Harvard University
Friday, December 17 at 4:00 p.m. in the Department of Linguistics Lounge, 3rd floor Boylston Hall (Harvard Yard).Gulsat Tosun
Harvard UniversityThe Structure of embedded clauses in Turkish:
Extraction possibilities and T-to-C Abstract: The possible complement clauses in Turkish consist of non-finite and restricted instances of finite (preverbal) clauses. With respect to their properties in allowing extraction of phrases to a higher clause, non-finite clauses pattern with ECM constructions; finite-clauses pattern with subjunctive infinitivals; and infinitival clauses with wh-in-situ I will present data illustrating these contrasts and discuss the contrasts in terms of T-to-C movement . I will argue that non-finite clauses in Turkish are indeed CPs but with null Tense contrary to Kural (1993),who assumes them to be[+ tense] and claims that the C filled by ^Vk. T-to-C allows both the subject and the object to be equidistant with respect to satisfying EPP (type feature) on C (in the sense of Miyagawa 1999) and extends the domain of extraction (Nichols 1999). This feature on C requires overt movement of an argument to its Spec position, which in turn allows this clause edge position to be an intermediary landing site. I will address the question ^S Why T-to-C cannot provide an intermediary landing site in finite embedded clauses?^T and try to account for it I will further discuss my observations on Switch-Reference Obviation Instances and availability of P(air) L(ist) readings in multiple wh-constructions.
MIT LING-LUNCH
Thursday, December 9, 1999 at 12:10, MIT E39-335 (conference room)Peter Svenonius Negative Movement in Icelandic and the Expression of Negation Abstract:In Icelandic, objects ordinarily follow verbs, including participial and infinitival complements to auxiliaries, but a negative object precedes such unmoved verb forms.
1. Eg hef lesidh bokina.
I have read the.book
2. Eg hef ekkert lesidh.
I have nothing read
"I have read nothing"
The leftward movement of a negative object (Negative Movement, or NM) is obligatory.
3. * Eg hef lesidh ekkert
I have read nothing
When there is no auxiliary, NM is string-vacuous, as the verb moves to the second position.
4. Eg las ekkert.
I read nothing
I analyze NM as forced by a licensing requirement on negation (cf. previous work by Christensen, Jonsson, Haegeman). Questions arise as to how languages like English permit sentences like (3).
More central to this talk, however, will be certain striking constraints on Icelandic NM. For example, although NM across a preposition is generally allowed, as in (5), it generally also requires that a verbal element be crossed, as indicated in (6).
5. Eg hef engum hlegidh adh
I have nobody laughed at
"I have laughed at nobody"
6. * Eg hlo engum adh
I laughed nobody at
I analyze this by taking Case to be checked at the level of the (lowest) vP, while NM is to the highest verbal projection. This means that in (6), the Case of the negative prepositional complement is inappropriately checked against the verb, since the NM position and the Case-checking position are the same.
For more information, contact karlos@MIT.EDU.
MIT RLE Speech Communication Group Seminar
Wednesday, December 8, 1999 at 3:00 PM at MIT 34-401BCorine Bickley
Speech Communication Group, MITTHE ENABL PROJECT -- A SPEECH USER INTERFACE FOR ENGINEERING DESIGN SOFTWARE Abstract: In the ENABL project, a Speech User Interface was developed for engineering design software. Implementation of the Speech UI included design of the grammar and vocabulary, incorporation of the speech recognizer into the interface architecture, and construction of a compiler to interpret the speech commands. Because no database existed, to our knowledge, of spoken phrases for the task of engineering design, we used existing drawings and engineering manuals for compilation of lexica. Based on feedback from users and trainers of an existing program for rule-based engineering design, a language for the ENABL Speech User Interface was devised. In parallel with the development of the language, construction of the system architecture and design of the interface were undertaken. Screen layouts and menus of information were designed for two tasks. One interesting feature of the Speech UI is its use of context. The Speech User Interface is controlled by a series of phrases. The context established by one phrase is used to interpret phrases that are spoken later during the session; context information is stored in a "dialog history". The core recognizer is a continuous-speech HMM-type recognizer for Swedish (Strom, 1997); the recognizer had been initially trained on the speech of approximately 50 speakers in the Stockholm area. The users in the ENABL project speak dialects that differ from Stockholm Swedish. Thus, a major effort in the project was focused on improving the performance of the recognizer for these users. Results of recognizer performance before and after adaptation will be presented. Another aspect of the project involved investigating the acoustic characteristics that would indicate good or difficult use of the speech recognizer by persons with dysarthric speech. We found that adaptation using dysarthric speech provided encouraging results for use of the speech recognizer by individuals with certain characteristics of dysarthric speech. The complete system is being evaluated in terms of performance of two engineering tasks. Preliminary evaluation results indicate that a usable "correct task" performance can be obtained for constrained tasks despite "correct word" recognition rates that are low for some words. Acceptable performance of the system for the user's vocational needs was accomplished, in part, because of limited task coverage, a highly-constrained grammar, and small lexica.This work was carried out in collaboration with Christel de Bruijn, Alice Carlberger, Rolf Carlson, Peter Cudd, Roger Gustavsson, Sheri Hunnicutt, Erland Lewin, Tina Magnusson, Lennart Nord, Frido Ordemann, Bengt Reimers, Mats Reimers, Kristin Rosen, Elisabet Rosengren, Nikko Strom, Nathalie Talbot, Sandra Whiteside, and Sasha Yampolsky
For more information, contact czoll@MIT.EDU
Harvard University
Friday, December 3 at 4:00 p.m. in the Department of Linguistics Linguistics Department Seminar Room, Boylston Hall 303 (Harvard Yard).Claire Bowern
Harvard UniversityAnd the Paradigm was made Karnic Abstract: Ms. Bowern will discuss the development of case morphology in Proto-Pama-Nyungan and the reflexes of some of these cases in the Karnic languages of the Lake Eyre basin of central Australia. No previous knowledge of Arabana-Wangkangurru required.
MIT Linguistics Colloquium
Friday, Devember 3, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Juliette Blevins
University of CambridgeUnderstanding Geminates:
The Evolution of Consonantal Length Contrasts Abstract: In Evolutionary Phonology historical explanation plays a major role in determining phonological typology. In this talk I present evidence that differences in the distribution, inventory, and phonological behaviour of geminate and non-geminate consonants originate in perceptual, acoustic, and articulatory properties of speech. I begin by exemplifying the many ways in which consonantal length contrasts can arise through regular sound change. I then demonstrate how these distinct origins provide the basis of understanding the language-specific nature of geminate weight, geminate integrity and antigemination.
For more information, contact asrackow@mit.edu.
MIT LING-LUNCH
Thursday, December 2, 1999 at 12:10, MIT E39-335 (conference room)Judy Baek and Ken Wexler The Syntax of Negation, Double-Checking and Development of Object Positions in Korean Abstract:The purpose of this talk is to understand a particular pattern of word order errors in child Korean and to use an analysis of these errors to help to solve a difficult and unresolved problem in the syntax of adult Korean, namely the structure of short-form negation.
There is a very well-known class of word order error in young (around 2 years old) child Korean. Namely the order of the object and the negation marker *an* (not) is reversed. In adult Korean the object always precedes *an* whereas children often produce *an* before the object. (1a) is the correct adult form and (1b) is a typical child error.
(1) a. na pap an mek-e
I rice Neg eat-Decl
b. na an pap mek-e
I Neg rice eat-Decl
"I don't eat rice."
(I). First we will present a detailed quantitative analysis of a young (2;0 - 2;8) Korean child's productions. We show two facts in detail:
(2) The child often (but not always) makes the word order error, putting the object after negation.
(3) The child often produces the Accusative Case marker on the object when the object precedes *an*, but never produces this accusative marker when the object (incorrectly) follows *an*. The accusative case marker is more or less optional in (adult) Korean, so the optionality of the marker in correct sentences is expected.
A general property of constructions in this period of child language is given in (4):
(4) A child often fails to raise a constituent that should be raised, but rarely raises a constituent that should NOT be raised.
Thus (4) suggests that the Korean child's error is to fail to raise the object to the left of *an*. Furthermore, the fact that *an* precedes the object in the child errors suggests that *an* starts out to the *left* of the VP, thus must be in a SPEC position (Spec,Neg) since Korean is strongly right-headed. In Part III we will give a detailed syntactic analysis of negation with these properties.
II. Assuming that *an* is in Spec,Neg and that the child often fails to raise the object, we want to ask *why* the child fails to raise the object. We assume that the child knows the complete syntax of negatioon in Korean. After all, the child often raises the object to the correct position and only assigns accusative case in this correct raised position. Thus we have to account for why the child (optionally) doesn't raise the object (and thus doesn't assign ACC case).
A natural place to look for an explanation is in the class of well-known child errors around this age. The Optional Infinitive stage characterizes children in this age range. Wexler (1998) has proposed the Unique Checking Constraint (UCC) as an explanation for the OI stage. Namely, the UCC says that a D-feature on a DP can only check *one* D-feature on a functional category. Therefore, the subject DP can only check the D-feature on AGR-S *or* on TNS, but not on both, and the child's grammar will therefore eliminate either AGR-S or TNS. The AGR/TNS Deletion Model, which Schutze and Wexler have shown to be empirically adequate for the OI stage, thus results.
Furthermore, Wexler (mss.) has shown that the UCC can also explain why Romance object clitics are omitted in this same age-range. Namely, since a clitic (or possibly pro, depending on which theory of clitics is assumed) must check *two* D-features (AGR-O, where it checks case, and INFL, where it winds up), the derivation with a clitic doesn't converge (given UCC), and the child will have to omit the clitic.
In other words, UCC applies to object positions as well as to subject positions. Given that the age range of the child error is the same as in the OI or clitic-omission stage, it is natural to look for an explanation for the Korean word order error in the UCC. In other words we should look to find an analysis of Korean short form negation in which
(5) a. *an* is in Spec,NEG
b. the object must move to a higher functional projection, checking
*two* D-features, i.e. two functional categories
Under (5) we would have an exact prediction of the child patterns, on the assumption that the child knew UG, knew the correct parameter values for Korean and was subject to the UCC. This would be an ideal solution.
III. The Syntax of short-form negation in Korean
We assume that Spec, AgrOP is the position to which objects move in Korean, a position where accusative case is checked.Following standard assumptions, AGROP dominates NEGP, and this will explain why the object preces *an*, which is in Spec,Neg. However, the child often leaves the object to the right of *an*, and we want to make this result from the child's doing a *single* movement, i.e. checking a single D-feature. Thus we suggest that there is an AspectPhrase, AspP, which NegP dominates. The incorrect productions which leave the object to the right of *an* have the object in Spec,Asp, i.e. they've raised once. We'll suggest some independent evidence in (adult) Korean for the existence of AspP.
In these terms (somewhat more complex than we present in this abstract), we work out the detailed syntax of short-form negation in Korean, predicting a number of phenomena, including position of adverbs. The analysis seems to be relatively successful at capturing the properties of negation in Korean.
Furthermore, the analysis predicts that subjects of unaccusative verbs will (optionally) not raise to the left of negation in children subject to the UCC, a fact which turns out to be true. Furthermore, as predicted, when the subject of the unaccusative fails to raise to the left of negation (in children), no (NOM) case-marking is placed on the subject.
For more information, contact karlos@MIT.EDU.
MIT RLE Speech Communication Group Seminar
Wednesday, December 1, 1999 at 3:00 PM at MIT 34-401BHarlan Lane, Melanie Matthies, Joseph Perkell, Jennell Vick, Majid Zandipour THE EFFECTS OF CHANGES IN HEARING STATUS ON THE ACOUSTIC VOWEL SPACE AND COARTICULATION Abstract: In order to examine the role of hearing status in controlling coarticulation, eight English vowels in /bVt/ and /dVt/ syllables, embedded in a carrier phrase, were elicited from seven postlingually deafened adults in repeated recording sessions both before and up to a year after they received cochlear implants and their processors were turned on. Measures were made of second formant frequency at obstruent release and at 25 ms intervals until the final obstruent. With the change in hearing status, four of the speakers significantly reduced the size of their vowel spaces in the formant plane, while the others increased them. All speakers but one also reduced vowel duration significantly. Four of the speakers had lower dispersion of vowel formant values around vowel midpoint means, but the other three did not.An index of coarticulation based on the ratio of F2 at vowel onset to F2 at midvowel target was computed. Changes in the amount of coarticulation after the change in hearing status were small and nonsystematic for the /bVt/ syllables, while those for the /dVt/ syllables averaged a three percent increase - within the range of reliability measures for two hearing control speakers. The slopes of locus equations and ratios of F2 onset in point vowels tend to confirm the conclusion that hearing does not play a direct role in regulating anticipatory coarticulation in adulthood.
For more information, contact czoll@MIT.EDU
MIT Linguistics Colloquium
Friday, November 19, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Maribel Romero
University of PennsylvaniaON SOME SIMILARITIES BETWEEN REDUCED CONDITIONALS AND SLUICING Abstract: This paper is concerned with two ellipsis constructions: Reduced Conditionals (in German) and Sluiced Interrogatives Clauses (e.g. in English), illustrated in (1) and (2) respectively.(1) Wenn ich wen besuche, dann (immer) den Peter.
If I somebody visit then always the Peter
"If / whenever I visit somebody, it's Peter."
(2) Something sneaked into the garbage can, but I don't know what.
In particular, this paper examines two peculiar characteristics that the two constructions share and that make them look rather different from other types of ellipsis, namely: (i) the restriction on possible antecedent phrases for the remnant NP / WhP, and (ii) the correlation between Quantificational NPs (QuNPs) in the antecedent clause and E-type expressions in the ellipsis clause. I will argue against current approaches in the literature that posit an idiosyncratic account for each of these types of ellipsis. I will show, instead, that these facts follow from the interaction of fairly standard assumptions about ellipsis-- presence of Focus in the remnant material-- and the semantics of Conditionals and questions.
Let us first examine the restriction on possible antecedents. Chung-Ladusaw-McCloskey (1995) (CLM, henceforth) claim that the correlate of a sluiced wh-phrase in the antecedent clause can be an indefinite (something in (2)), but not a name nor a QuNP, as the ungrammaticality of (3) suggests. A similar observation is made by Schwarz (1999) for Reduced Conditionals: the indefinite wen in (1) is a good correlate for the remnant den Peter, whereas the name den Karl in (4) is not. (Note that the full-fledged versions of (3) and (4) are perfectly grammatical.)
(3) a. *? I know that Meg's attracted to Harry, but they don't know (to) who(m).
b. * Each of the performers came in. We were sitting so far back that we couldn't see who. (CLM)
(4) * Wenn ich den Karl besuche, dann immer den Peter. (Schwarz)
If I the Karl visit then always the Peter
After providing new data that challenge CLM's approach, I will argue that, in both constructions, the restriction follows from the presence of Focus stress on the remnant WhP / NP. In the case of Sluicing, Focus requires the denotation of the sluiced interrogative to contrast with an alternative question in the discourse, from which the restriction of possible antecedents follows. In the case of Reduced Conditionals, the Focus on the remnant NP produces an exhaustivity implicature that, together with the semantics of Conditionals, derives the desired generalization.
The second observation is that QuNPs and indefinites in the antecedent clause correspond to E-type-like expressions in the ellipsis clause. It has been observed that, in Reduced Conditionals, indefinites in the antecedent clause correspond -at least semantically-- to definites in the elided clause: whereas the full-fledged conditional (5b) may talk about two different students, the reduced (5a) can only talk about one student. In Sluicing, Merchant (1998) and Romero (1998) argue that QuNPs in the antecedent clause often must correspond to E-type pronouns in the Sluice: e.g., (6a) has a pair-list-like reading, which cannot arise from the interrogative (6b) with a QuNP, but from the interrogative (6c) with plural definite.
