nominals. Correspondingly, instead of attributive adjectives we find reduced relative clauses based on stative verbs, including many examples of light verb plus nominal constant (Welmers 1973, Madu.ka-Durunze 1990, Hale et al. 1995).
Despite the foregoing, let us consider the possibility of a uniform adverbial syntax for both West Europe and West Africa, concentrating here on the African part (cf. D»chaine & Uchechukwu 2001) but keeping in mind the phrase structures proposed by Cinque (1999) for European -ly/-mente adverbs. Between them, Igbo and Yoruba have two ways to build a phrase meaning 'go quickly' from a VP 'go' plus a noun 'urgency' or 'speed'. Option (1) puts the nominal modifier in a higher instrumental phrase, while option (2) hypothetically fronts the VP around the modifier. This fronting is apparently obligatory in Yoruba: (1b) is out. A third option (3), a serial construction, is a highly productive but syntactically unrelated paraphrase of the preceding two, as suggested by the noncognate lexical material and different categorial input, and we set it aside. (4) gives counterparts of (1) - (2) in which the manner is wh-questioned by a movement strategy. Crucially, modifier extraction is incompatible with fronting, such that -- remarkably enough -- Yoruba matches Igbo exactly in the question form: a light predicator of manner, with a literal meaning 'use', 'from' or 'do' is required in (4) in both languages. The data in (4) are complicated by independent factors. In Igbo (4a), wh-chains are sensitive to the internal complexity of wh-expressions (cf. Goldsmith 1981). In Yoruba (4b), the progressive auxiliary [n] cannot precede a light predicator like [s.e]: only the lexical meaning 'behave' is available for the order [... n s.e ...], but the reverse order is independently available because Yoruba progressive merges below T (cf. Oyelaran 1992, Dechaine 1992).
(5) is our analysis, with an Adverbial shell between Tense and VP. For joint reasons of interpretation and pronunciation, the null head of AdvP, indicated in (5a) by X, raises to T where it is realised as the light manner predicator. This raising is strikingly confirmed in Yoruba by the linear order of overt X before progressive [n]; the situation in Igbo is less clear because progressive involves both an aux and a nominalized VP, and in manner questions the aux part is suppressed. In (5b), X can remain null because the fronted VP has scope over the Adverb. The type of position to which VP fronts is an open question; what is important for the present discussion is the impossibility of wh-movement combined with fronting (5b). This can be ascribed to Minimality (or in archaic terminology, 'proper government', cf. Carstens 1986).
Throughout the data in (1) - (4), the (a) sentence is Igbo and the (b) sentence Yoruba; both progressive and past terminative versions are given. For typographic reasons, the orthographic subdot [.] follows the vowel, and the tone pattern is cited after the sentence.)
(1)a.
O ji o.hu.hu. e-je. [HLLLLLH]
3S hold urgency NOM-go
'S/he is going quickly'O ji-ri o.hu.hu. je-e. [HLLLLLLH]
3S hold-TNS urgency go-ASP
'S/he went quickly'(1) b.
*O mu kiakia lo..
3s hold speed go(2)a.
O. na e-je (n') o.hu.hu.. [HLLHLLL]
3S PROG NOM-go LOC urgency
'S/he is going quickly'O je-re (n') o.hu.hu.. [HLLLLL]
3S go-TNS LOC urgency
'S/he went quickly'(2) b.
O n lo. (ni) kiakia. [HHM(H)HHHH]
3S PROG go LOC speed
'S/he is/was going quickly'O lo. (ni) kiakia. [HM(H)HHHH]
3S go LOC speed
'S/he went quickly'(3)a.
O. na a-gba o.so. e-je. [HLLHH!HLH]
3S PROG NOM-V race.GEN nom-go
'3S is going quickly'O. gba-(a)ra o.so. je-e. [HLLHHLH]
3S V-TNS race go-ASP
'3S went quickly'(3)b.
O n sare lo. [HHHHM]
3S PROG run go
'3S is/was going quickly'*O sare n lo.. [HHHHM]
3S run PROG goO sare lo. [HHHM]
3S PROG run go
'3S went quickly'(4)a.
Kedu. ka o si e-je? [LHLHLLH]
wh COMP 3S from NOM-go
'How is s/he going?'Kedu. ka o si je-e? [LHLHLLH]
wh COMP 3S from go-ASP
'How did s/he go?'Kedu. etu o ji e-je? [LHLHLLH]
wh manner 3S hold NOM-go
'How is s/he going?'Kedu. etu o ji je-e? [LHLHLLH]
how COMP 3S hold go-ASP\
'How did s/he go?'(4)b.
Bawo ni o s.e/ti n lo.? [HMMHMHM]
how FOC 3S use/do PROG go
'How is/was s/he going?'
(cf. Abraham 1958, p. 98)Bawo ni o n s.e lo.? [HMMHHMM]
how FOC 3S PROG use/do go
'How is/was s/he behaving while going?'Bawo ni o fi/s.e/ti lo.? [HMMHMM]
how FOC 3S use/do/from go
'How did s/he go?'(5)a.
[TP [T] [AdvP speed [X] [VP]]]
(where "X" = null head of AdvP)(5)b.
[TP [T] [VP]i [AdvP speed [X] ti]]
(where "ti" = VP trace)References
Abraham, R.C. [1958]. Dictionary of Modern Yoruba. University of London Press.
Carstens, V. [1986]. Proper government in Yoruba. M.A. thesis, University of California, Los Angeles.
Cinque, G. [1999]. Adverbs & Functional Heads, a Crosslinguistic Perspective. Oxford University Press.
Dechaine, R.-M. [1992]. Inflection in Igbo and Yoruba. MITWPL 17, 95-119.
Dechaine, R.-M. & C. Uchechukwu. [2001]. Adverbs, modals and the syntax of modifiers in Igbo. Presented at the 11th Niger-Congo Syntax & Semantics Workshop, Universita' Ca' Foscari, Venezia, 2 June.
Goldsmith, J. [1981]. The structure of wh-questions in Igbo. Linguistic Analysis 7, 367-93.
Greenberg, J.H. [1963]. The Languages of Africa. Mouton, the Hague. (= IJAL 29.1).
Hale, K., U..P. Ihio.nu. & V. Manfredi. [1995]. Igbo bipositional verbs in a syntactic theory of argument structure. Theoretical Approaches to African Linguistics, edited by A. Akinlabi, 83-107. Africa World Press, Trenton, N.J.
Igwe, G.E. [1999]. Igbo-English Dictionary. University Press Ltd, Ibadan.
Ihio.nu., U..P. [1988]. Serialization and consecutivization in Igbo. Presented at the 2nd Niger-Congo Syntax & Semantics Workshop, Lexicon Project, M.I.T., Cambridge, Mass., 11 April.
Jonas, D. & J. Bobaljik. [1993]. Specs for subjects: the role of TP in Icelandic. MITWPL 18, 59-98.
Madu.ka-Durunze, O. [1990]. Igbo adjectives as morphologised relatives. Studies in African Linguistics 21, 237-50.
Moro, A. [2000]. Dynamic Antisymmetry. MIT Press, Cambridge, Mass.
Oyelaran, O. [1992]. The category AUX in Yoruba phrase structure. Research in Yoruba Language & Literature 3, 59-86.
Pollock, J.-Y. [1989]. Verb movement, Universal Grammar, and the structure of IP. Linguistic Inquiry 20, 365-424.
Welmers, W.E. [1973]. African Language Structures. University of California Press, Berkeley.
MIT LING-LUNCH
Thursday, September 27, 2001 at 12:00, MIT E39-335 (conference room)Michela Ippolito
MIT"Presuppositions and implicatures in counterfactuals" I propose a pragmatic account of temporally mismatched past subjunctive counterfactuals such as "If Jack had got married tomorrow, they could have had the outdoors reception". First, I'll show that in case of temporal mismatches, Past has to be interpreted as constraining the time variable of the accessibility relation. Second, I'll propose specific Felicity Conditions (presuppositions) for subjunctive conditionals and I will argue that there is a strict correspondence between the time of the accessibility relation and the time relevant for the felicity conditions. Specifically, if the time relevant for the selection of the worlds (quantified over by the conditional operator) is past, then the conditional's presupposition will talk about the speaker's past epistemic states. I'll support these new felicity conditions empirically and discuss their repercussions on the puzzle of presupposition projection. Third, I'll argue that the intuition that the antecedent of a mismatched past subjunctive counterfactual is impossible is a Gricean scalar implicature arising from a competition between presuppositions (and not assertions).
SELECTED REFERENCES
Heim, I. 1994 "Presuppositions Projection and the Semantics of Attitude verbs". Journal of Semantics.
Iatridou, S. 2000 "The Grammatical Ingredients of Counterfactuality", LI
Ogihara, T. 2000 "Counterfactuals, Temporal Adverbs, and Association with Focus". SALT 10.
MIT Phonology Circle
Michael Kenstowicz
MIT"Paradigm Contrast:
An Example from Arabic"The phonology of base#clitic constructions in various colloquial dialects of Arabic has figured prominently in discussions of cyclic effects ever since it was first observed by Brame 1974. See recent papers by Kager 1999, Kiparsky 2000, among others. However, there is one aspect of the construction that has been overlooked in the generative literature. The 3 sg. fem. perfect of the verb (e.g. 9=E1llam-et 'she taught') when followed by a vowel-initial clitic is the locus of various phonological quirks in many Arabic dialects: unmotivated stress shift, gemination, lengthening, allomorphy. In his discussion of the Damascus dialect ,T.F. Mitchell (1993) proposed that stress shifts in the 3 sg. feminine 9allam-=E9t#o 'she taught him' in order to prevent the general process of open-syllable syncope of schwa from applying because to do so would merge this form of the verb with the 1 sg./ 2 sg.m. 9all=E1m-t#o 'I taught him'. If this is the correct interpretation of the phenomenon, it implies that the phonology of one form of the paradigm cannot be calculated in isolation from the others.