(5) a. Wenn ein Student was zum Trinken bestellt, dann Bier.
If a student something to drink orders, then beer
b. Wenn ein Student was zum Trinken bestellt, dann bestellt ein Student Bier.
(6) a. At least seven boys from your class danced the first waltz of the night, but I don't know who with. (Romero 1998)
Pair-list-like reading: ", but I don't know who each of them danced with."
b. Who did at least seven boys from your class dance the first waltz with?
* Pair-list-like reading.
c. Who did they / the boys dance the first waltz with?
Pair-list-like reading, derived from cumulative reading.
This definiteness effect has not been explained for Sluicing, and, as I will show, is not entirely captured by Schwarz' analysis for Reduced Conditionals without further modifications. It will be shown that the new proposal, instead, correctly derives all these data.
For more information, contact asrackow@mit.edu.
UMass, Amherst: Linguistics Colloquium
Friday, November 19, 1999, at 3:30pm in Machmer W-24, followed by drinks and munchies in the department lounge and dinner afterwards.Henry Davis
University of British ColumbiaPronominal Determiners Abstract not available. For more information, contact the Department of Linguistics at UMass.
MIT LING-LUNCH
Thursday, November 18, 1999 at 12:10, MIT E39-335 (conference room)Uli Sauerland
Tuebingen UniversityDonkey Anaphora: A new E-type analysis Abstract: Donkey anaphora involve a relationship to an antecedent that doesn't c-command them and to a quantifier mediating the dependency that c-commands both the donkey anaphor and its antecedent. I explore a new version of an E-type analysis; one that understands the relationship between donkey anaphor and antecedent as that of elided phrase and antecedent. A sketch of the result is (1), where both antecedent and donkey anaphor are interpreted as a Skolemized choice functions (Chierchia 1999) with the variable x bound by the mediating quantifier.(1) Every man who had a dime in his pocket put it in the meter.
every x f(dime,x) f(dime,x)
Evidence for (1) comes from weak crossover effects and a scopal behavior that parallels other choice function indefinites.
For more information, contact karlos@MIT.EDU.
MIT RLE Speech Communication Group Seminar
Wednesday, November 17, 1999 at 3:00 PM at MIT 34-401BEhan Goldstein and James Heaton AN EMG-EL INTERFACE SYSTEM PROVIDES HANDS-FREE CONTROL OF PROSTHETIC VOICE PITCH, ONSET AND OFFSET Abstract: Each year, thousands of people lose their voices as a result of laryngectomy to treat laryngeal cancer. For many such individuals, the only viable option for speech is to use a hand-held electrolarynx (EL). Unfortunately, currently available EL devices produce speech that is reduced in intelligibility and non-human sounding. In particular, EL users complain that their speech is monotonous and that using an EL sacrifices the function of one hand. In addition, without fundamental frequency (Fo) modulation, EL users have difficulty generating intonational and contrastive stress which markedly limits the informational content of their speech (Gandour & Weinberg, 1983).Improved control over EL devices might be possible using electromyographic (EMG) signals from skeletal muscles. I will demonstrate how a prototype EMG-EL interface can be used to trigger the onset/offset and control the Fo of an EL device through the conscious contraction of recorded muscles. Optimal EMG signals for EL control may be obtained from residual laryngeal motor nerve signals after laryngectomy. We have sought to acquire such signals by transposing the laryngeal nerves to alternate muscle targets at the time of laryngectomy (Heaton et al., 1999, in press). To date we have performed nerve transposition in 9 human volunteers at the time of total or partial laryngectomy. We have shown that transposed laryngeal signals can potentially enable more natural hands-free EL speech production than current commercial devices provide.
References:
Gandour, J., & Weinberg, B. (1983). Perception of intonational contrasts in alaryngeal speech. J Speech Hear Res, 26, 142-8.
Heaton, J.T., Kobler, J.B., Goldstein, E.A., McMahon, T.A., Barry, D.T., & Hillman, R.E. (in press). Recurrent laryngeal nerve transposition in the guinea pig. Ann Otol Rhinol Laryngol.
Heaton, J. T., Kobler, J. B., Goldstein, E. A., Zeitels, S. M., Randolph, G. W., Walsh, M. J. & Hillman, R. E. (1999) Laryngeal nerve transposition in humans. Soc Neurosci Abst, 25:1514.
For more information, contact czoll@MIT.EDU
Boston UniversityDeaf Organization
Tuesday, November 16, 1999 at 7 PM in the School of Theology, room 113.
CODA (Children of Deaf Adults)
panel discussion
Harvard University
Tuesday, November 15 at 5:15 p.m. in the Department of Linguistics Lounge, 3rd floor Boylston Hall (Harvard Yard).Balkiz Ozturk
Harvard UniversityTURKISH AS A NON-PRO-DROP LANGUAGE Abstract: The main aim of this study is to argue against the claim that Turkish is a pro-drop language. By examining the discourse dependent nature of overt personal pronouns in Turkish it will be argued that overt pronouns are in fact topic pronouns whose presence or absence within a particular structure is solely determined by discourse and that they are not generated in [Spec, VP], but in a topic position higher in the structure, namely [Spec, TopP]. It will also be claimed that the agreement morphology in Turkish is a pronominal category base-generated in [Spec, VP] as the VP-internal subject.
MIT RLE Speech Communication Group Seminar
Wednesday, November 10, 1999 at 3:00 PM at MIT Building E25, Rm 401Peter Jusczyk
Psychology and Cognitive Science, Johns Hopkins UniversityHOW INFANTS USE INFORMATION IN THE SPEECH SIGNAL TO SEGMENT WORDS Abstract: Several years ago, Jusczyk and Aslin (1995) reported that infants first display some abilities to segment words from fluent speech at around 7 months of age. Many subsequent studies have focused on the nature of the information that infants rely on to find words in fluent speech. The kinds of cues investigated include prosodic, phonotactic, allophonic, and statistical cues to word boundaries. English-learners appear to develop sensitivity to some of these types of cues earlier than they do for others. I will review some of these findings, along with some more recent attempts to investigate the relative weighting that infants give to the different types of cues. Although much remains to be learned about how infants come to integrate these different types of cues, it is clear that word segmentation abilities evolve considerably in the second half of the first year.Suggested Readings:
Jusczyk, P. W. & Aslin, R. N. (1995). Infants' detection of the sound patterns of words in fluent speech. Cognitive Psychology, 29, 1-23.
Brent, M. R. & Cartwright, T. A. (1996). Distributional regularity and phonotactic constraints are useful for segmentation. Cognition, 61, 93-125.
Saffran, J. R., Aslin, R. N, & Newport, E. L. (1996). Statistical learning in 8-month-olds. Science, 274, 1926-1928.
Jusczyk, P. W. & Hohne, E. A. (1997).Infants' memory for spoken words. Science, 277,1984-1986.
For more information, contact czoll@MIT.EDU
Harvard University
Monday, November 8, 1999, at 4:30 PM - in Boylston 105.Peter Kosta
Potsdam UniversityMinimalism and Free Constituent Order in Russian as Compared to German Abstract: The present paper deals with an analysis of so-called free word order languages. I shall discuss two alternative approaches: one in which both phenomena are treated as long vs. midrange scrambling (A-bar vs. A-movement), and one in which free word order phenomena one finds in Russian and German are base generated. If free constituent order is a base generated phenomenon, as I try to argue for, its properties should follow from independent motivated fundamentals of the Minimalist program.In the first part of my paper I shall concentrate on long scrambling constructions as they occur in Russian, but the paper also includes some remarks on the so-called "mid range scrambling phenomena" of the type one finds in German (cf. Mueller 1995, Fanselow 1996). Section two shows why unbounded movement such as scrambling cannot be included into the minimalist style syntax. Section three shows that the derivation of free constituent order in Slavic and German by a movement process is neither necessary nor possible. Instead, I try to advocate a position in which the basic facts of free word order are resolved in favor of Minimalist Theory.
Boston UniversityConference on Language Development
Friday, November 5 through Sunday, November 7, 1999For more information, go to http://www.bu.edu/linguistics/APPLIED/conference.html or send e-mail to langconf@louis-xiv.bu.edu.
MIT Linguistics Colloquium
Friday, November 5, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Lisa Green
University of Texas, AustinPREDICATE COERCION and BARE PLURAL SUBJECTS in 'BE'-TYPE CONSTRUCTIONS Abstract: This paper examines 'be'-type constructions in African American English (AAE). 'Be'-type constructions are distinguished from simple tense generics that are ambiguous between habitual/generic and capacity readings:1)a. This printer print fifty pages a minute.
'This printer prints fifty pages a minute'
b. This printer be printing fifty pages a minute.
'This printer usually prints fifty pages a minute'
The sentence in (1a) is ambiguous between the type of generic/habitual andcapacity readings that are discussed in Schubert and Pelletier (1989). On the generic/habitual reading, the interpretation is that the printer often prints fifty pages per minute, and on the capacity reading, the interpretation is that the printer has the capacity to print fifty pages per minute. The sentence in (1b) is unambiguous; it only has the habitual interpretation. It is a statement about the actual printing activity, and it indicates that the printer actually does what it has the capacity of doing.
The analysis proposed to account for 'be'-type constructions is one in which aspectual 'be' neutralizes the distinction between stage- and individual-level predicates. Following Kratzer (1995), I assume that stage-level predicates have a separate event argument associated with them, but individual-level predicates do not. Aspectual 'be' forces individual-level predicates to take an eventuality argument, which coerces them into stage-level predicates. As a result, individual-level predicates that occur in construction with aspectual 'be' have a stage-level reading:
2) Sue be knowing those songs.
(Lit: On different occasions, Sue shows that she knows those songs.)
The predicate 'know' indicates an individual-level property, yet it can occur in 'be'-type constructions to express an eventuality that is understood as holding on different occasions. The logical representations of these constructions are given a tripartite structure in which a habitual operator (HAB) binds variables ranging over eventualities.
This paper also considers the interpretation of bare plural subjects in stage- and individual-level predicate constructions. Diesing (1992) argues that there is an asymmetry between stage- and individual-level predicates with respect to subject interpretation such that subjects of stage-level predicates have generic and existential interpretations, and bare plural subjects of individual-level predicates have generic interpretation. Consider the sentences in (3) and (4):
3) Emergency room nurses be on call.
a. 'Emergency room nurses are usually on call'
b. 'There are usually emergency room nurses on call'
4) Emergency room nurses be patient.
a. 'Emergency room nurses are usually patient'
b. *'There are usually patient emergency room nurses'
In the sentence in (3), the predicate is stage-level, and the bare plural subject receives generic (3a) and existential (3b) interpretations. In the sentence in (4), the individual-level predicate occurs with aspectual 'be', so it is coerced into a stage-level predicate; however, the bare plural subject only receives a generic (4a) reading. It does not get the type of existential interpretation that occurs when bare plural subjects are in construction with predicates that are inherently stage-level, as in (3b). This paper considers the interaction between HAB and the eventuality argument in explaining why the reading in (4b) is ruled out.
The solution proposed here accounts for some well-known properties of aspectual 'be' constructions that have not been discussed in the literature.
For more information, contact asrackow@mit.edu.
MIT LING-LUNCH
Thursday, November 4, 1999 at 12:10, MIT E39-335 (conference room)Bruce Tesar
Rutgers UniversityOvercoming Ambiguity in Language Learning Abstract: One of the central problems in language learning is ambiguity in the overt data available to the learner. An overt form (the audible portion of an utterance) may be ambiguous between more than one full linguistic analysis, and different languages may choose different analyses. In metrical stress, a three syllable word with main stress on the middle syllable is ambiguous between an iambic analysis (in which the first two syllables are grouped together into a foot) and a trochaic analysis (in which the final two syllables are grouped together into a foot). Because linguistic principles evaluate entire linguistic analyses (and not just overt forms), the learner must overcome this ambiguity in order to use the overt forms to learn the grammar of their native language.This talk will present results from a learning algorithm capable of learning from ambiguous data. This research uses the linguistic framework of Optimality Theory, so that the primary goal of the learner is to learn the language-specific ranking of the universal constraints. The learner overcomes ambiguity by temporarily considering more than one analysis of the ambiguous form, and then eliminating those analyses which are inconsistent with other data from the language. The simulation results demonstrate that this strategy can succeed without explosive growth in the number of grammatical hypotheses the learner has to maintain during the course of learning.
The key to the success of this algorithm is the interrestrictiveness of the forms. Even though the stress patterns of two words each admit multiple analyses, the learner will commonly be forced to a particular analysis of the first word in order to have a grammar capable of producing a the correct stress pattern for the second word. In isolation, each word is ambiguous, but collectively, the words of the language force the learner to a single grammar (and thus a single analysis of each word). When working within Optimality Theory, the interests of linguistic theory and language learning align: constraint interactions which are simultaneously implicated in the analysis of many words makes for better linguistic theory, and also makes for more efficient learning.
For more information, contact karlos@MIT.EDU.
MIT RLE Speech Communication Group Seminar
Wednesday, November 3, 1999 at 3:00 PM at MIT 34-401BJanet Cahn
Dept. of Linguistics, Edinburgh UniversityADDING SPEAKER CHARACTERISTICS TO SYNTHESIZED SPEECH Abstract: Most uses of synthesized speech emphasize the structure and semantics of the text being synthesized. The result is speech whose prosody is predictable and uninteresting at best; more often, it is unnatural and recognizably artificial. The problem is that the speaker is missing from the speech. This speaker has a body, a brain, an emotion, an attitude, a memory. All affect the words that are uttered and their acoustical properties. In my talk, I will describe two projects that address the influence of the speaker on speech acoustics: a description model that generates recognizable affect and emotion in synthesized speech, and a simulation system that produces three styles of prosody - child-like, adult expressive and expert - by tuning the the recall parameter of a model of attention and working memory.References (available at http://www.media.mit.edu/~cahn/):
The Generation of Affect in Synthesized Speech. Janet E. Cahn. Journal of the American Voice I/O Society. July, 1990. Volume 8, pp. 1-19.
A Computational Memory and Processing Model for Prosody. Janet E. Cahn. Proceedings of the Fifth International Conference on Speech and Language Processing, Sydney, Australia. December, 1998. Volume 3, pp. 575-579.
For more information, contact czoll@MIT.EDU
Harvard University - mini conference
co-organised by the Dept. of Linguistics at Harvard and Eliot House.
Friday, October 29, 1999 in the Thompson Room (Barker Centre)30-minute talks (20 mins. + 10 mins. for questions) 2:25Opening
2:30-3:00Ute Bohnacker, "The acquisition of syntax in Icelandic-English child bilingualism"
3:00-3:30Jorun Hetland, "On the interpretation of the fall-rise accent"
Abstract: I will compare the fall-rise pitch accents used in English and German declarative sentences. Although there are phonetic differences between the two, and although the German fall-rise seems to be restricted to sentences with a two-peaked accent pattern, it will beshown that the fall-rise in English and the fall-rise in German have crucial properties in common.
It turns out that the fall-rise does not directly influence the information structure of the sentence in which it occurs (this goes for both languages). The accent functions as a marker of referential or cognitive status: It marks a constituent as belonging to a set in the addressee's knowledge-store.
This property of the fall-rise (marking constituents as hearer-old in the sense of Prince (1992)) is exploited in conversation: the fall-rise is used (and sometimes misused) by speakers to mark brand-new discourse units as if they were already part of the relevant common context.