More importantly, the effect cannot be formalized in terms of the traditional cycle because neither the 3 sg. fem. nor the 1 sg. is contained in the other: each is formed by adding phonologically distinct suffixes (typically -et vs. -t) to a common base. In this paper we review Mitchell's analysis of the Damascus dialect and then report the preliminary results of a survey of other dialects.
MIT Linguistics Colloquium
Paul Kiparsky
Stanford University"Stems, Words, and Phrases: Stratal or Parallel OT?" Inflected words in Finnish show a range of interdependent stem and suffix alternations, many in systematic free variation. The phonological constraints that interact to produce the alternations are conditioned by syllable structure and stress at the stem level, masked at the word level by rhythmic stress and resyllabification. I present an analysis of these data which supports stratal OT over non-stratal, fully parallel OT.
MIT LING-LUNCH
Thursday, September 20, 2001 at 12:00, MIT E39-335 (conference room)Danny Fox
MIT"Antecedent Contained Deletion and the Copy theory of Movement" Antecedent Contained Deletion raises a problem for theories of ellipsis, which, according to much literature, is resolved by Quantifier Raising. The resolution, however, conflicts with the copy theory of movement. This paper argues that this conflict is resolved with the aid of a theory of extraposition and covert movement proposed by Fox and Nissenbaum (1998), when accompanied with certain assumptions about the structure of relative clauses and the way chains are interpreted. The resolution makes various new predictions and provides an account of a wide range of otherwise puzzling facts.
MIT RLE Speech Communication Group Seminar
Chris Halpin
Massachusetts Eye and Ear Infirmary
"Clinical Speech Reception Modelling in the Damaged Cochlea" Audiograms are often thought of as if they were singular elements (ie microphones) with sub-optimal frequency responses. Such a view also implies a single relationship between severity and speech intelligibility as well. Also, hearing loss is often modeled as the disruption of the normal strategies used to process speech, with little attention paid to alternate abilities (not used by normals) which are uncovered by disease. These assumptions result in both diagnostic and hearing aid fitting concepts which do not seem to work well in the clinic. Conversely, if the standard clinical tests are re-evaluated with the assumption that the cochlea is an array of many receptors (some damaged), different and more clinically useful results are seen.
Boston University Linguistics Talk
co-sponsored by:
- the James Geddes, Jr., Lecture Series
- the Program in Applied Linguistics
- the American Sign Language Linguistic Research Project
Monday, September 17, 2001 - 7:30 PM - CAS (725 Commonwealth Ave.), room 222
Talk cancelled
Because of the tragedy on 9/11, Zucchi's plane to the US was forced to return to Italy after he had gotten part-way across the Atlantic, and he will not be able to get another flight in time to give the talks that had been scheduled in this country. We will try to reschedule for later during this academic year.
Alessandro Zucchi
Università di Salerno
Dipartimento di Scienze della Comunicazione"Tense, Time, and Adverbs in Italian Sign Language"
Refreshments will be served. ASL/English interpretation will be provided.
Second Edition of the MIT & Harvard Student Conference on Language Research
Invited speakers:
- Lila Gleitman (Penn)
- Morris Halle (MIT)
- Jay Jasanoff (Harvard)
Conference proceedings will appear with the MIT Working Papers in Linguistics.
Conference webpage:
THE CAMBRIDGE PROSODY GROUP
The Cambridge Prosody Group presents a tutorial for those interested in intonation (for any reason)...
** So that we can determine which room(s) we will use for the tutorial please reply to lisa@speech.mit.edu to let me know if you are planning to attend. **
Led by Laura Redi
MIT"TOBI PROSODY LABELING TUTORIAL AND HANDS-ON SESSION"
ToBI (tone and break indices) labeling is a widely-used system of prosody labeling that is based on Pierrehumbert's 1980 dissertation and later work by Pierrehumbert and Beckman. Numerous data bases have been ToBI-labeled here at MIT and elsewhere. Before the tutorial session, we recommend that you read two things:1. From Guidelines for ToBI Labelling (1994) by Mary E. Beckman & Gayle M. Ayers, the Preface, Overview and ToBI annotation conventions.
Available on the web at: ftp://julius.ling.ohio-state.edu/pub/phonetics/TOBI/ToBI/main.html
2. D. Robert Ladd (1992) 'An introduction to intonational phonology' in Papers in Laboratory Phonology II: Gesture, Segment, Prosody. eds. Docherty & Ladd, Cambridge University Press. [Let me know if you can't get your hands on this; we can probably circulate a few copies.]
Laura will give about an hour's worth of presentation and then we will move to computers to try out the ToBI exercises.
MIT LING-LUNCH
Thursday, September 6, 2001 at 12:00, MIT E39-335 (conference room)Shinichiro Ishihara and Ken Hiraiwa
Massachusetts Institute of TechnologyMissing Links: Cleft, Sluicing, and "No da" Construction in Japanese [Introduction]
Cleft construction, exemplified in (1), is one of the well-known constructions, but no consensus about its syntactic derivation has been established.
(1) Taro-ga tabeta no wa kono-ringo-o da
Taro-NOM ate Nominalizer(NL) TOP this-apple-ACC COP
'It is this apple that Taro ate./What Taro ate is this apple.'
Koizumi (1995, 2000) analyzes clefts as a result of the VP (or TP) remnant created by V-to-C movement. His proposal, however, does not clarify the derivation in detail. In this study, we make an explicit proposal for cleft construction in Japanese. We present a set of new data that is problematic for Koizumi's analysis, and show that our analysis provides an explanation for it, while it still captures what Koizumi's analysis tries to capture. Furthermore, our analysis reveals certain missing links that relate a few other constructions in an interesting way.
[Koizumi's (1995, 2000) Analysis of Clefts]
Koizumi proposes that Japanese has V-to-C movement. As a result, neither VP nor TP contains the head verb itself. He claims that the element that appears in the focus position of a cleft sentence (before the copula 'da') is actually this remnant VP (2) or TP (3), not a DP itself. This analysis easily captures the fact that Japanese allows so-called multiple clefts:
(2) Taro-ga tabeta no wa [VP kono-ringo-o t(tabeta)] da
Taro-NOM ate NL TOP this-apple-ACC COP
'It is this apple that Taro ate. / What Taro ate is this apple.'
(3) tabeta no wa [TP Taro-ga kono-ringo-o t(tabeta)] da
ate NL TOP Taro-NOM this-apple-ACC COP
'*It is Taro, this apple that ate. / *Who ate what is Taro, this apple.'
Also, his analysis explains why multiple cleft is subject to so- called "clause-mate condition", that is, phrases in the focus position must originate the same clause. Since he claims that phrases in the focus position is in fact a remnant VP/TP, all the phrases must be inside a single remnant VP/TP.
Under this analysis, however, it is not at all clear how the remnant VP appears in this position, and how the topicalized clause show up, guided with a nominalizer 'no'.
[Proposal]
Our analysis can be roughly described as follows: We first claim that clefts are derived by certain syntactic movements from so- called "no da" construction, another focus construction in Japanese. The "no da" construction is a sentence in which a whole TP is followed by a nominalizer no followed by a copula da.
(4) [TP Taro-ga kono-ringo-o tabeta] no da
Taro-NOM this-apple-ACC ate NL COP
'(Lit.) It is that Taro ate this apple.'
From the underlying structure (4), the phrase that is to be focused moves to the SPEC of C by an A-bar focus movement. More specifically, assuming a version of Rizzi's (1997) split-CP system(5), the to-be- focused phrase moves to SPEC FocP, which is headed by the copula 'da'.
(5) TopP
/ \
(Topic)-wa / \
FocP Top
/ \
(Focus) / \
FinP Foc('da')
/ \
TP Fin('no')
/____\
After the focus movement of the to-be-focused phrase, the nominalized TP, i.e., TP plus a nominalizer 'no' (FinP in (5)), is topicalized and realized with a topic marker '-wa'. This is done by a topic movement to the SPEC of TopP, which is right above the FocP under the split-CP system. Our analysis assumes multiple application of focus movement for the multiple clefts, instead of remnant VP/TP. As for the clause- mate condition, it is interesting to note that this condition is also observed in the multiple application of so-called A-bar scrambling. Therefore, we can say that the clause-mate condition is a property of A-bar movement in general.
(6)* sensei-ni kono-ringo-o [Hanako-ga t [Taro-ga t tabeta to]
iituketa]
Since our analysis assumes A-bar focus movement, it is natural that we observe the same condition for clefts.
This analysis provides natural explanations for the unclear points of Koizumi's remnant VP/TP analysis of clefts. The focused phrase appears right before the copula by focus movement followed the topicalization of the nominalized TP. Since the topicalized constituent is the nominalized TP, the nominalizer 'no' appears right before the topic marker.
[Lack of Clause-mate Condition with WH-phrases]
Furthermore, it is important to note that wh-phrases do not follow the clause-mate condition.
(7) a. Hanako-ga dare-ni [Taro-ga nani-o tabeta to] iituketa no?
Hanako-NOM who-DAT Taro-NOM what-ACC ate that told Q
'Who did Hanako tell that Taro ate what?'
b. dare-ni nani-o [Hanako-ga t [Taro-ga t tabeta to] iituketa no?
This is true for multiple clefts as well:
(8) Hanako-ga t [Taro-ga t tabeta to] iituketa no wa
Hanako-NOM Taro-NOM ate that told NL TOP
dare-ni nani-o na no?
who-DAT what-ACC COP Q
'*Who what is it that Hanako told that Taro ate?'
It is clear that Koizumi's analysis cannot explain the lack of the clause-mate condition with wh-phrases in a straightforward way, because there is no way to make the non-clause-mate clefts under the remnant VP/TP analysis. Our analysis, on the contrary, is still compatible with the data. We will show that the lack of clause-mate condition is due to a phenomenon called deaccenting, which obligatorily takes place after wh-phrase in Japanese.
[Sluicing]
Our analysis can be extended to the sluicing construction. In Japanese, multiple sluicing is possible, just like clefts, while sluicing does not exhibit the clause-mate condition, unlike the cleft. We claim that the sluicing is also derived from the no da construction. After the focus movement of the wh-phrase(s), the nominalized TP (FinP) in the sluicing is elided instead of being moved to the SPEC TopP. This analysis explains the fact that Japanese allows multiple sluicing, as well as the fact that it does not exhibit clause-mate condition, since sluicing involves (possibly multiple) focus movement, and the wh-phrases are exempt from the clause-mate condition due to deaccenting. Our analysis provides a straightforward account for the missing link between the two constructions.
University of Massachusetts, Amherst
Herter is the seven story classroom building which forms an L with the Fine Arts Center at the south edge of campus. They are both part of a U-shaped arrangement of buildings around a central grassy expanse called Haigis Mall, which sits, with flags at the end, alongside the road that both leads into the center of Amherst and which leads to Rte. 116, the latter giving access to the campus from Rte. 9. General driving directions to U Mass Amherst are available on their web site (www.umass.edu).
Bob Ladd
University of Edinburgh"SEGMENTAL ANCHORING OF INTONATIONAL TARGETS: IMPLICATIONS FOR PHONOLOGY"
THE CAMBRIDGE PROSODY GROUP
Presentation and Discussion led by
Lisa LavoieI will present some techniques for eliciting spontaneous speech and the kinds of speech that results. Then we will discuss/brainstorm other ideas for eliciting natural speech. If you have heard about an interesting way of eliciting speech, please plan to share the details.
Everyone is welcome -- please forward this notice to interested colleagues.
DIRECTIONS TO BUILDING 34MIT's Building 34 is located at 50 Vassar St. in Cambridge, between Mass. Ave. and Main Street. Street parking may be available in the afternoon, but give yourself plenty of time to find it. Also note that there is a huge amount of road construction at the intersection of Mass Ave. and Vassar St. so you might not be able to turn exactly where you want.
Via public transportation, take the red line to Kendall Sq. Upon exiting the station, walk up Main St. away from the river. [When you're facing away from the river, the MIT Coop will be on your right and the MIT Press Bookstore on your left.] When Main St. splits, veer left onto Vassar St. and walk along the huge construction site in the narrow chute created by the cement barriers (the Vassar St. sign has disappeared). Building 34 is the first building on the left after the construction.
You could also take a bus, such as the #1, to 77 Mass Ave. (main entrance to MIT) and walk away from the river, making a right on Vassar St.
For further information, contact lisa@speech.mit.edu.
THE CAMBRIDGE PROSODY GROUP
Duane Watson
Dept. of Brain and Cognitive Sciences, MITOur intent in starting this group is to provide a semi-regular forum for researchers who are working on aspects of prosodic structure. We hope to encourage informal networks of collaborators so that knowledge can be spread and shared.
Our first meeting on July 12 will consist of a short presentation and discussion. We will likely also find out what attendees might like to get out of such a group.
Future meetings may consist of presentation, tutorial, focused discussion, or some combination thereof.
DIRECTIONS TO BUILDING 34MIT's Building 34 is located at 50 Vassar St. in Cambridge, between Mass. Ave. and Main Street. Street parking may be available in the afternoon, but give yourself plenty of time to find it. Also note that there is a huge amount of road construction at the intersection of Mass Ave. and Vassar St. so you might not be able to turn exactly where you want.
Via public transportation, take the red line to Kendall Sq. Upon exiting the station, walk up Main St. away from the river. [When you're facing away from the river, the MIT Coop will be on your right and the MIT Press Bookstore on your left.] When Main St. splits, veer left onto Vassar St. and walk along the huge construction site in the narrow chute created by the cement barriers (the Vassar St. sign has disappeared). Building 34 is the first building on the left after the construction.
You could also take a bus, such as the #1, to 77 Mass Ave. (main entrance to MIT) and walk away from the river, making a right on Vassar St.
For further information, contact lisa@speech.mit.edu.
MIT 3rd Formal Approaches to Japanese Linguistics (FJAL) Conference
Invited speakers are: Noam Chomsky,
Masatoshi Koizumi,
Mamoru Saito and
Anna SzabolcsiThe program is available at: http://web.mit.edu/linguistics/www/FAJL3/program.html
Go to: http://web.mit.edu/linguistics/www/FAJL3/registration.html
MIT LING-LUNCH
Thursday, May 17, 2001 at 12:00, MIT E39-335 (conference room)Christopher Bader,
Massachusetts Institute of TechnologyTBA
MIT - WORKSHOP ON ALTAIC LANGUAGES
Thursday, May 17, 2001 at MIT Student Center, 307 Mezzanine Lounge (Student Center is W20, directly across Mass. Ave. from the Mass. Ave. entrance to MIT; see map of MIT at http://whereis.mit.edu/)PROGRAM
9:00 Opening Remarks9:10 Engin Sezer, Harvard University
"Finite inflection in Turkic: A comparative perspective"9:50 Nadya Vinokurova, Utrecht University
"Categorizing bare roots in Sakha"10:30 - 10:50 BREAK
10:50 Paul Hagstrom, Boston University
"Some implications of child errors for the structure of Korean negation"11:30 Judy Baek, Ken Wexler, MIT
"The unique checking constraint and the development of Korean negation"12:10 - 2:10 LUNCH
2:10 John Whitman, Cornell Univesity
"Korean and Japanese: Cognate morphology and how it has diverged"2:50 Ken Hale, MIT
"On the Dagur object relative: some comparative notes"(organizers' note: will include playing of field work tape on Dagur in Chinese)
3:30 - 3:50 BREAK
3:50 Meltem Kelepir, MIT
"Subject & object positions and scope in Turkish"4:30 Jaklin Kornfilt, Syracuse University
"Subjects and their Case in Turkish/Turkic embeddings"5:10 Shigeru Miyagawa, MIT
"Word order options in Japanese, Korean, and Turkish -- options without optionality"5:50 CLOSING
MIT RLE Speech Communication Group Seminar
Frank Guenther
Cognitive and Neural Systems, Boston University
"Neural Modeling of Speech Perception and Production"
For current details of our entire spring seminar series, please refer to http://web.mit.edu/speech and click 'Upcoming Seminars'.
Linguistics Department Lecture at Harvard
Place: Linguistics Department Seminar Room, Boylston 303
Prof.Jay Jasanofff
Harvard University"Why *woide?"
MIT LING-LUNCH
Thursday, May 10, 2001 at 12:00, MIT E39-335 (conference room)Susan E. Kalt
Dept. of Foreign Languages and Literatures, MIT"The Interpretation of Reflexive and Oblique Clitics in L2 Quechua Spanish" I used a picture selection and description task to evaluate how successful the two groups are at interpreting object clitics at various stages of acquisition, and to determine how closely these children follow the developmental path of other children acquiring similar elements in Spanish as a first language, as well as other first and second languages. The Full Access hypothesis of second language acquisition (Epstein, Flynn and Martohardjono 1996, Schwartz and Sprouse 1996, White 1996 among others) hold that L2 acquisition is inherently the same process as L1 acquisition, deriving from the same language acquisition device.
In previous studies (Deutsch, Koster and Koster 1986, Padilla 1990, Crysmann and Muller 2000), sentences involving reflexive anaphora were interpreted successfully earlier than sentences involving non-reflexive anaphora. Explanations offered for the earlier success with reflexives include: (1) Reflexives involve fewer sentential arguments than obliques (2) Binding of reflexives is resolved in a more local domain than obliques I observe a third possible explanation related to this particular task: (3) Unlike reflexive and no clitic sentences, obliques require binding to an argument identified only in the discourse; children tend to prefer an explicitly-mentioned bindee over a non-explicit one (D, K&K 1986)
Findings: L2 Southern Quechua Spanish speakers resemble L1 acquirers of Spanish, L1 acquirers of Dutch, and bilingual acquirers of French and German, in that reflexive and non-reflexive elements are interpreted with significantly different success from one another at various stages of acquisition. Successful interpretation of reflexive and oblique 3P clitics increases with age, and comes earlier for reflexives than for obliques.
Monolingual Spanish results did not pattern with the others. Performance on oblique and no clitic structures appeared to be at ceiling from the earliest stage tested, and target-like interpretation of reflexives decreased with age at the latest stage tested. Error scores and production data will be discussed in this regard.