3:30-3:45break (tea/coffee & biscuits)
3:45-4:15Lars-Olof Delsing, "Possessive constructions of the type 'huset hans Per' "
4:15-4:45Inger Rosengren, "Rethinking the adjunct"
4:45-5:00break (tea/coffee and biscuits)
5:00-5:30Christer Platzack, "The vulnerable C-domain"
Abstract: A close examination of the language of patients with Broca's aphasia and children with SLI reveals that large parts of their syntax are preserved, and that it is mainly the C-domain of the clause that is deviant. This observation, is in concord with recent ideas within the Minimalist program regarding the connection between internal grammar and other cognitive systems; cf. Platzack (1999), who proposes three interface levels corresponding VP, IP and CP, resp. Grammatical properties like V2, complementizers and topicali-zation are examples of syntactic phenomena related to the C-domain, whereas pre-/postposition, the order main verb - object and object - adverb-ials are accounted for at other domains. For Swedish, available data suggest that neither children with SLI nor patients with Broca's aphasia have problems with the later types of word order properties, whereas problems with verb second are reported in almost every study of these types of speakers (Hansson 1998, AhlsÈn & Dravins 1990). Grodzinsky (1999) concludes from a cross-linguistic study of patients with Broca's aphasia, that "the currently available data lead to the view that agrammatic aphasic patients produce [syntactic] trees that are intact up to the Tense node and 'pruned' from this node and up". Tense is presumably identical to C in my framework. For Broca's aphasics there is a direct connection between a brain deficit and the deviant syntax, and for SLI-patients, Kabani et al (1997) have reported a signi-ficant difference with a control group regarding the grey/white matter ratio. In this connection it is interesting to note that Håkansson & Nettelbladt(in press) have observed that L2 learners of Swedish and Swedish children with SLI are similar in acquiring verb second in a step-wise and effortful way, by making use of a canonical subject-verb order strategy. This behavior differs markedly from normally developing L1 children. If we infer from the findings of Dehaene et al (1997) and Kim et al (1997) that distinct cortical areas within Broca's area are activated for L1 and L2, it is perhaps possible to identify the area activated in L1 as an area responsible for the automatic use of the C-domain, and to identify this area with the area hampered in patients with Broca's aphasia, as well as the area hampered in children with SLI. The scientific study of any object in the real world involves both theoretical understanding and experimental proof. There is no lack of linguistic theories, but it has so far been hard to link the theoretical proposals to experimental data. As I hope to show in my talk, a combination of theoretical approaches, studies ofgrammatical impairment, and recent advances in neuroimaging technology, may offer a way to bridge the gap.
MIT Phonology Circle
Friday, October 29, 1999, 3:30-5:00, MIT E39-335 (conference room)Diane Brentari
Purdue UniversityThe role of phonetics in sign language phonology:
Differences between signed and spoken languages Abstract: This talk will address some specific ways that the articulatory and perceptual systems that serve signed languages have shaped the structures and constraints of sign language phonology. By doing this I will address the following two general questions: "How far into the phonology do the effects of the phonetics reach?" and "At what level of description are phonological units equally applicable to signed and spoken languages?" Examples of the differences covered will be: the relationship between root nodes and timing units, the nature of "consonant" and "vowel" information in the signal, and the operations for calculating complexity at the level of the syllable and the prosodic word. By providing evidence in each case for the penetration of phonetic structure on the phonological system of American Sign Language (ASL), the generalization that emerges is that phonetics does have a strong effect on the phonological system of languages, but this effect is balanced by an abstract set of requirements on the stream of linguistic information. These requirements divide the stream into units efficient for both grammatical productivity and for linguistic parsing.
For more information, contact czoll@mit.edu.
UMass, Amherst: Linguistics Colloquium
Friday, October 29, 1999, at 3:30pm in Machmer W-24, followed by drinks and munchies in the department lounge and dinner afterwards.Colin Phillips
University of DelawareGrammar, Parser, Resources -- Which is the Odd One Out? Abstract not available. For more information, contact the Department of Linguistics at UMass.
MIT LING-LUNCH
Thursday, October 28, 1999 at 12:10, MIT E39-335 (conference room)Christer Platzack
Lund UniversityShortest Move at the C-Domain Abstract: Depending on the form of the object (wh/DP), languages display both internal and cross-linguistic word order variation in object initial sentences. In my talk I will discuss four different systems, illustrated by English, Swedish, Italian and Finnish, respectively; in all the examples, Mary is the subject.(1) English: Who had Mary kissed? John, Mary had kissed
(2) Swedish: Vem hade Mary kysst? Johan hade Mary kysst
who had M. kissed J. had M. kissed
(3) Italian: Chi ha visto Maria? Giovanni Maria lo ha visto
who has seen Mary John Mary him has seen
(4) Finnih: Keta Marja rakastaa? Jussia Marja rakastaa
who Mary loves Jussi Mary loves
The purpose of my talk is to show how this variation follows from specific language properties in interaction with two independently motivated universal principles, Shortest Move and the Principle of Minimal Compliance (Richards 1998).
For more information, contact karlos@MIT.EDU.
MIT Phonology Circle
Friday, October 22, 1999, 3:30-5:00, MIT E39-335 (conference room)Lisa Lavoie
Harvard UniversityPhonetics and Phonology of Consonant Strength: Matches and Mismatches Abstract: In this talk, I will discuss some of the results of my very recent doctoral dissertation (Lavoie 2000). I will focus my talk on the acoustic results although the dissertation abstract I include below also mentions the articulatory findings.Consonant weakening, or lenition, is comprehensively examined in this dissertation which consists of acoustic and articulatory phonetic studies of consonant strength, as well as a cross-linguistic survey of phonological patterns of lenition and fortition. A goal of this study is to locate phonetic parallels to the phonological patterns. I examine the strength of consonants of American English and Mexican Spanish in the two cross-cutting environments of position in word and position with respect to stress to determine the relative contribution of each positional factor.
In the acoustic study, I analyze durations, intensity and spectral properties in closely-matched disyllabic stress pairs. In the articulatory study, I measure degree of linguopalatal contact using electropalatography (EPG) in real and nonsense words. I find clear evidence for strong and weak positions, as well as the phonetic patterns that characterize consonant strength. Somewhat surprisingly, I find that some phonetic characteristics pattern as general indicators of manner of articulation while others pattern clearly by position. The positional factors are found not to augment each other's effects, except in English flapping. Despite robust phonetic effects in the pre-stress position, stress is not a factor that conditions historical change whereas duration is.
In addition, I test numerous hypotheses about consonant strength and find evidence for aspects of several prevalent views of lenition. The acoustic and articulatory studies reveal both similarities to phonological weakening and differences. An elaborated sonority hierarchy best describes phonological weakening while a scaled down hierarchy best describes phonetic weakening. While gestural duration and magnitude are not found to decrease in tandem for phonetic weakening as Articulatory Phonology predicts, Articulatory Phonology accounts very nicely for the variation in weakening outcomes. Phonologically, voicing and fricativization are common steps in lenition, but the phonetic studies find little evidence of either. I attribute the frequency of voicing lenition to the documented tendency for shorter segments to be perceived as voiced and I attribute the frequency of lenition desribed as fricatization to reinterpretation of an incomplete seal as a fricative. I argue, based on the phonetics and phonology, that the prototypical weakening is approximantization.
For more information, contact czoll@mit.edu.
MIT Phonology Circle
Friday, October 22, 1999, 3:30-5:00, MIT E39-335 (conference room)Miklos Torkenczy
E. Lantos U in Budapest
For more information, contact czoll@mit.edu.
MIT Linguistics Colloquium
Friday, October 15, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Julie Sedivy
Brown UniversityInvestigating model-based on-line semantic interpretation Abstract: Much of the work in language processing over the past two decades has focused on questions involving the real-time computation of syntactic structures. We have known for a long time now that humans attempt to construct syntactic representations with a fine-grained degree of incrementality, making decisions based on partial information - a fact which results in frequent local syntactic ambiguities. The consequences of this basic property of the processing system have been and continue to be extensively studied in the psycholinguistic literature. Similar advances have not been made, however, in the on-line study of the semantic interpretation of linguistic expressions. For instance, the very general question of what constitutes the basic processing units for semantic interpretation has not even been addressed until very recently. In this talk, I will begin by presenting some data from foundational experiments in on-line interpretation conducted with former colleagues from the University of Rochester that demonstrate that referentially-based interpretation of language is conducted in a highly incremental (i.e. at least word-by-word) fashion that takes into account the structural relations between linguistic expressions. I will show that this general property of semantic processing leads to semantic indeterminacies that do not coincide with syntactic ambiguity. This is true, for instance, for the interpretation of scalar adjectives such as "tall", whose extension cannot be determined independently of information either about the head noun, or some salient contextual factors. I will present data showing that such expressions are indeed interpreted incrementally, with information from a visually presented context being recruited to resolve the indeterminacy. To the extent that systematic effects of context on incremental interpretation can be identified, it is desirable to specify the relationship between the linguistic representations and the contextual information. A number of possible hypotheses pertaining to this question will be raised and evaluated against further experimental data. In particular, I will investigate the possibilities that these contextual effects arise from direct presuppositions encoded as part of the meaning of nominal modifiers, or more indirectly, via the extremely rapid use of implicatures in the course of on-line interpretation.
For more information, contact asrackow@mit.edu.
UMass, Amherst: Linguistics Colloquium
Friday, October 15, 1999, at 3:30pm in Machmer W-24, followed by drinks and munchies in the department lounge and dinner afterwards.Norvin Richards
MITDependency Formation and Directionality of Tree Construction Abstract not available. For more information, contact the Department of Linguistics at UMass.
MIT LING-LUNCH
Thursday, October 14, 1999 at 12:10, MIT E39-335 (conference room)Ileana Paul Topic and focus in Malagasy Abstract: Drawing on data from Italian, French and English, Rizzi (1997) argues for an articulated CP structure where topic>focus>topic. In this talk, I will discuss data from Malagasy (western Austronesian) which support this proposal. In particular, I will show that Malagasy (a VOS language) has pre-verbal positions for topicalized and focussed XPs. Moreover, there are two distinct topic positions which sandwich the focus position.Although I discuss data from topicalization, the focus construction will be the central topic of this talk. A typical example is in (1).
(1) i Soa no nisasa tanana.
Soa foc washed hand
'It's Soa who washed her hands.'
'The one who washed her hands is Soa.'
Descriptively, focus is limited to subjects and adjuncts, like other extraction processes in Malagasy. I will show, however, that what is being fronted is in fact a predicate (Soa in (1)). The presuppositional clause is a headless relative in subject position (no nisasa tanana). In other words, the focus construction in Malagasy resembles an English pseudo-cleft. This analysis of focus parallels recent "predicate fronting" approaches to VOS word order (Pearson 1996, Rackowski 1998).
In the second part of the talk, I turn my attention to apparent multiple clefts, such as (2).
(2) Omaly i Soa no nisasa tanana.
yesterday Soa foc washed hand
'Yesterday it was Soa who washed her hands.'
I show, however, that only the first element (omaly 'yesterday) in (2) is focussed. The second (Soa) is a topic. I discuss a range of data to argue for this distinction. Thus as well as having a topic position above focus, Malagasy allows for a "lower" topic.
For more information, contact karlos@MIT.EDU.
MIT Phonology Circle
Friday, October 8, 1999, 3:30-5:00, MIT E39-335 (conference room)Bert Vaux
Harvard UniversityUyghur Raising and the Nature of (Under)specification Abstract: Transparency and blocking effects in harmony systems are traditionally analyzed in terms of underspecification and prespecification respectively (Clements 1976, Clements and Sezer 1982, etc.). However, the validity of underspecification theory has recently been called into question (cf. Mohanan 1991, McCarthy and Taub 1993, Calabrese 1995, Halle 1995, Steriade 1995), though these works do not specifically address the issue of harmony systems. In this paper I suggest that underspecification and prespecification are also unable to account for the range of attested harmony systems, and conclude that underspecification as a theoretical construct should be abandoned.The crucial piece of evidence on which this paper focuses involves the Uyghur disharmonic suffix -chae-. (Uyghur vowel harmony is very similar to the familiar Turkish system.) Even though ae has a [+back] harmonic counterpart a, the modal suffix /-chae/ invariably surfaces as [-back], regardless of the [back] specification of the root to which it attaches: cf. taG-chae 'as big as a mountain' (expected form *taG-cha) (Hahn 1991). The disharmonic vowel in -chae then spreads its own [back] specification to following vowels, e.g. /kita:b-chae-m-DA/ 'in my booklet' [kitapchaemdae]. (Capital letters denote harmonic segments.) Since the vowel in /-chae/ does not alternate for backness, it should be underlyingly specified as [-back] according to the conventional treatment of disharmonic vowels (cf. Clements and Sezer 1982).
Uyghur also possesses a rule that raises low vowels in medial open syllables, e.g. /bala-lAr-I/ 'his children' [baliliri] (Hahn 1991:84). The -chae suffix undergoes Raising, as expected: cf. /naey-chae-DA/ 'child-loc.' [naeychidae]. We predict that when -chae- undergoes Raising it should remain disharmonic, since its underlying [-back] specification is not affected. Forms such as /kita:b-chae-da/ 'booklet-loc.' [kitapchida] show that this prediction is incorrect: when the /ae/ raises to [i], it becomes transparent to [back] harmony, allowing the [+back] of the preceding /a/ to spread to the harmonic vowel in the following /-DA/ suffix.
The problem is that the prespecification analysis misses the connection between [i] produced by Raising and [i] derived from underlying /i/, which is transparent to [back] harmony (cf. /ishaeG-I-GA/ 'to his donkey' [ishaeghighae]). What is called for is a theory of harmony that evaluates the role of the [i] in the vowel system as a whole, regardless of its origins. Calabrese's 1995 theory of full specification does just this, allowing us to account straightforwardly for the Uyghur data. Uyghur [back] harmony, for example, is analyzed as spreading of contrastive [back]. Since the rule is sensitive only to contrastive [back] specifications, it ignores segments that are not contrastive for [back], such as the neutral vowel [i]. Crucially, this holds for [i] whether it results from underlying /i/ or from underlying low vowels that have undergone Raising. We therefore predict (correctly) that the [-chi-] allomorph of /-chae-/ should be transparent to [back] harmony, because its [i] is not contrastive for the feature [back]. Our theory furthermore predicts the nonexistence of a language exactly like Uyghur, save that the output of Raising a disharmonic suffix remains disharmonic.
For more information, contact czoll@mit.edu.
MIT LING-LUNCH
Thursday, October 7, 1999 at 12:10, MIT E39-335 (conference room)Alec Marantz,
MITThe Pieces of Derivation Abstract: The existence of a special word-building grammar would present a challenge to any syntactic theory. Why should there be two sorts of mechanisms for combining the atoms of language, one working inside and one outside words? Given that periphrasis is often used for the same compositional meanings as the combination of morphemes within words, how does the grammar draw the line between words and phrases? If all word-building can be shown to be syntactic, this challenge disappears and new challenges arise. While inflectional morphology has been well-integrated into syntax recently, derivational morphology still seems special. Apparently unlike syntax and inflectional morphology, derivational morphology seems often to be unproductive or semi-productive and to lack structure/meaning transparency.In this talk, I will argue that the theory of Distributed Morphology provides an account of derivational morphology that allows a full reduction of morphology to syntax, enabling the Minimalist Program's claim that the syntax is the unique generative engine in the grammar. Distributed Morphology combines the "late insertion" insights of Anderson's A-Morphous Morphology with the morphemes as pieces principle of Lieber's (and others') Lexical Morphology.