In addition to investigating how L2 Quechua Spanish relates to widely attested patterns in development, my study explores the role of Quechua morpho-syntax (if any) in determining the acquisition trajectory. The hypothesis that L2 functional feature values are limited to those defined in L1 (Hawkins and Chan 1997, Hawkins 1998) requires us to examine properties of 3P objects in Quechua.
Oblique 3P objects are null in Quechua while they are marked overtly in Spanish, but both languages mark reflexives overtly:
Southern Quechua Standard Spanish
1a) Ana (Jusi-man) jut'a-ta chura-0-n 1b) Ana le pone la sandalia (a Jose)
Ana (Jose-DAT) sandal-ACC put-3OBJ-3SUBJ Ana 3DAT puts the sandal (to Jose)
2a) Ana jut'ata chura-ku-n 2b) Ana se pone la sandalia
Ana sandal-ACC put-REF-3SUBJ Ana 3REF puts the sandal
Despite its non-overt status in morphology, the 3P oblique object in 1a) receives definite/specific interpretation. An interesting question emerges: do Quechua children insert a null 3P object pronoun in Spanish, as claimed in S·nchez (2000), and do they produce oblique target-like 3P clitics in their L2 at various stages?
Further evidence of potential L1 influence will be examined in the error scores and production data.
UMass, Amherst: Linguistics Colloquium
Young-mee Yu Cho
"A Historical Perspective on Nonderived Environment Blocking"
MIT Linguistics Colloquium
Jason Merchant
University of Chicago"Ellipsis and Repair Phenomena" The deletion view, however, brings with it a number of puzzles; in this talk, I concentrate on three of these, which fall into two general classes: instances of usually possible movements that are impossible under ellipsis, and impossible movements that become possible only under ellipsis. The first class includes I-to-C movement in matrix questions in the Germanic languages, which is impossible under sluicing, and which I argue provide support for a particular version of the strong lexicalist hypothesis. The second class includes island-violating wh-movement under sluicing and the obligatory presence of VP-ellipsis in comparatives with I-to-C movement. In both cases, I argue that the nature of the problem rests with certain ill-formed intermediate traces of movement.
MIT LING-LUNCH
Thursday, May 3, 2001 at 12:00, MIT E39-335 (conference room)Matt Shibatani
Center for Advanced Study in the Behavioral Sciences
Kobe University"On Non-Canonical Constructions"
A linguistics open house
for undergraduates
interested in linguisticsin Photonics (8 St. Mary's St.) room 203
to be followed by a talk by
Robert Hoffmeister, Boston University at 7:30
on Language and the Theory of Mind in Deaf Children
in CAS 226 (note room change)
For details, click here.
MIT RLE Speech Communication Group Seminar
Khalil Iskarous
Haskins Laboratories
"The Acoustics of Gestural Overlap"
Boston University Human Development Colloquium
Paul Bloom
Yale University"How children learn the meanings of words"
MIT The eighth meeting of AFLA
(the Austronesian Formal Linguistics Association)
Friday, April 27 - Sunday, April 29, 2001 MIT
Invited speakers are Sandra Chung, Melenaite Taumoefolau, Whitney Postman, and Paul de Lacy.
More information, including a program, can be found at http://mit.edu/linguistics/www/AFLA8/.
Communication & Culture Seminar
Friday, April 27, 2001 at Barker Center, 12 Quincy Street, Cambridge, Massachusetts 02138: room 133
"Ethnomethodological Approaches to Learning and Instruction"
The April meeting of the Communication & Culture Seminar Series will be an all day mini-conference. Following is the schedule for the day's events:
9:00-12:30: Morning Session (papers/talks):Doug Macbeth (Ohio State University): "Reflections on 'reflexivity' in educational studies"
Michael Lynch (Cornell University): "Following instructions', membership, and blame: an alternative ethnomethodological treatment of institutional discourse."
Reed Stevens (University of Washington): "Within and across moments: the help ethnomethodology provides for respecifying 'learning'"
12:30-2:00: Lunch
2:00-5:00: Afternoon Session (Demonstration/Working Sessions)
Dusan Bjelic (University of Southern Maine) and David Bogen (Emerson College): "Praxitoria: the incarnate and practical hermeneutics of demonstrably rational action"
(Demonstration of praxiological orders involving audience participation)
Robert Sanders (SUNY-Albany) and Anita Pomerantz (SUNY-Albany) (Roundtable discussion with reference to video-taped materials of children playing with Lego blocks)
The Communication & Culture Seminar Series explores the cognitive, sociological, anthropological, linguistic, and philosophical dimensions ofdiscourse in social contexts. Meetings feature presentations by invited speakers on topics ranging from theoretical and methodological discussions to studies based on empirical research. The Communication & Culture Seminar is open to the public, and anyone with an interest in these areas of scholarship and research is encouraged to attend.DIRECTIONS TO THE BARKER CENTER: The Barker Center is in the main quad at Harvard, just off of Harvard Square. It is bordered by Prescott, Harvard, and Quincy Streets. Enter the courtyard from Quincy Street. The Barker Center is a large brick building facing the faculty club.
PARKING: Parking is available at no charge, on a space available basis, at the Broadway Garage, located on Felton St. between Cambridge St. and Broadway. All parkers should identify themselves as participants in the Communication and Culture Seminar at the Humanities Center. If there are no spaces available the guard will direct you to another Harvard parking facility.
For further information about Communication and Culture Seminar contact: David Bogen, Communication & Culture Co-Chair DBogen@emerson.edu or George Psathas, Communication & Culture Co-Chair geops1@bu.edu
For more information about the Center and its programs or to be put on the CLCS mailing list go to: http://www.fas.harvard.edu/~clcs
Boston University Graduate Program in Applied Linguistics
Spring Speaker Series 2001:
Topics in Computational Linguistics
Rukmini Iyer
Manager of Speech Engineering, Speech Solutions Group, GTE/BBN Industries"Modeling Domain Knowledge in Speech Recognition Applications"
Ford Foundation GSAS Workshop on Syntax at Harvard
Place: Sever Hall 202, Harvard University
Map: http://www.map.harvard.edu/level2/2Yard.shtml
Prof.Edward Gibson
Department of Brain and Cognitive Sciences
Massachusetts Institute of Technology"Syntactic integration difficulty in language comprehension"
UMass, Amherst: Linguistics Colloquium
Gita Martohardjono
"Tense and Aspect in Second Language Acquisition"
MIT LING-LUNCH
Thursday, April 26, 2001 at 12:00, MIT E39-335 (conference room)Rita Manzini
Universitý degli Studi di Firenze"Primitives of C-HL: A restrictive theory of morphosyntactic categories (features)" (1) [N l ... [Q i ...
The discussion of so-called datives in fact allows us to make the point that it is not just the notion of Case that is called into question. By analyzing (1) as we do, we imply that Number is also a descriptive artefact; there is no Number but only Q properties which may subsume plurality.
Let's draw some conclusions. First, even in instances where one commonly speaks about morphological nominative, accusative and dative, it is clear that the morphology in no way reflects an independent underlying primitive of `abstract Case'; it reflects strongly motivated independent divides such as those between D, Q and N categories. Second, there truly are no separate realms of syntactic categories and morphological features; in the syntax we are used to think of Q as hosting one as well as two, three, many, but so evidently it is in the morphology. The 'i' formative in (1) corresponds to a category that covers plurality but does not in any way identify with it. In both cases the conclusions to be drawn correspond to considerable simplifications with respect to current views; which clinches the case in their favor.
The key restriction seem to be that each lexical element - hence in (1) the formatives 'l', 'i' - belongs to a single category and that categories are not multivalued (i.e. they do in fact coincide with what are traditionally called categories and not features). In the rest of the paper we shall concentrate on the consequences of this for a minimalist-type computational system.
Harvard University
Talk sponsored by The Linguistics Department, the Department of Germanic Languages and Literatures, and the Committee on Folklore and Mythology
Place: Kates Room on the second floor of Warren House. Warren House is located between the Barker Center and the Faculty Club; it houses the Department of Celtic Languages and the Committee on Folklore and Mythology.
Kristjan Arnason
Professor of Old Icelandic
University of Iceland (Reykjavik)"Generative Metrics and Germanic Poetry"
MIT RLE Speech Communication Group Seminar
Tecumseh Fitch
Department of Psychology, Harvard
"How Speech is Special II:
More Data from Nonhuman Primates"
Harvard University
Place: Yenching Institute, room 18, 2 Divinity Avenue, Harvard University
Prof. Lord Renfrew of Kaimsthorn
McDonald Institute for Archaeological Research, Cambridge University, UK" Towards a Prehistory of Languages:
Archaeology, Genetics, and Linguistic Diversity."
MIT RLE Speech Communication Group Seminar
Gloria Waters
Boston University
"How Specialized is Working Memory?
Evidence from Normals, Brain-Damaged Patients and Neuroimaging"Working memory is considered to be a specialized component of memory that is responsible for the temporary storage and manipulation of information necessary to accomplish a cognitive task. Language processing is an example par excellence of a task that requires temporary storage and manipulation of information, and that therefore provides a domain in which ideas about working memory can be explored. In this talk, I will review our current studies of the involvement of the working memory system in language processing. These studies suggest a model in which there are several mechanisms that are responsible for temporary storage and manipulation of information related to different aspects of language processing.