An explanatory approach to derivational morphology requires making two distinctions. First, following R. Beard's insights about "separation," we must separate the phonological forms of derivational morphemes from the syntactic/semantic features that they spell out. This separation makes derivational morphology paradigmatic in a deep sense and changes the questions one asks about possible combinations of morphemes.
Second, we must distinguish the syntactic heads that create the "lexical" categories N, V, and A out of roots from heads that attach to constituents that already have syntactic categories. Much of the apparent unproductivity and idiosyncracy of derivational morphology is attributable to the heads that attach directly to roots.
In sum, the pieces of derivation look just like the pieces of inflection: terminal nodes in syntactic trees consisting of bundles of syntactic and semantic (but no phonological or idiosyncratic "lexical") features.
For more information, contact karlos@MIT.EDU.
Boston UniversityPhonology Circle
The BU Phonology Circle will meet on the first Monday of each month, starting October 4th, at 7 PM, in the BU Linguistics Library (704 Commonwealth Avenue). Anybody who is interested in phonology is welcome. The main focus of the last couple of meetings has been language variation and change within the framework of Optimality Theory.
For more information, please contact Larry Ichimura, LIchimura@aol.com, 978-531-6102.
MIT Linguistics Colloquium
Friday, October 1, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Norbert Hornstein
University of MarylandSidewards Movement; Adjunct Control and Parasitic Gaps Abstract: In 'Movement and Control' , I argue that obligatory control (OC) is a function of movement. Cases of adjunct control have all of the diagnostic properties of obligatory control, hence, must be formed by movement. The paper starts with a discussion of how these cases can be integrated into a movement analysis by exploiting sidewards movement in the sense of Nunes (1995). The generalization that emerges is that sidewards movement is possible just in case it involves movement via a theta position. This is illustrated with a discussion of parasitic gap phenomena which are also seen to be amenable to a sidewards movement analysis as Nunes (1995) argued. The origin of CED effects is then discussed as is the structure of nominal complementation and the prohibition against raising in noun.
For more information, contact asrackow@mit.edu.
UMass, Amherst: Linguistics Colloquium
Friday, October 1, 1999, at 3:30pm in Machmer W-24, followed by drinks and munchies in the department lounge and dinner afterwards.Anna Szabolcsi
New York UniversityDirect and Inverse Scope in Overt Syntax Abstract not available. For more information, contact the Department of Linguistics at UMass.
MIT Phonology Circle
Friday, September 24, 1999, 3:30-5:00, MIT E39-335 (conference room)Ben Bruening
MITMorphological Templates and Phonological Opacity
For more information, contact czoll@mit.edu.
MIT LING-LUNCH
Thursday, September 23, 1999 at 12:10, MIT E39-335 (conference room)Elena Benedicto and Diane Brentari
Purdue UniversityVerbal Classfiers as Heads of Functional Projections: Evidence from ASL Abstract: This paper explores verbal classifiers in ASL and their effect on the predicate argument structure and its aspectual interpretation, providing further support for a universal, syntactically encoded theory of the interpretational aspectual properties of predicates.We propose a syntactic analysis for two types of classifiers in ASL: instrumental classifiers (ICLs) and extent descriptive classifiers (eDCLs). We claim that these two classifiers head the functional projection above VP and encode the semantic property of telicity. We further suggest that both classifiers are associated with the internal argument of the verb. The selectional restrictions that exist between them hold via Spec-Head agreement within the (telic) functional projection. The difference between these two types of classifiers is that ICLs have a [+Case] feature, while eDCLs do not; thus, the constituent in their Spec can check Case with ICLs, but not with eDCLs. As a consequence, the single argument of eDCLs will have to further raise to a position where it can check Case. This will result in eDCLs manifesting properties associated with unaccusativity.
MIT Phonology Lecture
Monday, September 20, 1999, at 3:00: MIT, 4-159Aditi Lahiri
Unversity of KonstanzProcessing of morphologically complex words Abstract:There are two sides to the discussion of the comprehension of morphologically complex words: processing and representation. The original debate, based primarily on visual word recognition, was focused on two entirely different hypotheses: (a) the prefix-stripping or decomposition hypothesis where initial access was made on the basis of the stem, and (b) the full-listing hypothesis which assumed left-to-right access with all derived words listed, presumably alphabetically. In recent years, particularly in the auditory modality, the decomposition view has gained ground. Very briefly, the notion of a 'stem' has steadily gained weight, but here the stem is not the access unit, but a representational unit. That is, on encountering the first segment, the comprehension system does not necessarily wait to locate a stem before accessing the lexicon. Nevertheless, on the recognition of a prefixed word, the decomposed entry with the stem is activated. Further, the decomposed representation (and hence access) of the affix+morpheme is not affected by any phonological 'mutation' of the stem or affix (e.g. {tone} - {tonic}) The crucial assumption is that morphological decomposition is acceptable only if the relationship is semantically transparent. This is based on priming evidence where {flatter} does not prime {flat} (being semantically unrelated), but {writer} primes {write}. One problem is that in most models, it is unclear why morphological decomposition is necessary given that semantic transparency is the crucial factor. Second, it is also unclear precisely how morphologically (and semantically) related forms which are phonologically altered are represented. Third, since partial phonological (auditory) information is sufficient to activate words which are *not* necessarily semantically related, it is disconcerting to find that there is no priming in semantically unrelated words with phonetic overlap. This talk presents the phonologist's attempt to resolve some of the above issues in the comprehension of morphologically complex words.
For more information, contact Michael Kenstowicz, kenstow@MIT.EDU.
MIT Phonology Lecture
Friday, September 17, 1999, at 4:00: MIT, 56-154Aditi Lahiri
Unversity of KonstanzThe dental preterite in Germanic: reanalysis and levelling Abstract:The origin of the dental in the preterite of the weak verbs in Germanic, has been the source of a great deal of controversy. One assumption is that it comes from IE *dhe- 'to do', whereby a complex verbal construction like 'save + do' has led to a reanalysis incorporating the dental of the stem of *dhe- as part of the inflectional ending. The evidence has come mainly from the different manifestations of the dental in the various Germanic languages. In this talk I am interested in the reanalysis of the verb as an inflectional suffix. Based on phonological and morphological evidence, I will argue that the reanalysis is not a simple matter of grammaticalization where an auxiliary is reduced to a suffix. The reanalysis involves the interpretation of the dental in the auxiliary not only as a marker of the past tense, but also as a 'class' marker for the weak verbs. This means, that the dental consonant is separated from its person/number endings and treated as an independent morpheme. The reanalysis is crucially linked to the behaviour of other nominal 'stem extensions' and the verb-deriving suffix /j/, which trigger umlaut, gemination and syncope, all governed by metrical structure. This analysis accounts for the lack of gemination in the 2nd and 3rd person singular, the imperative, and the preterite, and also explains the discrepancy between some inflectional classes of the /ja-/ nouns and the weak verbs. Apparent analogical levelling in Old High German and Old Norse weak verbs, and the development of the 'special' past tense forms in Middle English are also accounted for.
For more information, contact Michael Kenstowicz, kenstow@MIT.EDU.
UMass, Amherst: Linguistics Colloquium
Friday, September 17, 1999, at 3:30pm in Machmer W-24, followed by drinks and munchies in the department lounge and dinner afterwards.Paul Smolensky
Johns Hopkins UniversityOpacity, Turbid Representations, and Output-Based Explanation Abstract not available. For more information, contact the Department of Linguistics at UMass.
MIT Phonology Lecture
Thursday, September 16, 1999, at noon: MIT, 56-114Aditi Lahiri
Unversity of KonstanzVariability in speech: place of articulation and duration Abstract:'Noise' in the speech signal can come from various sources. Both contextual assimilations leading to neutralization as well as allophonic alternations change the idealised norm. For instance a coronal nasal [n] could easily assimilate in place to a following labial stop [b]. If in a language, a 'real' [m] is also present, contrasting with [n], then the assimilation leads to neutralization. If, on the other hand, there is no phoneme [m], the assimilated consonant would be an allophone. The literature has focused in general on the completeness of assimilation - i.e., if the assimilation is not complete, it is possible that the acoustic cues for a 'real' labial nasal [m] and one with a partial coronal gesture would be different. Very little has been discussed on the possibility of variation between neutralisation and allophonic alternations. One could imagine that the acoustic distinction between two phonemes is more striking than when there is only an allophonic variation to be distinguished. In this talk, I will discuss whether the acoustic properties differentiating two segments differ as a function of phonological status. Two issues will be discussed. First, do the acoustic properties distinguishing two segments differ depending on whether they contrast phonemically or not? Second, in the case of neutralizing a phonemic contrast, is there a difference depending on whether the assimilation is lexical or postlexical? For instance, would it matter if the assimilation takes place across morphemes but within a word, as against assimilations across words? Several crosslinguistic experiments will be discussed involving the phonetic dimension of duration for geminate consonants and the place of articulation feature [anterior]. The general claim is that the phonological status does not affect the realisation of a phonetic distinction, which could account for why a postlexical allophonic alternation can lead to the introduction of a phonemic contrast when there is a loss of the assimilatory trigger.
For more information, contact Michael Kenstowicz, kenstow@MIT.EDU.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department.Ling-lunch will take place every Thursday at 12, starting September 9.
Anybody interested in presenting during the coming semester should send Karlos Arregi, karlos@MIT.EDU, e-mail.
Thursday, September 9, 1999 at 12:10, MIT E39-335 (conference room)Sharon Inkelas and Cheryl Zoll
MITReduplication as Double Stem Selection Abstract: Phonological Theories of Reduplication (including Marantz 1982, Clements 1985, Mester 1986, McCarthy and Prince 1986, Steriade 1988, Kiparsky 1986 and 1997) consider RED to be a morpheme whose segmental content is filled in by the phonology. This theme continues in Correspondence Theory (McCarthy & Prince 1995), which claims further that phonological identity between RED and BASE is mediated and enforced by correspondence between the two constituents. We show that (1) there is no morpheme RED, since so called reduplicants are often morphologically complex; and (2) identity between Base and Reduplicant must be at the morphosyntactic level, not in the phonology. We propose a morphological theory of reduplication: Reduplication as Double Stem Selection (DSS) (developing proposals by Downing 1997 et seq. and Hyman, Inkelas & Sibanda 1999; cf. (Yip 1995)'s REPEAT(Stem).) that better accounts for the range of patterns of reduplication found cross-linguistically. A re-examination of phenomena that have been claimed to require phonological correspondence further reveals that once the morphology of reduplication is clearly understood, ostensible cases of "back-copying" are in fact straightforward normal application of phonological processes, leaving no role for BR correspondence in phonology.
Harvard International Symposium on Korean Linguistics
Friday, July 16 - Sunday, July 18, 1999 at Harvard UniversityFields: Korean Phonetics, Phonology, and Morphology/ Korean Syntax/ Korean Semantics/ Korean Acquisition/ Korean Historical Linguistics/ Korean Computational Linguistics/ Korean Lexicography
The 11th North America Conference On Chinese Linguistics
at Harvard University
Friday, June 18-20, 1999 at Harvard University Papers on a wide variety of topics, such as historical linguistics, phonology, syntax, semantics, discourse/pragmatics, sociolinguistics, computational-linguistics, psycholinguistics, and language acquisition. Our goals are to provide participants with an opportunity to present their works in Chinese linguistics and to keep up with recent developments in the field.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, May 13, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Beatriz Fernandez & Ken Hale "Ergative displacement" Abstract:This talk will be concerned with the phenomenon known as Ergative Displacement in four unrelated languages: Navajo, K'ichee', Chukchi and Basque. Under certain conditions, these languages exhibit behavior in their agreement systems which in all probability is governed by the same grammatical principles. Briefly, except in the Navajo case, what is involved is the reduction of agreement morphology in transitive clauses, leaving only the so-called nominative (also called absolutive) agreement, forcing the ergative argument to "shift" to absolutive agreement. In the first twenty minutes of the period, an essentially descriptive overview of the K'ichee', Chukchi, and Navajo will be presented (framed in the Case Binding Theory of Bittner, 1994). The remainder of the period will be devoted to Basque, shifting the theoretical perspective to the Minimalist Program of Chomsky (1998). The following ideas will be developed.
Firstly, assuming that third person DPs lack person agreement features, we will claim that absolutive agreement features are obligatorily checked by the absolutive argument in ergative languages, or, when this fails for some reason,by the ergative argument as a Last Resort operation. Secondly, it will be claimed that some verbal prefixes can behave as agreement-checkers, satisfying the requirement that absolutive agreement features must be checked obligatorily. The immediate consequence of these claims is that Case Checking is linked to agreement only in "optimal" cases. Under other conditions, Case and Agreement may separate, departing from the canonical Case-Agreement configuration.
For more information, contact mhcote@mit.edu.
MIT RLE Speech Communication Group Seminar
Wednesday, May 12, 1999 at 3:00 PM at MIT in the Grier Room B, MIT 34-401Joanne L. Miller
Department of Psychology, Northeastern University"Some characteristics of the internal structure of phonetic categories" Abstract: A widely held assumption in the speech perception literature for many years was that during the course of processing listeners derive an abstract phonetic representation and, in doing so, discard information about the fine-grained detail of the speech signal. However, more recent research has shown that the representations of speech are much richer than this emphasis on abstract categories would suggest, and that listeners apparently retain in memory a substantial amount of fine-grained acoustic-phonetic information. One line of evidence for the richness of phonetic representations comes from research showing that members of a given phonetic category are not perceptually equivalent. Instead, phonetic categories are internally structured in a graded fashion, with some members of the category perceived as better exemplars (as more "prototypical") than others. In this talk I will first review some of our work on this issue that establishes the generality of such internal structure across a variety of phonetic categories and, in addition, demonstrates that this structure is highly sensitive to acoustic-phonetic contextual factors (such as speaking rate). I will then present some new findings showing that internal category structure is much less sensitive to a higher-order linguistic contextual factor (lexical status). Finally, I will present data from ongoing experiments that begin to assess the role that internal category structure plays in on-line phonetic categorization.References
Volaitis, L.E., and Miller, J.L. Phonetic prototypes: Influence of place of articulation and speaking rate on the internal structure of voicing categories. Journal of the Acoustical Society of America, 1992, 92, 723-735.
Hodgson, P., and Miller, J.L. Internal structure of phonetic categories: Evidence for within-category trading relations. Journal of the Acoustical Society of America, 1996, 100, 565-576.
Miller, J.L., and Eimas, P.D. Internal structure of voicing categories in early infancy. Perception & Psychophysics, 1996, 58, 1157-1167.
Miller, J.L., O'Rourke, T.B., and Volaitis, L.E. Internal structure of phonetic categories: Effects of speaking rate. Phonetica, 1997, 54, 121-137.
For more information, contact MARILYN@speech.mit.edu
MIT MorphBeer
Monday, May 10, 1999 at noon in E39-335 (conference room) at MIT.Gunnar Olafur Hansson
U.C. Berkeley"The Synchrony and Diachrony of Opaque Sound Patterns: Clues from Yowlumne (Yawelmani) Vowel Harmony" gunnar@socrates.berkeley.edu
http://socrates.berkeley.edu/~gunnar
Abstract: As a consequence of the general shift towards output-oriented models in phonological theory in recent years, especially with the advent of Optimality Theory, the problem of phonological opacity has come into renewed focus. Whereas serialist models were able to handle opaque sound patterns quite straightforwardly by mimicking their relatively simple diachronic origin, this option is not available in an output-based model. In order to decide how to accomodate opaque sound patterns in phonological theory, it is important to take into account all available evidence on their nature and synchronic status, such as evidence on how opaquely conditioned alternations are learned and internalized in the synchronic grammar. In this paper I present what I take to be such evidence, namely the particular pattern in which the system of Vowel Harmony in Yowlumne appears to have been breaking down over the past 60-70 years.As is well known, the conditioning of Yowlumne (Yawelmani) Vowel Harmony -- left-to-right rounding harmony between vowels OF THE SAME HEIGHT -- is rendered opaque by a general phenomenon of Long-Vowel Lowering. All descriptions and analyses of this system are based on data gathered in the 1930s. I will present new data, elicited from two of the last remaining native speakers, which show that the system has undergone significant and systematic changes. In several suffixes, the harmony alternation is retained intact; in others, it has either been lost altogether or redefined as transparent (i.e. sensitive to output rather than input height). All the suffixes that have been affected in either of these two ways are low-vowel suffixes (with a/o rather than i/u), virtually all of them non-subcategorizing, i.e. not imposing a specific prosodic template on the preceding root.
I argue that a simple, unified, learnability-based explanation underlies this particular diachronic development. For any given suffix, the alternation has been lost if its conditioning could not be captured by some surface-true (and thus transparent) generalization. All things being equal, this would have been true of ALL low-vowel suffixes, but in several cases the alternation has been "rescued" as an accidental side effect of the prosodic morphology. If a particular prosodic template is imposed on the root, Long-Vowel Lowering may be inapplicable, in which case the harmony alternation is transparently conditioned and thus retained.
This development suggests that opaque alternations should perhaps be treated on a one-by-one (affix-by-affix) basis. At the stage where learners are failing to acquire the "correct" pattern, they appear to be internalizing not a general pattern of vowel harmony (and the constraints/rules/rankings that drive it), but simply surface alternations in the exponence of individual suffixes. I will discuss the implications that this may have for synchronic phonological theory.
For more information, contact czoll@MIT.EDU.
MIT Linguistics Colloquium
Friday, May 7, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Susan Ehrlich
York University"(Re)Constructing Innocence: Language in Sexual Assault Trials" Abstract: This paper explores the grammatical means by which a defendant in both a sexual assault criminal trial and a university tribunal attempts to construct himself as innocent. I assume that language, particularly in these kinds of adjudication settings, does not simply reflect some pre-existing reality -- that is, it is not some transparent medium through which stable 'facts' or 'accurate' descriptions are conveyed. Rather, through language, all of the participants in these adjudication processes actively construct their version of the past events under investigation. In this case, the defendant's testimony does not simply reflect his defense of innocence, but is also involved in constituting the 'facts' of the case. Capps and Ochs (1995) discuss the role of narrative in legal settings: "Particularly with respect to psychological issues such as alleged abuse and harassment -- personal testimonies -- narrative accounts of what happened are often the sole basis for determining a verdict. Such cases can involve radically different accounts of what happened. And on the basis of divergent versions of events, juries, judges and adjudicators construct a narrative that is plausible and coherent in their eyes, but the truth is beyond their reach. In this sense rendering a verdict is analogous not to ascertaining the facts of a case but to determining an official story." Official stories, however, are not just ascertained on the basis of their plausibility and coherence. Following Gal (1994), I assume that a society's institutions are not neutral arenas for talk: "they are structured along gender lines to lend authority to reigning classes and ethnic groups but specifically to men's linguistic practices." Thus, in this paper, I am not only interested in the defendant's construction of innocence through what I'm calling a grammar of non-agency, but also the way in which his particular characterization of events is afforded legitimacy and authority -- the way it becomes the 'official story.'
For more information, contact asrackow@mit.edu.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, May 6, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Cristina Altman
University of São Paulo"The Early Days of the 'Cercle Linguistique de New York' -- Jakobson, Lévi-Strauss and others in the eyes of a Brazilian linguist" Abstract:The New York Linguistic Circle was created in 1943 as an extension of the 'École Libre des Hautes Études' which had been organized in New York city by a group of French and Belgian refugees from the Nazi-occupied Europe. According to the School's 'Catalogues', conserved nowadays at the MIT Institute Archives and Special Collections, the École Libre offered various courses in the humanities which attracted various scholars such as Paul L. Garvin (1919-1994), Charles F. Hockett (b. 1916), Thomas Sebeok (b.1920) and also the Brazilian linguist Joaquim Mattoso Câmara (1904-1970). It was in connection with the École Libre that Roman Jakobson gave for the first time, between May and June of 1942, his lessons on 'Le son et le sens des mots', which appeared in book form only many years later, in 1976. These lectures have been celebrated in linguistic historiography for their subsequent influence rather than for their contents. In fact, it was on this occasion that Jakobson and Claude Lévi-Strauss, who had come from Brazil where he had done ethnological work, first met and exchanged ideas. This singular event, however, has cast a big shadow on any other contacts between Brazil and American linguistics, of which, the one between Jakobson and Mattoso Câmara was much more consequential, at least as far as the implementation of 'modern linguistics' in Brazil and in South America during the 1950s and 1960s is concerned.
During his stay in United States between 1943-1944, besides attending Jakobson's lectures on general linguistics, Mattoso Câmara made contact with several other linguists, with whom he would correspond later when back to Brazil, among others: Leonard Bloomfield, Morris Swadesh, George Trager, Louis Herbert Gray, Giuliano Bonfante, Yuen-Ren Chao, Edgar H. Sturtevant and Franklin Edgerton. In addition, Mattoso Câmara participated in the first meetings of the Linguistic Circle of New York in the quality of one of its founders. These activities have been dully recorded by him in a notebook, conserved today at the Catholic University of Petrópolis, close to Rio de Janeiro. Mattoso Câmara's notes bear witness not only to what extend he had absorbed Jakobson's conception of linguistic structure, but also aspects of American intellectual life in the mid-20th century. The impact of Jakobson and of American Linguistics on the work of Mattoso Câmara and its subsequent influence on developments within the description of Brazilian Portuguese and of Brazilian indigenous languages are the focus of this talk.
Selected Bibliography
Altman, Cristina. 1998. A Pesquisa Lingüística no Brasil (1968-1988). São Paulo: Humanitas.
Chomsky, Noam. 1997. "Knowledge of History and Theory Construction in Modern Linguistics". São Paulo Conference. DELTA 13:103-128.
Jakobson, Roman. 1963. "Le langage commun des linguistes et des anthropologues" Essais de linguistique générale, 25-42. Paris: Minuit. ("Results of the Conference of Anthropologists and Linguists", Indiana Univ., 1952; transl. from English and preface by Nicolas Ruwet.)
----------. 1976 [1942]. Six leçons sur le son et le sens. Paris: Minuit. (Préface de C. Lévi Strauss; Portuguese version by Luís Miguel Cintra, Lisboa: Moraes, 1977; English version by John Mepham. MIT Press, 1978.)
Mattoso Câmara Jr., Joaquim. 1941. Princípios de Lingüística Geral como fundamento para os estudos superiores da língua portuguesa. Rio de Janeiro: Briguiet. (Repr. in 1942; 2nd rev. and enl. ed. Rio de Janeiro: Acadêmica, 1954, 3rd ed. 1958, 4th ed., 1967.)
----------. 1953 [1949]. Para o Estudo da Fonêmica Portuguesa. Rio de Janeiro: Organização Simões.
----------. 1965. Introdução às línguas indígenas brasileiras. Rio de Janeiro: Museu Nacional.
For more information, contact mhcote@mit.edu.
MIT RLE Speech Communication Group Seminar
Wednesday, May 5, 1999 at 3:00 PM at MIT in the Grier Room B, MIT 34-401Sheila E. Blumstein
Brown University"A Theory of Acoustic Invariance Revisited" Abstract: It is the goal of this presentation to present a theoretical framework for consideration of the acoustic properties which define the phonetic categories of speech, and to consider the results from a cross-linguistic study of syllable-final fricatives (Lee and Blumstein, 1999) in terms of this framework. The development of the current framework has its roots in a number of earlier theories of the sound structure of language including the quantal theory of speech (Stevens, 1989), a theory of acoustic invariance (Stevens and Blumstein, 1981), and the acoustic basis of distinctive (phonetic) features (Jakobson et al., 1963). In the current theory the acoustic properties associated with the phonetic categories of speech are: time varying (not static); relative (not absolute); graded (not all or none); and perceptually weighted (not perceptually primary). We propose that for any given phonetic feature there is at least one criterial attribute, which is distinguished from other acoustic cues in that it is minimally affected by sources of variability such as speaker, phonetic context, and speaking rate, and in this sense is 'invariant' or acoustically stable. As such, the criterial attribute defines the boundary constraints of the phonetic category and also provides the means by which listeners can categorize sounds spoken by different speakers, in different phonetic contexts, and at different speaking rates as members of the same phonetic category. It also provides an organizational framework for the sound inventories of language, defining not only the limits of those categories but also how those sound inventories may be structured.The current framework was considered in relation to data obtained from a study of voicing characteristics of syllable-final fricatives (Lee and Blumstein, 1999). We conducted acoustic analyses of syllable-final voiced and voiceless fricative consonants [s z] in American English and French preceded by one of 3 vowel contexts. Each VC stimulus was produced in 3 conditions including isolation and in contexts with following voiced/voiceless stop consonants. An acoustic measure of glottal excitation during the fricative noise based on Stevens et al. (1992) was evaluated, as were three measures of duration, including the fricative noise duration, the preceding vowel duration, and the ratio of fricative duration over the preceding vowel duration. Each stimulus was analyzed for each parameter separately and also in relation to the parameters jointly. Results showed that the measure of glottal excitation was the most 'highly valued' parameter in that it accounted for the highest categorization results (at 91%). The measure of glottal excitation and the duration ratio taken together accounted for 96% of the data. The duration measures, taken alone and together, accounted for no more than 79% of the data. Implications of these findings for the current theoretical framework will be discussed.
References
Jakobson, R., Fant, G., and Halle, M. 1963. Preliminaries to Speech Analysis. MIT Press.
Stevens, K.N. 1989. On the quantal nature of speech. J. Phonetics, 17, 3-45.
Stevens, K.N. and Blumstein, S.E. 1981. The search for invariant acoustic correlates of phonetic features. In P.D. Eimas and J.L. Miller (ed.). Perspectives on the Study of Speech. Lawrence Erlbaum Press.
Stevens, K.N., Blumstein, S.E., Glicksman, L., Burton, M.W., and Kurowski, K. 1992. Acoustic and perceptual characteristics of voicing in fricatives and fricative clusters. JASA, 91, 2979-3000.
For more information, contact MARILYN@speech.mit.edu
Phonology 2000 Symposium - Harvard and MIT
April 30-May 1, 1999 The purpose of the symposium is to gather some 35 leading phonologists to initiate a substantive debate concerning the relative empirical and theoretical merits of the two dominant models of human phonological knowledge, derivational phonology (DP) and Optimality Theory (OT).Central to DP, which was employed by most phonologists until this decade, is the proposition that the surface representation of words is derived in a deterministic fashion from their underlying representations by the application of a series of ordered rules. The introduction of OT has resulted in a drastic realignment of this outlook and of the field of phonology as a whole, in terms both of the questions that are being asked and of the way in which these questions are being addressed. In OT the underlying and surface representations are related by means of violable constraints, which reflect aspects of universal well-formed outputs, and the differences among languages are attributed exclusively to differences in the rankings of the constraints.
To this point, proponents of DP have failed to mount a systematic response to the problems raised or implied by OT. By the same token, the adherents of OT have yet to tackle the challenge of demonstrating that rule-based phonology is inviable. As a result, many important issues that have been raised by the coexistence of these two phonological perspectives have not been adequately addressed. Perhaps the most important among these is the lack of a systematic comparison of the two competing theories, with the aim of determining their relative merits on both formal and empirical grounds.
The purpose of the present symposium is to serve as a starting point for such a comparative evaluation of OT and DP. In particular, we expect to discuss data and formal issues that highlight critical differences between the two theories, with the ultimate goal of determining not only which of the two approaches is to be preferred but also-and more importantly-why. Ideally the participants will leave the symposium with a deeper appreciation of the problems that each model can solve and of the problems that each may have difficulty in solving.
The Phonology 2000 Symposium will be held on Friday, April 30, and Saturday, May 1, 1999. The sessions on the first day will be held at Harvard, and those on the second day will be at MIT. Admission is free, but in order to gain admission to the building on Friday, it will be necessary for individuals not affiliated with Harvard to sign in at the front desk. We apologize for this inconvenience.
SCHEDULE
Friday, April 30, 1999
Lamont Library, Harvard University
9-9:15 Coffee and snacksSESSION 1: General Issues
9:15-9:45 Stuart Davis, Indiana University, "OT: Empirical problems and insights"
10-10:30 David Odden, Ohio State, "Ordering"
10:45-11 Break for snacks
SESSION 2: Syllables
11-11:30 Donca Steriade, UCLA, "Word phonotactics vs. syllabic intuitions"
11:45-12:15 Francois Dell, CNRS, Paris, "Syllabification with epenthesis and without: syllable structure in two Berber dialects"
12:30-2 LUNCH
SESSION 3: Reduplication
2-2:30 Sharon Inkelas, Berkeley, and Cheryl Zoll, MIT, "Reduplication as morphological doubling"
2:45-3:15 William Idsardi and Eric Raimy, Delaware, "Reduplication and underapplication"
3:30-4 Break for more snacks
SESSION 4
4-4:30 Andrea Calabrese, UConn, "Glide formation, gemination, and Sievers' Law in Vedic Sanskrit"
4:45-5:15 Charles Reiss, Concordia, "Acquisition and post-OT phonology"
5:30-6 Michael Hammond, Arizona, "Poetic meter, feet, and acquisition"
Saturday, May 1
MIT 66-110
SESSION 5: Stress9-9:30 Ellen Broselow, SUNY-Stonybrook, "Stress-epenthesis interactions in a constraint-based theory"
9:45-10:15 John Frampton, Northeastern, "On iterative rules"
10:30-11 Break for snacks
SESSION 6: Morphology and phonology
11-11:30 Bruce Hayes, UCLA, "Burnt and Splang: Some Issues in Morphological Learning Theory"
11:45-12:15 Rolf Noyer, Penn, "The basis of bases"
12:30-2 Lunch
SESSION 7: Features
2-2:30 Diana Archangeli, Arizona, "On Constraint Motivation"
2:45-3:15 Keren Rice, Toronto, "Featural markedness"
3:30-4 Break
SESSION 8: General Issues
4-4:30 Michael Kenstowicz, MIT, "Transderivational relations"
4:45-5:15 Mark Hale, Concordia, "Historical Phonology and Phonological Theory in the 21st Century"
There will be 15 minutes for discussion after each paper.
In addition to the speakers above, the invited discussants will include:
Abby Cohn, Cornell
Ben Hermans, Tilburg
Paul Kiparsky, Stanford
Mark Aronoff, SUNY-Stonybrook
Nick Clements, Paris
Elan Dresher, Toronto
Eulalia Bonet, Barcelona
Jim Harris, MIT
Ellen Kaisse, Washington
Alec Marantz, MIT
David Pesetsky, MIT
Sylvain Bromberger, MIT
Steve Anderson, Yale
Jean-Roger Vergnaud, USC
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, April 29, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Beatriz Garza Cuaron
Universidad Nacional Autonoma de Mexico"Linguistic and cultural problems: the case of Chiapas" Abstract:On New Year's Day of 1994, Mexican society discovered that an armed uprising had erupted in the southern state of Chiapas, which shares a border with Guatemala. The so-called Ejercito Zapatista de Liberacion Nacional (EZLN), or the Zapatista National Liberation Army declared war on the Government of Mexico.