Harvard University Indo-European Workshop
Place: 303 Boylston Hall, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Prof. Sylvia Luraghi
Universita di Pavia"Null Objects in Greek and Latin: Their Relevance for Indo-European Linguistics and for Language Typology"
UMass, Amherst: Linguistics Colloquium
Laura Benua
"Markedness and Harmony Preconditions"
Ford Foundation GSAS Workshop on Syntax at Harvard
Place: Linguistics Department Lounge, Boylston Hall, 3rd Floor, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Prof. Victoria Anne Lotridge
Department of Linguistics, Harvard"The Cautious Liaison - Where and Why it happens"
MIT RLE Speech Communication Group Seminar
Dav Clark
Dept. of Brain and Cognitive Sciences, MIT
"More is Better: Predicting Word Memory with "Phonological Loop" Activation"
UMass, Amherst: Linguistics Colloquium
Mark Baker
" On the Lexical Category Distinctions"
MIT Linguistics Colloquium
Gennaro Chierchia
Università degli Studi di Milano"Polarity phenomena and the syntax/pragmatics interface"
MIT LING-LUNCH
Thursday, April 5, 2001 at 12:00, MIT E39-335 (conference room)Professor Alison Henry
University of Ulster at Jordanstown"Expletives in Belfast English" (1) There have lots of students written the assignment already.
(2) There have some people built houses beside the lake.
and the availability of a range of positions for associates:
(3) There should have been lots of students completing the assignment by now.
(4) There should have lots of students been completing the assignment by now.
(5) There should lots of students have been completing the assignment by now.
(6) For some time there have seemed to be several students in a panic about the work.
(7) For some time there have seemed several students to be in a panic about the work.
(8) For some time there have several students seemed to be in a panic about the work.
These facts about Belfast English can, it will be argued, throw light on some current questions about the nature of expletive structures, the licensing of transitive expletives, and the nature of movement outside the narrow syntax. A number of proposals have been put forward about how transitive expletives are licensed, including proposals of a relationship to agreement, to verb raising to T, and/or to the possibility of object shift of full DPs. Belfast English however differs little from standard English in relation to agreement; it certainly does not have richer subject-verb agreement. Moreover, it does not have verb raising to T. It does have some evidence of object shift, but only for weak pronouns (a fact which only shows up in imperatives, the only structures where lexical verbs move out of VP). Thus it behaves much more like Mainland Scandinavian than Icelandic along the dimensions argued to license object shift in the latter but not the former, but nevertheless transitive expletives are possible. A careful consideration of Belfast English transitive expletives and their relationship to other aspects of the dialect is used to throw some new light on the debate about the licensing of these structures. Another aspect of Belfast English expletives which appears to be significant for current theory, and which unlike transitive expletives does not appear to have been widely studied in other language varieties, is the ability of associates to appear in different positions in the structure - in effect, as shown in (3) - (8) above, they can appear in any intermediate specifier position between the position where they appear in standard English, and the specifier position immediately below that occupied by the expletive. Examples like (7) appear to present problems for any analysis which seeks to exclude such structures in standard English through a universal constraint (such as a preference of Merge over Move) because there is clearly a variety which violates such a constraint. The 'movement' of the associate from its base position however shows some characteristics of being movement outside the narrow syntax: the moved associate must be stressed, and is much better if it is quantified, for example. Thus, (6) above is much better than (9) or (10), which are only grammatical with strong contrastive stress on the associates.
(9) For some time there have seemed students to be in a panic about the work (10) For some time there has seemed a student to be in a panic about the work
However, if this is movement in the phonology, it is clear that such movement cannot be as highly local or constrained as sometimes envisaged; in particular, there is some evidence that wh-movement can occur from a moved associate:
(11) How many have there of the students been asking for help?
(12) How good has there seemed of a guide to be showing people around?
It is argued that movement 'in the phonology' is in fact a type of movement subject to different constraints from movement in the narrow syntax, deriving from the need to satisfy information/discourse requirements, as distinct from the need to satisfy purely syntactic requirements, and as such syntactic boundaries/phases do not have the same role to play in constraining this type of movement.
MIT RLE Speech Communication Group Seminar
Zhiqiang Li
Department of Linguistics and Philosophy, MIT
"A Prosodic Account of Neutral Tone in Chinese
UMass, Amherst: Linguistics Colloquium
Marcel den Dikken
"A polar whole. Dutch heel 'whole' as a special kind of negative polarity item"
Boston University Graduate Program in Applied Linguistics
Spring Speaker Series 2001:
Topics in Computational Linguistics
Nanette Veilleux
Simmons College"Computational Approaches to Discourse Tracking:
using the Centering Paradigm"
MIT LING-LUNCH
Thursday, Mach 22, 2001 at 12:00, MIT E39-335 (conference room)Ileana Paul
MIT/UQAM"Malagasy extraction asymmetries" (1) a. ny vehivavy izay namaky ilay boky
det woman REL PST.AT.read def book
`the woman who read the book'
b. * ny boky izay namaky Rasoa
det book REL PST.AT.read Rasoa
`the book that Rasoa read'
(1b) shows that an object cannot undergo A-bar movement in Malagasy. Instead, the object must first be promoted to subject, as in (2). In (2), the verb is marked with Theme Topic morphology.
(2) ny boky izay novakin-dRasoa
det book REL PST.TT.read.GEN.Rasoa
`the book that was read by Rasoa'
Recent work on Malagasy accounts for the ungrammaticality of (1b) by invoking the A-bar status of the subject. In this talk, however, I argue that the ungrammaticality of (1b) is due to the fact that there is no derived object position in Malagasy. I also discuss Tagalog and Indonesian and suggest that the same account applies to the former, but not to the latter.
Data from binding show that the subject is an A position. The complex reflexive `ny tenany' can be the subject of a Theme Topic verb; (3) is often cited as evidence for reconstruction and hence for the A-bar status of the subject.
(3) Hajain'i Sahondra ny tenany.
TT.respect.GEN.Sahondra DET self.GEN.3
`Sahondra respects herself.'
Further data show, however, that this "reflexive" is not subject to Condition A (it can take a discourse antecedent, for example). Therefore (3) is not evidence for reconstruction. I also provide evidence from Condition C against the A-bar status of the subject.
To account for the lack of object A-bar movement, I draw on Travis (2000), who argues that Malagasy lacks a derived object position. For example, although Malagasy has apparent subject-to-object raising, it can be shown that the final position is not the object of the matrix clause.
(4) Mino azy(i) ho tsara tarehy i Koto(i).
AT.believe 3(ACC) PRT good face Koto
`Koto believes himself to be beautiful.'
A pronoun in the apparently derived object position can be coindexed with the matrix subject. This is not possible for standard direct objects, hence the pronoun in (4) is not in the direct object position. Travis cites further data from applicatives, possessor raising and object shift to support her claim.
I follow Travis and propose that the logical object is "trapped" within vP. Adopting Chomsky (1999), a [+wh] C probe cannot target the logical object as the logical subject is closer. Economy therefore dictates that an object can never A-bar move past a subject. For languages such as English that allow object wh-movement, the object must first adjoin to vP. From this position, the subject and the object are equidistant to C. In Malagasy, the only "escape hatch" for object extraction is the grammatical subject position, [Spec, TP]. The object must therefore be promoted to subject with Theme Topic morphology, as in (2).
Time permitting, I discuss Tagalog and Indonesian, languages which have similar extraction patterns to Malagasy. I show that Tagalog lacks derived objects (Maclachlan and Nakamura 1997). Hence the present analysis accounts for Tagalog. Indonesian, however, does have derived objects. I suggest that the A-bar analysis is appropriate for Indonesian (see Soh 1998 on Malay).
References
Chomsky, Noam. 1999. Derivation by phase. MIT Occasional Papers in Linguistics 18.
Keenan, Edward 1972. Relative clause formation in Malagasy. In: The Chicago which hunt: Papers from the Relative Clause Festival, 169-189. Chicago: CLS.
Maclachlan, Anna and Masanori Nakamura. 1997. Case-checking and specificity in Tagalog. Linguistic Review 14: 307-333.
Soh, Hooi-Ling. 1998. Certain restrictions on A-bar movement in Malay. In Recent papers in Austronesian linguistics. M. Pearson (ed). Los Angeles: UCLA.
Travis, Lisa. 2000. Derived objects in Malagasy. To appear in Objects and other subjects. W. Davies and S. Dubinsky (eds).
MIT RLE Speech Communication Group Seminar
Chris Halpin
Mass Eye and Ear Infirmary
"electing Patients for Speech Production Effects Based on Cochlea Damage Models and Hearing Aid Performance"
Harvard University GSAS Workshop on Syntax
Place: 303 Boylston Hall, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Claire Bowern
Department of Linguistics, Harvard"Bardi root faithfulness: or, Why I haven't learned to stop worrying and love paradigm uniformity"
MIT Linguistics Colloquium
Hagit Borer
University of Southern California""Plurals as Classifiers" b.All N predicates are mass, because nouns, as such, are encyclopedic items without any grammatical structure. Unless given structure, they will acquire none. Mass is simply a default interpretation in the absence of `count' structure (and likewise, atelicity is a default interpretation in the absence of `telic' structure)
c.English (and Hebrew, and many other languages) does have classifier morphology. It is called `plural'. The reason plural morphology and classifier morphology do not co-occur is because they range over the same functional value. So, contrary to common wisdom, plurality is not a number specification or a quantity specification, nor is it a function from singulars. Rather, it is a divisional function on mass. The difference between the classifier system of English and that of Chinese is that while both divide, the latter also defines a possible portion.
These claims, as it will turn out, will account in a straightforward way for the similarities, often observed, between (bare) mass nouns and bare plurals, including within the domain of (a)telicity; for the properties of singulars; for the distribution of so-called `plural' /s/ in English, and for a non-trivial range of interlanguage variations as regarding the behavior of singular expressions and cardinals. Finally, they will pave the way for a unified description of quantity which applies, homomorphically, within the nominal and the aspectual domain, overcoming many of the problems faced by Krifka's (1989, 1992, 1998) notion of quantized.
MIT LING-LUNCH
Thursday, March 15, 2001 at 12:00, MIT E39-335 (conference room)Linnaea Stockall
Massachusetts Institute of TechnologyTeal Bissell
Massachusetts Institute of Technology"Coalescence in Tangale" "Miya Pluractional Allomorphy" Abstract: Tangale phonology is characterized by a complete lack of any syllable internal consonant clusters (see Kidda 1986). In OT terms, *Complex is very highly ranked. Yet an independent process of final vowel deletion before a following morpheme creates many consonant clusters that would violate this constraint. Triggered by the high ranked *Complex constraint, the illicit structures undergo one of the possible repairs illustrated in (1): CC.CVC (a) or CC.C (b).
(1) a. [bagd+no] . /bagudno/ b. [kold+mu] . /kolmu/
Kidda (1985) proposes that the epenthesis strategy (1a.) is the default repair and that the repair in (1b.) is deletion, a last resort strategy employed when epenthesis is impossible. She argues that epenthesis is blocked in forms like (1b) due to a form of geminate integrity. Homorganic sequences are underlyingly linked to a single place node, therefore epenthesis, as in koludmu, would result in a line-crossing violation. Where epenthesis is blocked, the unsyllabifiable consonant is deleted.
I argue that the Kidda (1985) account is inadequate. Kidda's assumption of underlyingly double-linked homorganic Cs, wrongly predicts that all such sequences should pattern with the [koldmu] . /kolmu/ cases in (1b.) This is not the case as the alternation: [korn+go] . /korungo/ (*korgo) shows. As there is no principled reason to distinguish the two cases representationally the answer must lie elsewhere.
My account for this difference makes the repair strategy choice follow from basic properties of the segmental inventory of Tangale. I argue the following:
(A) Homorganic sonorant obstruent sequences trigger the 'deletion' repair, all other consonant clusters trigger epenthesis.
(B) The 'deletion' repair is preferred to epenthesis, where possible, because it is not deletion. An ld input sequence mapping to an l output is a case of consonant coalescence. Coalescence violates neither Max nor Dep and is hence more faithful to the input than either epenthesis or deletion.
(C) High ranked Max(feature) and Ident(feature) constraints determine which input features are realized in the coalesced output form. Ident(place) in Tangale restricts a coalescence repair to homorganic consonant sequences (coalescence of md.m, for example, would violate Ident(coronal)) and Max(sonorant) requires that all input sonorant segments map to sonorant segments in the output.
This account has the following desirable consequences: first, the choice between two competing repair strategies is understood as a preference for the fewest faithfulness violations, constrained by a restriction on correspondence relations (McCarthy 1995) -two segments must share a place node to correspond. Second, the precise nature of the coalesced output segment is determined by a markedness-as-faithfulness hierarchy (Pulleyblank 1997), prioritizing the retention of less marked sonorant codas over more marked obstruent codas. I argue that this simultaneously affords us a better understanding of the basic phonotactics of Tangale than previous accounts, and raises interesting questions about the nature of formal correspondence relations in OT.
Abstract: This talk investigates the Miya pluractional, which can be realized via apparent CV reduplication, vowel lengthening, or vowel change, depending on root shape. This talk argues that the CV doubling in (a-b) is not true reduplication, and that the pluractional prefix is an mora. The empirical basis for this claim comes from the unified analysis of pluractional allomorphy it affords. How the mora is realized is predictable by the phonological shape of the root: it is optimally realized as a prefix, when this violates minimal markedness constraints; however, when realizing it in this position creates a too greatly marked structure, it surfaces as vowel lengthening or vowel change. The mora is realized before C+C and Caa roots, as a prefix, in accordance with ALIGNR, an alignment constraint demanding prefixation. ONSET, which prohibits onsetless syllables, forces non-reduplicative copying of the root initial consonant, which is realized at the beginning of both the word and root, a violation of Integrity (no splitting). The apparent CV reduplication is therefore derived via the ranking ONSET, ALIGNR >> INTEGRITY, not via faithfulness between base and reduplicant (Correspondence Theory, McCarthy & Prince 1995).
Vowel quality of the prefix is also predictable. It is optimally realized as [a] in C+C roots, because a set of harmonically ranked constraints name [a] the best vowel, as the most sonorant. In Caa roots it is realized as schwa in order to avoid an OCP violation: prefixing [a] to Caa roots creates a configuration containing two identical vowels adjacent on the vowel tier. Provided the ranking OCP >> [Sonorancy Hierarchy], the prefix is realized as schwa in order to be distinct from the root vowel.
AlignR is violated in CaC and CVCC roots, in order to avoid marked structure. As in CV roots, prefixing [a] to CaC roots creates an OCP violation. In effort to avoid this illicit configuration, the moraic morpheme is realized by lengthening the root vowel. Caa roots do not exploit this strategy because the resulting form would not be distinct from the input of the root. AlignR is violated in CVCC roots in effort to satisfy a set of prosodic structure constraints (ProsCon) which limit verb size to 3 moras. A ranking of PROSCON >> ALIGNR, MAX forces the pluractional morpheme to occur root internally.
Harvard University GSAS Workshop on Syntax
Place: 303 Boylston Hall, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Balkiz Ozturk
Department of Linguistics, Harvard"DP's in Turkish" The main aim of this study is to show that DP also exists in Turkish nominal phrases. Based on the semantic and syntactic distinctions between two kinds of number expressions, namely, ones with and without classifiers, this study will try to show that there is a need for a DP projection in Turkish.
There are two ways for a noun to co-occur with a numeral in Turkish, that is, either with a bare numeral or with a numeral plus the classifier tane 'grain'. However, there are differences between the two forms. Semantically, the forms with the classifiers only refer to the quantity/number for the occurrence of the entity that the noun denotes. They do not refer to any individual in the world, hence they have no referential power. The form without the classifier, on the other hand, can refer to definite individuals in the world, hence it is referential. (1a) illustrates that the form without the classifier enables to list the names of the individuals forming the group indicated by the number, whereas (1b) shows that with the classifier such a listing is not possible as the nominal expression simply denotes the number of the people without any reference to any individuals in the world.
(1)a. Uc ogrenci keki yedi: John, Tom and Bill
Three student cake-acc ate
Three students ate the cake
b. Uc tane ogrenci keki yedi: * John, Tom and Bill
three cl. Student cake-acc ate
Demonstratives, which can only occur with definites, can also occur with the form without the classifier (2a), but not with the numeral with a classifier (2b), indicating again that the form with the classifier has no referential power.
(2) a. Bu uc ogrenci geldi b. *Bu uc tane ogrenci geldi
these three student came these three cl. student came
these three students came
The D head is considered to be the locus of reference (Longobardi 1994). Hence, this study will argue that a DP should be present in the forms without the classifier, which are referential, in order to enable the noun to have referential value. The forms with the classifier, on the other hand, are non-referential and therefore it will be claimed that DP is not projected. Hence, based on the data from numeral expressions with and without classifiers, it will be argued that all referential arguments in Turkish, namely the referential numeral expressions, bare definites, specific indefinites, proper names, pronouns and generics are introduced in the form of a DP. Non-referential numeral expressions and non-specific indefinites, on the other hand, lacking referential power also lack DP.
While this analysis will support the claim that DP is a universal category, it will still challenge the omnipresent DP account of Longobardi (1994). He claims that DP is always projected both for definite and indefinite NPs, as long as the NPs are in argument positions, even if the D head is not overtly filled. Hence he explains the asymmetries between the syntactic distribution of definites and indefinites in Romance and Germanic in terms of government relations of the D head, which he claims to be empty in indefintes, but filled in definites. However, as the data in (3) indicates there are no restrictions on the syntactic distribution of the referential and non-referential forms in Turkish: They both may occur in subject position outside the VP (3a), as well as directly adjacent to the verb (3b). This also implies that there are no empty heads which are subject to government.
(3)a. uc (tane) ogrenci kitabi hizla okudu.
three student book-acc quickly read
Three students read the book quickly
(3)b. Kitabi hizla uc (tane)ogrenci okudu
book-acc quickly three student read
three students read the book quickly
Application of Longobardi's omnipresent DP account to Turkish would require to establish complicated government relations to explain the syntactic distribution of Turkish (non-)referential argumental NPs, whereas the minimally projected DP analysis proposed in this study would account for the data in a simpler way.
MIT RLE Speech Communication Group Seminar
Chao-Yang Lee
Department of Cognitive and Linguistic Sciences, Brown University
"Lexical Tone in Spoken Word Recognition:
A View from Mandarin Chinese"
MIT - Special Phonology Circle Presentation
Tuesday, March 13, 2001, 12:30-2, Room E39-335
Hubert Truckenbrodt
MIT and Rutgers"Downstep, upstep, and register levels" The talk first treats the assignment of prosodic structure in German, and the assignment of tones relative to this prosodic structure in my Southern German material.