On the same day, January 1st., in a state of surprise and general disbelief, Mexico and the rest of the world found out the truth about the armed uprising along with its causes: hunger, misery, lack of justice and the marginalization that for centuries the Indian population has had to bear.
Marcos is the figure from whom we have heard the most through incisive and poetic communiques, which happen to be not only important and interesting, but also intelligent, profound and well-written.
Particularly extraordinary about their organization is the Clandesitine Revolutionary Indigenous, made up of representatives from all of the zapatista communities, including such ethnicities as the Tzotzil, Tzeltal, Tojolabal, Chol, Mam, and Zoque.
In this lecture, we shall see the amerindian languages decrease, its causes and a historical explanation of this problem.
We also will talk about the endangered languages of Chiapas, and, in general, of Mexico and whow the goverment has reacted, in order to have bilingual eduction.
We finish talking about the situation of actual indian writers, its goals and how the society is receving them.
For more information, contact mhcote@mit.edu.
Harvard University - Ford Foundation Acquisition and Learnability Colloquium Series
Thursday, April 29, 1999, at 7:30 PM - Emerson 108, Harvard Yard.There will be a reception after the talk where wine and food will be served and to which everyone is invited. The reception will be held in the Ticknor Lounge of Boylston Hall (Harvard Yard) Directions will be provided.Luigi Rizzi
University of Sienna"Relativized Minimality Effects" Abstract not available at this time.
Boston Colloquium for Philosophy of Science: Norms of Language
Tuesday, April 27, 1999, 2 - 5 PM, Boston University Terrace Lounge, George Sherman Union, 775 Commonwealth Avenue
M o d e r a t o r: Jaakko Hintikka, Boston University RUTH MILLIKAN, University of Connecticut
"Why There Are No Rules of Language"ROBERT BRANDOM, University of Pittsburgh
"Normativity and Modality"
Harvard University - Ford Foundation Acquisition and Learnability Colloquium Series
Friday, April 23, 1999, at 5:20 PM - Sever 103, Harvard Yard. (NOTE the time and room change!)There will be a reception after the talk (7 PM) where wine and food will be served and to which everyone is invited. The reception will be held in the Thompson Room of Barker Centre (intersection of Harvard Street and Quincy Street).
Richard Kayne "Clitic Doubling in French" Abstract not available at this time.
Boston UniversityApplied Linguistics Colloquium
Friday, April 23, 1999, at 4:00 PM in SED Room 259Marina Sbisà
University of Trieste, Italy"Illocutionary Force: Degree of Strength" Abstract: Not available at this time.
For more information, contact Liz Hughes.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, April 22, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Calixto Aguero
MIT"On Pair-List Readings in Questions with Quantifiers" Abstract: Since May (1985) it has been noticed that the sentences in (1) and (2) differ in that the former can be answered in any of the ways in (3) whereas the latter only admits the answer in (4a).(1) Which boy did every girl meet at the park?
(2) Which boy met every girl at the park?
(3) a. Bill (single answer(SA))
b. her boyfriend (functional answer(FR))
c. Mary met Bill, Susan met John, ...(pair-list(PL) answer)(4) a. Bill did
b. *Bill met Mary, John met Susan,...In this talk I will investigate the distribution of Pair-list(PL) readings in questions with quantifiers examining the syntax and semantics of interrogative and non-interrogative quantifiers that support such readings. In particular I will argue that given any of the structures in (5), PL readings arise iff the restriction of the wh-phrase can be reconstructed to some position below the QP.
(5) a. [Whi.....QPj.....tj.....ti]
b. [Whi.....QPj.....ti.....tj]With examples like those in (6), I will show that reconstruction is necessary for PL readings to be available.
(6) a. To which picture of John did he assign every frame? (PL)
b. To which picture of Johni did hei assign every frame? (*PL)(7) a. Which boy seems to have met each girl? (PL)
b. Which boy seems to his mother to have met each girl? (*PL)
c. Which boy seems to my mother to have met each girl? (PL)
d. Which boy seems to his mother to have met which girl? (PL)These examples show two things: 1) that when something blocks reconstruction (e., Condition C (6), PL readings are not available; and 2) that PL readings do not arise when the quantifier can QR and adjoin to IP without intervening the chain of the wh-phrase and its trace since otherwise (6) should have a PL reading. I will further argue that the reason why reconstruction is necessary is that PL readings obtain when a wh-phrase is interpreted as a higher order existential quantifier over functions of type <e, e> (i.e., Skolem functions) in which case there will be a free individual variable providing the argument for the functional variable inside the restriction of the wh-phrase. This individual variable needs to be bound somehow if the sentence is going to be interpretable, reconstructing the restriction below some c-commanding quantifier provides the right opportunity for binding the individual variable with the index of the quantifier.
For more information, contact mhcote@mit.edu.
MIT RLE Speech Communication Group Seminar
Wednesday, April 21, 1999 at 3:00 PM at MIT in the Grier Room B, MIT 34-401Jay Moody
Speech and Hearing Sciences,
Health Sciences and Technology, MIT"Visualizing speech with a recurrent neural network trained on human acoustic-articulatory data" Abstract: A new method has been developed for converting speech sounds into a visual representation of articulation. Images of the tongue and lips are captured using ultrasound and video, respectively. Principal component analysis of the images produces a compact representation of speech production dynamics. When this representation was used to train a recurrent neural network on a corpus of voiced stops, the network produced movement estimates which captured the appropriate bilabial, alveolar, and velar gestures. Because the network averaged over disparate targets associated with similar inputs, it tended to under-represent articulatory variations which did not have significant acoustic effects, in a sense "cleaning up" articulatory variability. In a direct test of the usefulness of the estimated articulatory images, human subjects were shown images of moving lips produced by a network trained on a corpus of spoken digits. Under conditions of simulated hearing impairment, recognition performance rose from less than 50% to over 80% correct when the acoustically-driven network output movies were presented. In light of this performance, and inspired by the rough similarities between artificial and biological neural networks, a new hypothesis for human speech perception is formulated in which articulatory representations are recruited. The hypothesis differs from established motor theories of speech perception in that articulatory recruitment is taken to be learned rather than innate and not functionally essential but nonetheless biologically common.
For more information, contact MARILYN@speech.mit.edu
MIT Phonology Circle
Friday, April 16, 1999, 2:00-3:30, MIT building 66 room 156Elsa Gomez-Imbert, CNRS
Michael Kenstowicz, MIT"Barasana Accent" Abstract: Barasana is a Tucanoan language from the Amazon Basin in Columbia. This presentation is based on the extensive data documented and collected by the first author in her research on Barasana over the past ten years. We concentrate on the pitch-accent system of the language. Morphemes lexically contrast for H or HL pitch accents. One of these survives per accentual domain (generally the word) from which the pitch pattern of the entire word is predictable. After describing the basic system we survey some interesting accent patterns that arise in particular morphological constructions including compounds, possessive phrases and verbs. Processes of particular interest include an accent copy phenomenon in possessives, an accent shift in verbal constructions, and a polarity prefix. If time permits we end the presentation with certain alternations in which branching structure is key: augmentation to achieve bimoraic feet as the optimal morpheme size and accentual domains that do not exceed three feet.
For more information, contact czoll@mit.edu.
MIT Linguistics Colloquium
Friday, April 16, 1999, at 3:30, in MIT Room 56-114Peter Ludlow
SUNY, Stonybrook"LF and Natural Logic: The Syntax of Directional Entailing Environments" Abstract: During the past century, the focus of work in logic has principally been on the development of logical calculi that bear little apparent relation to the structure of natural language. That is, the syntactic forms postulated for purposes of logical reasoning (e.g. the logical forms of the propositional and predicate calculi) show little relation to the forms postulated within current grammatical theory. This apparent mismatch between logical form and grammatical form is ordinarily taken as an unsurprising consequence of the fact that, since Frege and Russell, one of the central goals of logic has been to help clarify (if not establish) the foundations of mathematics. Mathematics aside, however, it has also been held that natural language is a rather poor medium in which to couch logical reasoning, since natural language is ambiguous, vague, etc. Yet if we step back a bit and view the development of logic over the last two millennia, we find a rather different picture. Up until the beginning of this century logic was very much concerned with representing the inferences that are made in natural language. So, as everyone knows, classical Aristotelian logic attempted to characterize the valid syllogistic arguments, and the strategy was to elucidate the natural language forms that underwrite valid inferences. It is less well known is that in the 2000+ years after Aristotle a number of successful efforts were made to expand the scope of classical logic as well as to find ways of generalizing and simplifying the rules of inference to a few core cases. Let's use the phrase 'natural logic' to cover this broad research program.Among the interesting insights of the medieval logicians was the idea that the valid arguments of the syllogism could be generalized to fall within two basic inference paradigms, which they called the dictum de omni and the dictum de nullo, and much of the effort expended in natural logic concerned the characterization of the grammatical environments that corresponded to these two dicta. More recently, it has been observed by Hoeksema (1986) that the dicta de omni et nullo correspond to the upward entailing and downward entailing environments widely discussed in linguistics literature (e.g. in Ladusaw 1979). In effect, the medievals were looking for the syntactic reflexes of directional entailingness.
Can we find the holy grail of natural logic?; can there be syntactic accounts of directional entailingness? S·nchez (1991, 1995) and Dowty (1994) have offered proposals that introduce "monotonicity markings" into the Lambek calculus and categorial grammar respectively, but these efforts, or so I shall argue, have an ad hoc character. Taking an alternative tack, I will describe a formal language L* from Law and Ludlow (1985) in which directional entailingness can be defined in terms of independently motivated features of the syntax of L*. In particular, as proved in Ludlow (1995) a term in L* is in a downward entailing environment if all its occurrences are in the scope of negation and an upward entailing environment if it has no occurrences in the scope of negation. I will then argue that the LF representations in current linguistic theory can easily reflect all the relevant syntactic properties of L*, and I will show how the relevant LF representations can be derived utilizing only "off the shelf" syntactic resources (e.g. functional heads for polarity, number, and conjunction, a copy theory of movement, basic ideas about feature checking and chain formation, and unselectively bound free variables). Along the way I will have some speculative remarks about the syntactic environments which license negative polarity items, and will make a plea for austerity within semantic theory.
References
Dowty, D., 1994. "The Role of Negative Polarity and Concord Marking in Natural Language Reasoning." Proceedings of SALT IV.
Hoeksema, J. 1986. "Monotonicity Phenomena in Natural Language." Linguistic Analysis 16, 235-250.
Ladusaw, W. 1979. Polarity Sensitivity as Inherent Scope Relations. Doctoral dissertation, University of Texas at Austin.
Law, D. and P. Ludlow. 1985. "Quantification without Cardinality." In Berman, Choe, and McDonough (eds.) Proceedings of NELS XV, Amherst: GLSA.
Ludlow, P., 1995. "The Logical Form of Determiners," The Journal of Philosophical Logic 24, 47-69.
S·nchez Valencia, V., 1991. Studies on Natural Logic and Categorial Grammar. Doctoral Dissertation, University of Amsterdam.
S·nchez Valencia,V., 1994. "Monotonicity in Medieval Logic," In A. de Boer et al (eds), Language and Cognition 4, Yearbook 1994 of the Research Group for Theoretical and Experimental Linguistics at the University of Groningen.
S·nchez Valencia,V., 1995. "Natural Logic: Parsing-driven Inference." Linguistic Analysis 25, 258-285
For more information, contact asrackow@mit.edu.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, April 15, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Akihiko Uechi
The Reischauer Institute of Japanese Studies, Harvard U."What does phonology see in syntax?: A case study in Japanese" Abstract: In this talk, I show that phonological phrasing is determined not only by phonological and semantic constraints but also by its focus status (whether an element is focused or not, and if focused, contrastively focused or presentationally focused) and by the syntax, i.e., phrase structure and argument structure. In particular, I demonstrate that phonology refers to different levels of syntactic representation, which are manifested as different degrees of downstep, i.e., as the different types of phonological phrasing.The talk is divided into two parts. In the first half, I establish the syntax-prosody mapping in Japanese in terms of the revised End-based Analysis. In particular, I modify the End-based Analysis proposed by Selkirk & Tateishi (1988, 1991) by proposing two additional constraints: the Branching Constraint and the Minimization of Structures. I claim that at the start of a discourse, the domain of a Major Phonological Phrase (the domain of partial downstep) is determined by the left edge of branching XPs.
In the second half, I demonstrate that phonological phrasing is further regulated by argument structure and government relations, which determine total downstep. In order to explain the two degrees of downstep, I propose a new prosodic level--the Middle Phonological Phrase (MidP)--below the Major Phonological Phrase (MajP). By assuming that the prosodic hierarchy includes this intermediate phrase, total donwstep in Japanese and focus projection in English may be explained in a parallel manner. Additionally, so-called rhythmic effects may then be derived from strict phonological binarity.
For more information, contact karlos@mit.edu or mhcote@mit.edu.
MIT RLE Speech Communication Group Seminar
Wednesday, April 14, 1999 at 3:00 PM at MIT in the Grier Room B, MIT 34-401Joseph Perkell
Research Laboratory of Electronics, MIT"Speech motor control: Acoustic goals, saturation effects, auditory feedback and internal models" Abstract: A theoretical overview and supporting data are presented about the motor control of the segmental component of speech production. We hypothesizE that segmental speech movements are programmed to achieve sequences of phoneme-related goals that consist of regions in a multi-dimensional auditory perceptual space -- the "task space" in motor control terminology. To a first approximation, auditory parameters of speech sounds can be described by their acoustic correlates. Findings of "motor-equivalent" trading relations between the contributions of two constrictions to the same acoustic transfer function provide preliminary support for the acoustic nature of segmental articulatory goals. The goals are determined partly by non-linear, quantal relations, called "saturation effects" between motor commands and articulatory movements and between articulation and sound. Saturation effects are basic characteristics of speakers' production systems which make it possible to produce a sound output that has some relatively stable acoustic properties with a somewhat variable motor input.Since processing times would be too long to allow the use of auditory feedback for closed-loop error correction in achieving acoustic goals, the control mechanism must use a robust "internal model" of the relation between articulation and the sound output that is learned during speech acquisition. Studies of the speech of cochlear implant and bilateral acoustic neuroma patients provide evidence supporting two roles for auditory feedback in adults: maintenance of the internal model, and monitoring the acoustic environment to help assure intelligibility -- by guiding relatively rapid adjustments in "postural" parameters underlying average sound level, speaking rate and amount of prosodically-based inflection of fundamental frequency and sound level.
Reference:
Perkell, J.S., Matthies, M.L., Lane, H., Guenther, F.H., Wilhelms-Tricarico, R., Wozniak, J. and Guiod, P. (1997). Speech motor control: Acoustic goals, saturation effects, auditory feedback and internal models, "Speech Communication" 22, 227-250.
For more information, contact MARILYN@speech.mit.edu
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, April 8, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Anders Holmberg
University of Tromsø / MIT"Why aren't there any possessive suffixes in Scandinavian?" Abstract: Given fairly standard, reasonable assumptions about language change and language variation in general, and given what we know about the evolution of the noun phrase in the Scandinavian languages in particular, we expect at least some varieties of Scandinavian to have possessive suffixes, as a result of reanalysis of postnominal possessive pronouns: All the preconditions seem to have been met. But there are no such varieties. In fact, although possessive affixes are extremely common world-wide, no Indo-European language except the Iranian languages have possessive affixes (as far as I know). This implies that there is something about these languages which prevents reanalysis of possessive pronouns as affixes. The claim is that the problem is the structure of the pronouns, specifically the gender feature, characteristic of Indo-European (except Iranian). It makes it impossible for the pronouns to agree with the possessor, identifying small pro, the way a possessive affix must do.
For more information, contact karlos@mit.edu or mhcote@mit.edu.
MIT RLE Speech Communication Group Seminar
Wednesday, April 7, 1999 at 3:00 PM at MIT in the Grier Room B, MIT 34-401D. R. Ladd
Dept. of Linguistics, Edinburgh University"Segmental Anchoring of Tonal Targets: Some Consequences " Abstract: Arvaniti, Ladd and Mennen (1998) show that in Modern Greek, rising pitch accents have neither constant slope nor constant duration, but rather are quite precisely influenced by the segmental make-up of the accented word. Specifically, such rises begin at the end of the preceding unstressed syllable and end at the beginning of the following unstressed vowel. This is consistent with a theory in which the rise is analysed as a sequence of independently aligned and scaled L and H tones.In subsequent work my collaborators and I have found that this "segmental anchoring" of tones is a rather general phenomenon. In this talk, I will summarize several of our recent studies, including one on the effects of speech rate in English, one on the effects of phonological vowel length in Dutch, and one on the effects of syllable boundary location in English, as well as additional data from various other languages.
The phenomenon of segmental anchoring has clear implication for speech technology, for speech production, and for phonology. For speech technology, modelling segmentally anchored targets should increase the naturalness of synthesized intonation. In the study of speech production, segmental anchoring may lead to better understanding of production constraints on the coordination of laryngeal and supralaryngeal gestures. In phonology, if segmentally anchored targets correspond to phonological "tones", then the correct analysis of intonation systems becomes a much more straightforwardly empirical matter, and various current problems in intonational phonology have relatively clear solutions.
Reference:
Arvaniti, Amalia; Ladd, D. Robert; Mennen, Ineke (1998). Stability of Tonal Alignment: the case of Greek prenuclear accents. Journal of Phonetics 26: 3-25.
For more information, contact MARILYN@speech.mit.edu
MIT Phonology Circle
Friday, April 2, 1999, at 2:00 p.m. in room E66-156 (note room change), MIT.Liina Pylkkanen "Reduplication and Prosodic Misalignment"
For more information, contact czoll@mit.edu.
MIT Linguistics Colloquium
Friday, April 2, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Mats Rooth
University of Stuttgart"Focus licensing of anaphoric and conjunctive ellipsis" Abstract: Previous work has claimed that VP ellipsis is licensed by a focus feature (contrastive prominence morpheme). The talk will review the arguments for this, and refine the analysis: VPE has the tree geometry which also licenses second occurrence focus phonology. Then I will show that the non-local sloppy readings which motivate the focus analysis of VPE are also observed for ellipsis of the conjunctive type (e.g. gapping and comparative ellipsis). This motivates an LF for conjuctive (bare argument/remnant) ellipsis with a focus feature having scope over the remnant. Contrary to initial impressions, this representation is supported by the phonology of the remnant in comparative ellipsis.I'll consider an implementation of a two-level redundancy analysis of VPE, with reduncancy at the VP level in addition to propositional redundancy expressed by the focus. Though no conclusive argument will be presented, bringing conjuctive ellipsis into the picture tends to support two-level redundancy.
For more information, contact asrackow@mit.edu.
Boston UniversityApplied Linguistics Colloquium
Friday, April 2, 1999, at 4:00 PM in Room SED 259John (Haj) Ross
University of North Texas"The Grammar of Path" Abstract: Not available at this time.
For more information, contact Liz Hughes.
MIT LING-LUNCH
Thursday, April 1, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.
Two WCCFL practice talks:1. Idan Landau, Massachusetts Institute of Technology
"Psych-Adjectives and Semantic Selection" 2. Miriam Engelhardt, Hebrew University of Jerusalem"Properties and Arguments" Abstracts:1. Idan Landau: Psych-Adjectives and Semantic Selection
Faraci (1974) observed that psych-adjectives cannot take 'for'-complements with deleted objects, whereas non-psych adjectives require an object gap in their complement:
- (1)
- a. The patient is anxious/willing/eager for the doctor to operate on *(him).
- b. The patient is ready for the doctor to operate on (him).
- c. The tumor is ready for the doctor to operate on (*it).
Notice that 'ready' is ambiguous between a psychological reading and a "material" reading. The two readings correspond in (1b) to the options of leaving the object 'him' in or out, respectively.
Faraci's paradigm can be extended, as in (2), to infinitival complements with null subjects.
- (2)
- a. Bill1 is afraid [e1 to tell the truth].
- b. Bill1 is afraid [e1 to be told the truth].
- c. * Bill1 is afraid [e to tell the truth to e1].
- d. The book1 is available [e to read e1].
- e. * The book1 is available [e to read it1].
- f. The book1 is available [e1 to be read].
The overall state of affairs can be summarized by two generalizations:
- (A) Given an adjective A that takes an infinitival complement C:
- i. If A is psychological, C contains at most one gap - in subject position.
- ii. If A is non-psychological, C contains at least one gap (subject or object).
I argue that the two generalizations follow from independently motivated assumptions about the argument structure of the two kinds of adjectives and the semantic types to which different infinitives are mapped. Psychological adjectives denote two-place relations between individuals (experiencer) and propositions (target/subject-matter; cf. Pesetsky 1995); non-psychological adjectives denote one-place properties of individuals (theme). This is schematized in (B-i,ii).
- (B)
- i. Psych-A(<DP,CP>) DP type e; CP type
- ii. NonPsych-A(DP) DP type e
- iii. (CP(NonPsych-A))(DP) DP type e; CP type <e,<s,t>>
Next, consider clause types. An object gap can only correspond to a variable bound by a null operator (English lacking object-pro). Therefore, a clause with an object gap denotes a predicate. A subject gap can be interpreted as PRO (controlled or arbitrary) or as a bound variable too (as in 'the man to do the job'; see Clark 1990). The resulting semantic typology of clauses is schematized in (C):
- (C)
- i. [CP Op1[IP DP/PROarb...[VP ... e1] ]] CP denotes a predicate <e,<s,t>>
- ii. [CP Op1[IP e1... [VP ... ] ]] CP denotes a predicate <e,<s,t>>
- iii. [CP [IP DP/PRO...[VP ... ] ]] CP denotes a proposition <s,t>
The system straightforwardly derives the generalizations in (A): Because it selects a proposition, a psych-adjective can only combine with clauses with no gaps or just a subject gap (interpreted as PRO); this is option (C-iii). We assume that obligatory control is not simple predication (contra Williams 1980, Chierchia 1984), hence it is PRO that saturates the embedded predicate. A non-psych adjective can be modified by a (type-shifted) predicative CP, as shown in (B-iii), forming a complex predicate which then applies to the single argument DP; these are options (C-i,ii).
A clear prediction we make is that infinitivals under psych-adjectives should behave like arguments, but those under non-psych adjectives like modifiers. A series of diagnostics confirms this prediction: Infinitivals under psych-adjectives are transparent to extraction, can be extraposed after VP-fronting and can strand the preposition 'for'; those under non-psych adjectives display the opposite properties. Of particular interest is the fact that clauses of type (C-ii) pattern with those of type (C-i) and not (C-iii): This suggest that contrary to appearance, subject-gap complements to non-psych adjectives do not involve obligatory control but rather complex-predicates.
The latter conclusion is strengthened by cross-linguistic evidence. Whereas all languages appear to have subject-gap complements to psych-adjectives (2a), some languages lack subject-gap complements to non-psych-adjectives (2f). Interestingly, the presence of the latter correlates with the presence of subject-gap infinitival relatives. This strongly suggests that (C-ii) and (C-iii) exploit distinct UG resources, the former subserving null-operator constructions and the latter obligatory control constructions.
References:
Chierchia, Gennaro. 1984. Topics in the Syntax and Semantics of Infinitives and Gerunds. Doctoral dissertation, Umass, Amherst MA.
Clark, Robin. 1990. Thematic Theory in Syntax and Interpretation. Routledge, London.
Faraci, Robert. 1974. Aspects of the Grammar of Infinitives and For-Phrases. Doctoral dissertation, MIT.
Pesetsky, David. 1995. Zero Syntax. MIT Press, Cambridge MA.
Williams, Edwin. 1980. Predication. Linguistic Inquiry 11, 203-238.
2. Miriam Engelhardt, "Properties and Arguments"In this talk I challenge the claim that the NP/DP distinction in noun phrases corresponds to the predicate/argument distinction (Abney 1987, Szabolcsi 1987, 1994, Stowell 1989, 1991, Vergnaud and Zubizarreta 1992, Longobardi 1994). I argue that bare NPs are essentially argumental and that a predicative use is a derivative one.
1. Drawing on data from Modern Hebrew, I present evidence that argument positions can be occupied by noun phrases which do not project to D. The evidence is adduced from the domain of nominals which take obligatory internal arguments (=process, or event, nominals in the sense of Grimshaw (1990)).
2. It is further shown that NPs can be predicated of the internal argument, and that predication of the internal argument necessitates a structurally realized open position. This indicates that some mechanism is required to turn an NP into a predicative expression. I put forth the idea that the lexical/functional distinction in the nominal system matches the different denotations of noun phrases as referential expressions and properties, but the latter do not necessarily correlate with the different uses of nominal expressions as arguments and predicates. Bare NP arguments are analyzed as nominalized properties. The event variable present in these nominals is bound by a generic operator thus yielding a generic interpretation of the event. When D is projected, the event variable is bound by D. The binding of the event variable by D is linked to specificity. Process nominals headed by D are shown to be interpreted as specific events.
For more information, contact karlos@mit.edu or mhcote@mit.edu.
Boston UniversityApplied Linguistics Colloquium
Monday, March 29, 1999, at 4:00 PM in Room SED 259Ken Turner
University of Birmingham, England"The AntiGrice" Abstract: Not available at this time.
For more information, contact Liz Hughes.
MIT Phonology Circle
Friday, March 26, 1999 - talk postponed until later in the term!Elsa Gomez-Imbert, CNRS
Michael Kenstowicz, MIT"Barasana Accent" Abstract: Barasana is a Tucanoan language from the Amazon Basin in Columbia. This presentation is based on the extensive data documented and collected by the first author in her research on Barasana over the past ten years. We concentrate on the pitch-accent system of the language. Morphemes lexically contrast for H or HL pitch accents. One of these survives per accentual domain (generally the word) from which the pitch pattern of the entire word is predictable. After describing the basic system we survey some interesting accent patterns that arise in particular morphological constructions including compounds, possessive phrases and verbs. Processes of particular interest include an accent copy phenomenon in possessives, an accent shift in verbal constructions, and a polarity prefix. If time permits we end the presentation with certain alternations in which branching structure is key: augmentation to achieve bimoraic feet as the optimal morpheme size and accentual domains that do not exceed three feet.
For more information, contact czoll@mit.edu.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, March 25, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Akiniko Uechi
Harvard UniversityTitle TBA Abstract: Not available at this time.
For more information, contact karlos@mit.edu or mhcote@mit.edu.
Harvard University
Friday, March 19, 1999, at 4:00 PM - Room 303, Boylston HallKazuhiko Yoshida
Kyoto University"Assimilation in Hittite and Related Problems" Abstract not available at this time.
MIT Phonology Circle
Friday, March 19, 1999, at 2:00 p.m. in room E66-156 (note room change), MIT.Jay Rifkin presenting work of Paul Boersma:
a discussion of his thesis, Functional Phonology Preview: We will focus on Chapters 1, 6, and 19. Some of the architecture is introduced in Chapter 2, and the most of the constraints relevant to the discussion are introduced in Chapters 7 & 9.Chapters 8 & 12 may also be of interest, so I've included them in the hard copy as well.
Except for chapter 6, versions of these are all available on the Rutgers Optimality Archive. A hard copy for copying will also be available on the file cabinets outside the first-year office.
Here is a description of Boersma's book, taken from the blurb on the back cover:In Functional Phonology, Paul Boersma develops a theory that seeks to explain and describe the data of the languages of the world from general capabilities of human motor behaviour and perception. By separating the roles of the articulation and the audition of speech sounds, it predicts and clarifies generalizations about the organization of human speech, and solves several outstanding controversial phonological issues.
Providing a synthesis between the "phonetic" and "phonological" standpoints, the theory of functional phonology expresses explanatory functional principles like the minimization of articulatory effort and the minimization of perceptual confusion directly in a descriptive formal grammar, and offers a typologically and empirically adequate alternative to generative theories of autosegmental phonology and feature geometry. The subjects covered in this book include articulation and perception models, constraint-based accounts of phonetic implementation, the acquisition of articulatory and perceptual phonological feature values, an algorithm for learning stochastic grammars, the construction of phoneme inventories, circular optimization in sound change, and a determination of the fundamental principles that underlie the surface phenomena sometimes ascribed to the primitive phonological operations of spreading and the Obligatory Contour Principle.
This book will appeal to phonologists interested in the possibility that the grammar directly reflects common principles of efficient and effective communication, to phoneticians interested in the idea that phonetic explanations can be expressed as constraint interactions in a formal grammar, and to any linguist interested in the innateness debate.
For more information, contact czoll@mit.edu.
MIT Linguistics Colloquium
Friday, March 19, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Keren Rice
University of Toronto"Templatic Morphology? The Athapaskan Evidence Revisited" Abstract: In examining the order of morphemes in the verb of Athapaskan languages, two properties stand out. First, we observe what can be called "global uniformity" -- there is a remarkable similarity in the ordering of morphemes within the verb across languages of the family. Second, we observe what can be called "local variability" -- the order of morphemes in the verb differs in some ways across the languages of the family. I argue that the global uniformity exhibited within the family is a consequence of general principles of grammar rather than being accidental. In particular, I argue that there is an overarching principle that determines the ordering of morphemes within the verb of Athapaskan languages, a principle of scope, or semantic compositionality, which requires that morphemes of greater scope occur in a fixed position with respect to morphemes within their scope. This principle accounts for the global uniformity found within the verb of the languages of the Athapaskan family: given a universal scope relationship, the morphemes could not be ordered in any other way. It also accounts for a certain amount of local variability: given two morphemes A and B, either A may occur in the scope of B, or, on a different reading, B may occur in the scope of A, leading to different orderings. There are times when this principle is irrelevant -- morphemes may have no scopal relationship to each other. In this case, variability exists across the family and, at times, within a single languages.Traditionally, the Athapaskan verb is argued to exhibit template morphology, with each lexical entry having associated with it a marking for its position. Word formation follows a standard model of derivation followed by inflection (see, for instance, work by Kari), and the morphemes are then ordered by a template, opacifying the order in which they were added to the verb. Under this account, the global uniformity becomes accidental, finding an account simply in common origin.
In this talk, I focus on arguments as to why the similarities and differences in the verb should be attributed to universal principles that function to regulate both synchronic morpheme order and language change rather than simply to common origin.
Athapaskan languages differ from each other in a myriad of ways -- the languages show distinctions from each other in phonological, morphosyntactic, and syntactic areas. Thus, it is not the case that the languages are impervious to change. In the absence of principles to explain why certain changes are allowed and not others, the fact that so much global unfiormity exists within the verb is surprising -- one might expect random changes in morpheme order across the family rather than the systematic differences that actually are found.
The lexicon also provides evidence for the scopal hypothesis. Under a hypothesis where similarities are due simply to common origins and no principles explain the differences, few predictions are made concerning morpheme placement of new morphemes, of morphemes that change in meaning, and so on. Examining languages across the family, we find the following types of situations.