It then turns to the question of the relation of the prosodic structure to changes in the phonetic register in which the tones find their phonetic values. The core phenomenon of this kind is downstep, lowering the phonetic register under certain conditions. However, the properties of downstep can be studied in an interesting way in positions in which chains of downstep are interrupted and a new tonal height is assigned (reset and upstep). The Southern German material here allows for new insights into the nature of register-shifts in that it combines the presence of a partial reset in a new intonational phrase with the regular upstep of a H boundary tone, immediately preceding the partial reset. I will argue that the material supports the following conclusions:
o The crucial trigger for downstep is not so much the presence of a L tone often involved in conditioning downstep, but primarily the presence of a prominent prosodic element preceding the downstep: H tone in Yoruba (Connell and Ladd 1990, Laniran 1992), stress in English Beckman and Pierrehumbert 1986), Dutch (van den Berg, Gussenhoven, and Rietveld 1992) and Southern German, focus in Mandarin Chinese (Xu 1999).
o The model of van den Berg, Gussenhoven, and Rietveld 1992 is a good representation of our current understanding of downstep and reset. Integrating insights of Ladd (1988), it postulates register lowering on two separate levels: phrasal downstep among larger prosodic units and, embedded within that, accentual downstep. My data supports a unification of the two kinds of downstep, such that lower stress is the trigger for accentual downstep, and nuclear stress is the trigger for phrasal downstep. It will be seen that this can be captured in a model in which tones and stress are assigned to particular prosodic levels (Pierrehumbert and Beckman 1988), and in which register subordination is contingent on prosodic levels in a general way.
MIT RLE Speech Communication Group Seminar
Marc Hauser
Department of Psychology, Harvard
"What's Special about Speech? How Studies of Nonhuman Primates Inform the Debate"
Harvard University Linguistics Department Talk
Claire Bowern "Bard arr miidin biinmal burungan milon milonjun arra uliidina 'away walk to-go strongly place-allative before not one-go-pst' (A Foray into Proto Nyulnyulan Morphosyntax)"
All are welcome, and refreshments will be served. No previous knowledge of Baadi, Djawi, Dyaberdyaber, Dyugun, Nimanbur, Nyigina, Nyulnyul, or Yawuru required!
MIT Phonology talk
Thursday, March 1, 2001 at MIT, details TBAArto Anttila
Boston University and National University of SingaporeTHE BLOCKING OF PHONOLOGICAL ALTERNATIONS In this talk, I evaluate various theories of blocking in the light of Vowel Coalescence (VC) in Colloquial Helsinki Finnish, based on an electronic corpus of approximately 13,000 naturally occurring vowel sequences (Paunonen 1995). The alternation shows strange split behavior: VC is blocked in nonderived environments, but only if the structural change is phonologically marked; if the structural change is unmarked, VC applies even in nonderived environments. In addition, a parallel effect emerges in derived environments, but in a weaker form: VC is quantitatively dispreferred if the structural change is phonologically marked; if the structural change is unmarked, VC is quantitatively preferred. Finally, blocking is more prevalent in nouns than in adjectives or verbs. These facts reveal a large-scale global interaction of grammatical constraints the surface outcome of which is sometimes categorical blocking, sometimes quantitative dispreference.
I develop an optimality-theoretic analysis of these facts based on partially ordered grammars (Anttila 1997) that capture both the categorical and quantitative blocking effects. We will see that (i) nonderived environment blocking is a genuine phenomenon that requires a grammatical explanation; (ii) despite their superficial similarity, blocking of the "nonderived environment type" and blocking of the "noun type" are fundamentally different and require different analyses; (iii) nonderived environment blocking is not a property of a class of rules, hence explanations such as the Strict Cycle Condition and the Elsewhere Condition fail; (iv) blocking cannot be uniformly treated in terms of prespecification (e.g. Ringen 1975, Kiparsky 1993, Inkelas et al. 1997, Inkelas 1998), because different kinds of blocking have quite different sources.
Harvard University GSAS Workshop on Syntax
Place: 303 Boylston Hall, Harvard University (Boylston Hall is a gray building next to Widner Library in Harvard Yard.)
Dominika Baran
Department of Linguistics, Harvard"English loanwords in Polish and the question of gender assignment"
MIT RLE Speech Communication Group Seminar
David Caplan
MGH and Harvard
"Functional Neuroimaging Studies of Syntactic Processing"
MIT - Special Phonology Circle Presentation
Tuesday, February 27, 2001, 12:30-2 in Room 2-190Kie Zuraw
University of Southern California"Exceptions and Regularities in Phonology" Lexical regularities are represented as low-ranking constraints, their rankings learned through exposure to the lexicon using Boersma's Gradual Learning Algorithm. High-ranked constraints ensure the primacy of listed pronunciations; but when a speaker produces a novel word, these high-ranking constraints are irrelevant and the constraints that encode lexical regularities take over. The subterranean constraints are stochastically ranked; speakers' behavior on novel words probabilistically reflects the lexical regularities. The listener uses the same grammar to produce well-formedness judgments for novel words and to reconstruct inputs from interlocutors' outputs. The model's well-formedness judgments reproduce the experimental result that although the productivity of nasal coalescence on novel words is low, nasal-coalesced novel words are judged more acceptable than non-coalesced words in certain cases.
Bayesian reasoning by the listener favors novel nasal-coalesced words-they are disproportionately likely to become listed. A computer simulation of the speech community confirms that although nasal coalescence is the minority pronunciation for novel words, a word may eventually enter the lexicon as nasal-coalesced.
I argue that the distinction between "regulars" and "exceptions" (which has been taken as a reason evidence for leaving lexical regularities out of the grammar) reduces to a difference between listed words and synthesized words; this difference can arise through listener reasoning of the type I propose, without a prior qualitative difference.
MIT Linguistics Colloquium
Dan Jurafsky
University of Colorado at BoulderThis talks summarizes a number of results from our lab on the role of probability and statistical knowledge in human language processing in comprehension, learning, and production. These include our earlier results that in comprehension, humans compute the probability of an interpretation in order to resolve lexical, syntactic, and semantic ambiguities. I will discuss recent work that extends these results to show that aphasics are also sensitive to these same probabilities. In the area of learning, I will describe our work on the machine learning of phonology and morphology, and the way that linguistic "learning biases" can be combined with distributional and statistical inductive mechanisms. Finally, I will focus most on our recent work on lexical production, especially on our tests of the "Probabilistic Reduction Hypothesis". This hypothesis, which draws from our work and that of colleagues and predecessors from Schuchardt (1885!) and Jespersen, to Bybee and Pierrehumbert, claims that words are reduced or shortened in lexical production exactly when they are highly probable. In other words, humans compute the probability of words in language production to help determine the surface form the words should take. The implications are far-reaching and exciting. Human language processing is "probabilistic all the way down".
If time permits I hope to also spend a little time discussing the role of computation in the linguistic curriculum.
This talk describes joint work with Alan Bell, Eric Fosler-Lussier, Susanne Gahl, Daniel Gildea, Cynthia Girand, Michelle Gregory, Lise Menn, Srini Narayanan, William D. Raymond, Doug Roland, Patrick Schone, and others.
MIT - Special Phonology Circle Presentation
Thursday, February 22, 2001, 12-1:30, at MIT, Building 2-190Adam Albright
UCLA"The Identification of Bases: A Computational Approach" In this talk, I suggest that the problem of base identification is more fruitfully approached as an acquisition problem, rather than as a typological problem. In particular, I propose that base identification is part of a strategy employed by language learners to develop a system that will allow them to project inflected forms that they have not encountered before. I present a computationally-implemented model of paradigm acquisition that attempts to use a single surface form to project the rest of the paradigm, using stochastic morphological rules. It compares the predictive power of various forms within the paradigm, and selects as a base the form that allows the remaining forms to be projected as confidently and accurately as possible. As evidence for this approach, I discuss three cases in which an unexpected, marked form served as the base of an analogical change: Yiddish present tense paradigms (in which all forms were remodeled on the 1st sg), Latin noun paradigms (in which nominatives were remodeled on oblique forms), and Lakhota verbs (in which unsuffixed forms are being remodeled on suffixed forms). In each of these cases, I show how the model correctly selects the typologically marked form as the base, and also correctly predicts the direction of subsequent paradigmatic changes. Finally, I consider some of the implications of a model that is restricted to using a single surface form to project morphological paradigms.
MIT RLE Speech Communication Group Seminar
Irene Pepperberg
Media Lab, MIT
"Cognitive and Communicative Abilities of Grey Parrots"
Boston University Graduate Program in Applied Linguistics
Spring Speaker Series 2001:
Topics in Computational Linguistics
James Pustejovsky
Brandeis University"Constraining Inferences with Event-based Language Models"
MIT RLE Speech Communication Group Special Seminar
James E. Flege
Department of Rehabilitation Sciences
University of Alabama at BirminghamSOURCES OF INDIVIDUAL DIFFERENCES IN SECOND LANGUAGE SPEECH ACQUISITION
Boston University Linguistics Talk
Cecile an der Weert
University of Reading, UK"Native Elements of Discourse "
MIT LING-LUNCH
Thursday, February 15, 2001 at 12:00, MIT E39-335 (conference room)Ken Hiraiwa,
Massachusetts Institute of TechnologyVivian Lin,
Massachusetts Institute of Technology"Deriving Parametric Differences: EPP and Object Shift in the Scandinavian Languages" "A Way to Undo A-movement" Abstract: It is well-known that Scandinavian languages (Icelandic vs. the Mainland Scandinavian (MSc.); Danish, Swedish and Norwegian) show various parametric differences: Icelandic allows object shift of full DPs and weak pronominals, whereas MSc. allows only the latter; Icelandic also has Stylistic Fronting, Transitive Expletive Construction, and Quirky Subject Construction, whereas MSc. does not (but instead allows Raising-over-Experiencer Construction cotra Icelandic) (cf. Holmberg and Platzack 1995).
In this talk, I will demonstrate that all of these parametric differences in the Scandinavian languages naturally reduce to a simple single parametrization on T proposed in (1).
(i) Icelandic:
EPP on T is not contingent on Agree.MSc. :
EPP on T is contingent on AgreeIt will be shown that under (i) availability of full DP object shift in Icelandic but not in MSc. is naturally predicted by Defective Intervention Constraints of Chomsky (1999, 2000).
I will further argue that locality/minimality is a condition on each syntactic operation Agree/Move, showing that Chomsky's (1999) "phase-based" locality theory in fact makes empirically wrong predictions.
Abstract: Johnson (1996) suggests that A-movement is not subject to the CSC, based on an analysis of Gapping which involves independent A-movement of subjects out of first VP-conjuncts:
(1) [TP Desmond-i played-j [VP1 t-i t-j bass], and [VP2 Molly t-j guitar]].
This paper argues on the basis of scope interactions that, while A-movement itself is not subject to the CSC, the CSC does have an effect on the interpretation of structures like (1). Specifically, the CSC, formulated as a constraint on representations, forces independent A-movement to be undone by quantifier lowering (QL).
Sentences incorporating scopal elements show that QL is indeed required in Gapping/conjoined VP structures. For example, ' a single student' can only be interpreted as a Negative Polarity Item in (2).
(2) ? [TP A single student-i did not [[VP1 t-i ride the elevator],
or [VP2 a single professor climb the stairs]].
(2) is particularly striking because QL of the subject 'a single student' is not permitted below negation in other structural contexts. In (3), for example, a single student can only be interpreted as a specific indefinite.
(3) A single student did not ride the elevator.
Hence, QL is unexpected in conjoined structures like (1, 2). I conclude that QL must be forced in (2) by the CSC, understood as a constraint on representations.
MIT RLE Speech Communication Group Special Seminar
James E. Flege
Department of Rehabilitation Sciences
University of Alabama at BirminghamTHE PERCEPTION OF SECOND-LANGUAGE VOWELS We evaluated these general approaches in a series of categorial discrimination tests. The stimuli were multiple natural tokens of English vowels presented in an oddity format. To estimate sensitivity to between-vowel contrasts, "catch trials" (three physically different instances of one vowel category) were included to generate a false alarm rate. Separate studies were carried out to examine the discrimination of English vowels by native speakers of Italian, Korean, and Japanese, all of whom were learning English naturalistically in North America. Three general conclusions can be drawn. Learning to discriminate L2 vowels takes place slowly, over the course of years. Individuals who begin began learning an L2 as children ("early" bilinguals") often discriminate L2 vowels more accurately than individuals who began learning their L2 as adults ("late" bilinguals). However, some late bilinguals eventually learn to discriminate L2 vowels in a native-like manner. Taken together, the results are more consistent with the view that L2 vowel perception learning is limited by the quantity and quality of L2 phonetic input that has been received, and that the primary "constraint" on learning is the effect of previous learning, not maturation.
MIT RLE Speech Communication Group Special Seminar
Albert Costa
Department of Psychology
Harvard UniversityFUNCTIONAL ARCHITECTURE AND PROCESSING DYNAMICS IN SPEECH PRODUCTION: THE CASE OF COGNATES AND HOMOPHONES
MIT Phonology Circle
Marie-Helene Cote
University of Wisconsin, Madison"A perceptual approach to the resolution of consonant clusters: the French schwa revisited" In this talk I present and exemplify the perceptual approach to deletion and epenthesis using a single set of data: the distribution between schwa and zero in French. The behavior of the French schwa is notoriously complex and provides a good illustration of various aspects of the proposal. No available analysis entirely captures the intricacies of the data, but I show that an approach based on perception achieves substantial progress, from a descriptive, explanatory and conceptual point of view, over syllabic approaches. Empirically, the perceptual approach accounts for a number of generalizations that remain mysterious under a syllabic account. Conceptually, it integrates under the more general notion of perceptual salience principles that were thought to be independent: on the one hand, principles of syllable well-formedness, on the other hand, principles that operate on sequences of segments, in particular the Obligatory Contour Principle. It also eliminates the need for exceptional (and undesirable) mechanisms such as extrasyllabicity in the segmental domain.
The talk is structured as follows. I first provide some basic facts about the French schwa, and then show how syllabic analyses fail to account for its distribution. The major part of the talk is devoted to the presentation of the alternative approach. We will see how this proposal naturally explains the major segmental generalizations observed in the alternation between schwa and zero. Then I outline a formal implementation of this proposal in the framework of Optimality Theory. I conclude with a summary and some open issues, including the status of perceptual (and other functionally-based) constraints in the grammar.
MIT LING-LUNCH
Thursday, February 8, 2001 at 12:00, MIT E39-335 (conference room)Paul Elbourne
Massachusetts Institute of TechnologyNorvin Richards
Massachusetts Institute of Technology"On the Semantics of Pronouns and Definite Article" "A Distinctness Condition on Linearization" Abstract: Lots of semanticists would have you believe that pronouns can be interpreted in two ways, as variables and as E-type pronouns. The trouble is that no language that we know of makes a lexical or morphological distinction between these two meanings, so we really seem to be dealing with one thing.
This paper attempts to show how a unified semantics can be given for the two uses by using Postal's idea that pronouns are definite articles. As I argued in some previous work, donkey pronouns are definite articles followed by an NP that has undergone NP-deletion: 'Every man who owns a donkey beats it' is really 'Every man who owns a donkey beats it donkey', where 'it' means more or less the same as 'the'. Extending this line of thought, I argue in this paper that indices are phonologically null NPs. This allows us to deal with binding and reference in the same terms that we used for donkey pronouns: pronouns are definite articles taking phonologically null NPs.
Finally I argue that 'it' means what Fregeans thought 'the' meant, and that 'the' means something a bit different.
Abstract: The syntax literature contains a number of bans on "overcrowding" of various kinds: cases in which things "of the same type" cannot be "too close together" (where all of these terms will have to be defined). For example, quotative inversion in English is subject to a restriction that could be stated (as it is by Alexiadou and Anagnostopoulou, to appear) as a requirement that the vP not contain two DPs by the end of the derivation:
(1) a. "It's raining," said the weatherman
b. "It's raining," said the weatherman to the anchorwoman
c. *"It's raining," told the weatherman the anchorwoman
In this talk I'll try to unify the various conditions of this type that we seem to find. Adopting Kayne's (1994) LCA, I'll argue that the ordered pairs used for linearization can make reference only to node labels, and that this restriction sometimes leads to such ordered pairs being insufficiently informative for linearization to succeed.
Human Development Program Colloquium
Helen Tager-Flusberg, Ph.D.
University of Massachusetts
Eunice Kennedy Shriver Center"The role of language in theory of mind: Evidence from preschooler, autism, and Williams Syndrome."
All are welcome.Graduate Program in Human Development
Department of Psychology
Boston University
64 Cummington Street
Boston 02215For more information, see http://www.bu.edu/psych/graduate/hd/humtalks.html or contact gagnej@bu.edu
MIT LING-LUNCH
Thursday, February 1, 2001 at 12:00, MIT E39-335 (conference room)Ben Bruening
MITThe Hidden Syntax of Discontinuous NPs in Passamaquoddy (1) peskuw-ok n-nomiya-k kiwhosuw-ok wolaku
some.of-3P 1-see-3P muskrat-3P yesterday
`I saw some of the muskrats yesterday'
This construction raises many questions for syntactic analysis, chief among them how the two parts are related and interpreted. In exploring the syntax of discontinuous NPs, I will argue for a simple analysis involving movement of the determiner away from the NP, and show that more complicated facts follow, including intervention effects by quantificational elements and binding facts.
Boston University
lecture on language acquisition
Holly Storkel
Indiana University Speech Research Laboratory
Learnability Project
"The Emerging Lexicon:
Phonological Influences on Learning and Structure"Abstract: Children acquiring a language are faced with at least two lexical tasks: learning new words and organizing these in memory to uniquely distinguish each word from every other. During this process, children are also learning the sound or phonological system, and it is likely that phonology may influence lexical development. These issues are addressed in three studies. Studies 1 and 2 examine how the likelihood of sound occurrence, termed phonotactic probability, influences word learning in school-age and pre-school children. Study 3 investigates the structure of words in the lexicon of pre-school children and compares these findings to previous studies of infants and adults. The results indicate that phonology does influence word learning and that lexical structure changes across the lifespan. Future theoretical and clinical directions will be discussed.
MIT Appl(icative) Fest 2001
Independent Activities Period Workshop on Applicative Constructions and Related Issues
For further information, please cotnact marantz@mit.edu
Boston University Graduate Program in Applied Linguistics
Spring Speaker Series 2001:
Topics in Computational Linguistics201222, College of Arts & Sciences, 725 Commonwealth Ave, Boston.
Note the room changeJeff Adams
Senior Speech Scientist at Lernout and Hauspie"Language Modeling for Speech Recognition"
"On Neo-Darwinian Linguistics:
The Creolist as Myth-Maker"Prof. Michel De Graff of MIT Refreshments and ASL interpreting will be provided.