First, in cases where a particular semantic form has more than one phonological realization across the family, the semantics of the item determines its placement. For instance, an aspectual preverb meaning undirected movement has several different forms across the family, yet its position is identical in all languages. Similarly, the D-quantifier (variously a distributive or a collective) occupies a similar location despite variation in its phonological realization. Languages differ in whether they differentiate first person plural and second person plural object pronouns phonologically, but, no matter what, the object pronouns have a shared location. In all of these cases, the common origins hypothesis is silent as to why these changes have the outcome they do, while the scopal account offers an explanation.
Second, in some cases, new morphemes have been added in a particular language. These do not necessarily locate themselves at word edge, but are positioned where one might expect given their semantics. For instance, Hupa has innovated a viewpoint/situation type, customary, with overt marking by a functional morpheme; this morpheme is located in the same position as other aspect-marking functional items.
Third, morphemes have shifted function in some languages. For instance, in Hupa and Kato, there is a third person pronoun that marks discourse topics that is cognate to an unspecified human subject in other languages. The Hupa/Kato form is a lexical item, whereas cognate is a functional item; these differ in positions. Again, the common origins hypothesis does not predict a change in position co-occurring with a change in function while the scopal hypothesis does.
Finally, in some cases morphemes have developed opacity in their meaning. These particular morphemes are more mobile than those with fixed meanings. For example, in Navajo, there is an aspectual morpheme that marks inception. It is extremely productive, and is basically invariant in position. There is a second morpheme that historically marked termination, but it is not used productively, and its function as a marker of termination is now quite opaque; this latter morpheme is far more variable in its position. The scopal acocunt predicts that this could be the case; the common origins hypothesis has nothing to say about the difference between these morpheme types.
I conclude that the ordering of morphemes within the verb of Athapaskan languages is, largely, a consequence of universal scopal principles. These principles control the ordering of interacting items, and regulate change in that items within a fixed scopal relationship are not subject to changes in ordering.
For more information, contact asrackow@mit.edu.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, March 18, 1999 from 12:10 to 1:10 pm in E39-335 (conference room) at MIT.Maya Arad "On the nature of "v": the case of Object Experiencer verbs" Abstract: In this talk I address two issues regarding the "transitivity head" v (Chomsky 1998):
1. Is there only one type of v?
2. What does it mean for v to be a "transitivity head"? Is there a relation between its two functions ^V introducing an external argument and case-marking of the object?To get some insight about these questions, I will look at a class of verbs which alternates between an agentive and a stative reading, namely, Object Experiencer verbs (frighten, annoy, etc.). I will show that verbs like frighten can take two readings, stative and agentive, and that these two readings differ in their syntactic behavior: on the agentive reading the verb behaves like a standard transitive verb, while on the stative reading it exhibits syntactic peculiarities (cf. Belletti and Rizzi 1988).
I argue that the two readings are formed from the same root (cf. Marantz 1997), which is combined with different verbal heads: on the agentive reading it is combined with (standard) little v, while on the stative reading it is combined with a transitive stative head, which introduces a stative causer and assigns dative case. Thus, "little v" is only one type of possible verbal head. I assume that the syntax manipulates roots, and that category features are assigned in the syntax (Marantz 1997, Chomsky 1998). Given that the same root can form different verb types (e.g. unaccusative, transitive and stative verbs as in Semitic and Romance), there have to be different types of verbal heads with which it combines on each reading (cf. Harley 1995).
Next, I discuss the "transitivity" of v. I suggest that a "transitivity head" mediates between two different properties: the semantic role of the subject and the case assigned to the object. Using data from ObjExp verbs, I will show that both (active) v and "stative little v" are transitive: the case and theta properties of these heads are interdependent, and when one is affected, so is the other.
For more information, contact karlos@mit.edu or mhcote@mit.edu.
Boston UniversityApplied Linguistics Program and Department of Psychology Colloquium
Wednesday, March 17, 1999, noon, Psychology Seminar Room (room 150), 64 Cummington St.Elizabeth Bates
University of California, San Diego"Structural priming: Cross-linguistic studies of lexical access in context" Abstract: Not available at this time.
For directions or more information, call 353-2583/2580.
MIT LING-LUNCH
Ling-Lunch is a series of weekly talks, open to all linguistics topics. It is held in an informal setting, and everybody is welcome to present their work, but preference is given to members of the MIT Linguistics Department. Thursday, March 11, 1999 from at 12 noon in E39-335 (conference room) at MIT.Uli Sauerland
Kanda University"A Uniform Approach to Determiner Quantification" Abstract: Current theories of DP quantification posit a dichotomy between interrogative quantifiers, which are allowed to quantify over choice or Skolem functions (e.g. Engdahl 1980, 1986, Chierchia 1992), and non-interrogative quantifiers, which are restricted to quantification over individuals. Evidence from the distribution of Condition C effects (Chomsky 1993, Fox 1994) and identity of trace effects in Antecedent Contained Deletion (Sauerland 1998) shows that the dichotomy is not only conceptually, but also empirically problematic. This talk develops a uniform semantics for DP quantification, in which all quantifiers uniformly quantify over choice functions. Von Stechow (1996)'s earlier proposal along these lines is shown to make wrong predictions for cardinal quantifiers, as well as cases involving a bound variable. As a refinement of von Stechow's proposal, I propose to restrict quantification over pointwise different choice functions.
For more information, contact karlos@mit.edu or mhcote@mit.edu.
MIT Linguistics Colloquium
Friday, March 5, 1999, from 3:30 to 5:30 p.m. in room 56-114, MIT.Chris Wilder
Zentrum fur Allegmeine Sprachwissenschaft"Shared constituents and the LCA" Abstract: I. Shared constituents like 'the book' in (1) have been analyzed as: (2) ATB movement out of the coordination, i.e. Right Node Raising (Postal 74); (3) ellipsis, i.e. phonological deletion = (...) of one constituent under identity with another (Backward Deletion rule of Wilder 97); (4) multiple dominance, where OB* is a single constituent with two mothers, one per conjunct (Goodall 87, Muadz 91, Moltmann 92).
- 1)
- John bought and Mary read the book
- 2)
- [ [John bought tj ] and [Mary read tj ] OBj ]
- 3)
- [John bought (OB1) ] and [Mary read OB2 ]
- 4)
- [John bought OB* ] and [Mary read OB* ]
Assuming (2) is wrong--rather, the shared constituent is inside the last conjunct (Wilder 97)--the choice reduces to whether there are two constituents, one deleted (3); or only a single constituent (4), in which case the Single Mother Condition must be given up. I shall argue for (4) over (3).
II. The ellipsis approach must stipulate (5), illustrated by (6) vs (7)--unlike the DO in (6), the IO in (7) cannot stand in VP-final position (Oehrle 91):
- 5)
- If a shared constituent X surfaces in the final conjunct, then gaps corresponding to X in nonfinal conjuncts must be at the right edge of their respective conjuncts.
- 6)
- I [invited into my house _ ] and [congratulated [the boy who lives nextdoor]]
- 7)
- * I [gave _ a present] and [congratulated [the boy who lives nextdoor]]
Where X surfaces in or to the left of the initial conjunct --e.g. X has been ATB-moved leftwards, or X is the shared V in Gapping (8)-- X-gaps in non-initial conjuncts underly no such edge restriction.
- 8)
- John likes beer and [Mary _ wine]
III. Johnson (97) proposes that gapping is ATB movement of V out of &P, accompanied by CSC-violating subject-raising out of the initial conjunct (9).
- 9)
- Johnk likesj [&P [ tk tj beer ] & [ Mary tj wine ]]
More generally, I suggest that in all cases where shared X appears to surface inside the initial conjunct, X has in fact ATB-moved out of &P. Then the only case where shared X surfaces inside &P is RNR--X surfaces inside the final conjunct, with X-gaps in non-final conjuncts restricted to the right edge, as per (5).
IV. A question arising from (4) is: what determines that OB* surfaces in the final and not the initial conjunct? I will show that Kayne's LCA, suitably modified to cope with multiple dominance trees, explains both this property and the restriction (5).
V. Multiple dominance relaxes the Single Mother Condition, permitting e.g. (4), (10).
- 10)
- [ [A ... X* ... ] [B ... X* ... ] ]
X is 'shared' by A and B iff (i) neither of A and B dominates the other, (ii) both A and B dominate X. Under multiple dominance, plus assumptions (11), (5) can be derived from a modified LCA. The deletion rule needed for (3) and the associated stipulation (5) can be dispensed with.
- 11)
- a. A c-commands X contained in B (=A's sister) even if A also dominates X.
- b. Trees encode only dominance, with precedence defined only for terminals (traces invisible) by the LCA (Chomsky 1995).
- c. Coordination is asymmetric: the first conjunct asymmetrically c-commands the second, etc.
VI. LCA (Kayne 94): The 'image' of a category X, d(X), is the unordered set of terminals that X dominates. The image of an ordered pair of categories <X,Y> is the set of ordered pairs of terminals that is the Cartesian product of <d(X), d(Y)>. For a tree T, the LCA takes the image of the set CC of all pairs of categories <X,Y> in T s.t. X asymmetrically c-commands Y. The output is wellformed only if d(CC) = a linear (asymmetric, irreflexive, transitive, total) ordering of terminals of T. Under these assumptions, the LCA cannot linearize shared constituents: given sisters A, B sharing X as in (10), where A asymmetrically c-commands (into) B, the terminals of X, included in both d(A) and d(B), inevitably precede themselves when the terminals of A and B are ordered (x<x violates the irreflexivity requirement).
VII. Modified LCA: Assume that terminals of constituents shared by A and B are not included in the image of A or of B. Reflexive orderings of terminals of X shared by A and B no longer arise. In (12), X* is ordered by c-command within A (w<X*<x) and within B (y<X*<z). C-command by A into B orders only non-shared terminals of A (w,x) w.r.t daughters of B (including X*, which A c-commands): {w,x}<{y,X*,z}, but not X*<X*. X* is thus placed among the terminals of B (='shared constituent surfaces in the final conjunct').
- 12)
- [ A B ] = [ [ w X* x ][ y X* z ] ]
(12) still cannot be linearized, since X* also c-commands x within A. The output, though irreflexive, is neither asymmetric (X*<x, x<X*) nor transitive (y<X*, X*<x but not y<x). But if there is no x within A s.t. X* asymmetrically c-commands x, the symmetry/transitivity violations disappear, and a linear order results, e.g. (1) for (4), with X*=OB*. This derives (5). The LCA orders X* within the final conjunct, leaving a 'gap' in non-final conjuncts where X* 'ought to be'. If X* 'ought to' precede any other terminal in a non-final conjunct, the structure is out (not linearizable).
References
Chomsky, N. 1995. The Minimalist Program. MIT Press.
Goodall, G. 1987. Parallel Structures in Syntax. CUP.
Johnson, K. 1997 Gapping. Ms., UMass.
Kayne, R. 1994. The Antisymmetry of Syntax. MIT Press.
Moltmann, F. 1992. Coordination and Comparatives. PhD, MIT.
Muadz, H. 1991. Coordinate Structures: a Planar Representation. PhD. Arizona.
Oehrle, R. 1991. Categorial Frameworks, Coordination and Extraction. WCCFL 9.
Postal, P. 1974. On Raising. MIT Press.
Wilder, C. 1997. Some Properties of Ellipsis in Coordination. In A. Alexiadou/T. Hall eds. Studies in Universal Grammar and Typological Variation. John Benjamins. .
For more information, contact asrackow@mit.edu.
MIT Lecture Series
Thursday, March 4, 1999, at noon, at MIT in 56-154Rose Marie Dechaine "Agreement and Event Structure" Abstract: not available at this time. For more information, contact Michael Kenstowicz.
Boston UniversityLinguistics
Lecture
NEW TIME: Monday, March 1, 1999, 9:30-10:30, CAS 533 (the video studio near Geddes Language Center)Judy Bernstein
Syracuse University"(Un)natural Grammars and Economy in Language Acquisition" Abstract: I argue based on child language data that basic preposition pied-piping in English is an unnatural prescriptive artifact. In a study of 115 children and 20 adults, which included an elicited production task and a grammaticality judgment task, we investigated the status of stranding and pied-piping alternatives in preposition relative clauses. Our general findings were: a) children avoided and rejected preposition pied-piping; b) adults also avoided preposition pied-piping, although they accepted it in their judgments. I demonstrate how the child data on preposition pied-piping are expected under an economy (Chomsky 1995) account, where only one derivation from a given set of lexical items is predicted to be grammatical. The adults' acceptance of preposition pied-piping is a result of a prescriptive rule learned through direct instruction.
For more information, contact carol@bu.edu.
MIT Lecture Series
Monday, March 1, 1999, at noon, at MIT in E25-117Lisa Matthewson
MIT"Inside Quantificational Noun Phrases" Abstract: not available at this time. For more information, contact Michael Kenstowicz.
Boston UniversityLinguistics
Lecture
NEW TIME: Friday, February 26, 1999, 3:00-4:00 PM, CAS 326Dawn MacLaughlin
Boston University"Universal Grammar and adult second language acquisition:
Evidence from the acquisition of reflexives" Abstract: Not available at this time.
For more information, contact carol@bu.edu.
MIT Lecture Series
Thursday, February 25, 1999, at noon, at MIT in 56-154Matthew Pearson
UCLA"X(P)-Movement and Word Order Typology: Direct and Inverse Languages" Abstract: not available at this time. For more information, contact Michael Kenstowicz.
Communication & Culture Seminar (formerly MIDAS)
Wednesday, February 23, 1999, at 7:30 PM - Center for Literary and Cultural Studies at Harvard University: Barker Center, 12 Quincy Street, Cambridge, Massachusetts 02138.Bruce Fraser
Boston University""Discourse Markers and Types of Coherence" Abstract: not available at this time. 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 MIDAS Seminar at the CLCS. If there are no spaces available the guard will direct you to another Harvard parking facility.
MIT Lecture Series
Monday, February 22, 1999, at noon, at MIT in E25-117Rajesh Bhatt Title TBA Abstract: not available at this time. For more information, contact Michael Kenstowicz.
Harvard University
Wednesday, February 17, 1999, at 4:30 PM - Warren House, Room 201 (Kates Room). Reception to follow at Warren House.Eugeniusz Cyran, Catholic University of Lublin, Poland "Understanding the Phonemic Systems of Celtic Languages" Abstract: not available at this time.
MIT Phonology Lecture Series
Tuesday, February 16, 1999, at noon, at MIT in E25-117Joe Pater, University of Alberta "Developmental Speech Perception and the Phonological Lexicon" Abstract: not available at this time. For more information, contact Michael Kenstowicz.
Harvard University
Friday, February 12, 1999, at 4:00 PM - Boylston Auditorium. Reception to follow in Departmental lounge.Arild Hestvik, University of Bergen, Norway
and William Philip, Harvard University"Long-Distance Reflexives in Norwegian" Abstract not available at this time.
MIT Phonology Lecture Series
Thursday, February 11, 1999, at noon, at MIT in 56-154Ayako Tsuchida, Rutgers University "The Phonetics-Phonology Interface: Laryngeal features and Glottal Opening" Abstract not available at this time. For more information, contact Michael Kenstowicz.
MIT Phonology Lecture Series
Monday, February 8, 1999, at noon, at MIT in E25-117Eric Raimy, University of Delaware "Representing reduplication" Abstract not available at this time. For more information, contact Michael Kenstowicz.
Workshop on Japanese Prosody at UMass
Friday, January 22, 1999 at University of Massachusetts, Amherst
10:00-11:00Hisao Tokizaki (UMass/Sapporo Univ.)
"Prosody and information in Japanese and English" 11:00-12:00Masako Hirotani
"Deaccenting on verbs?" Lunch 13:00-14:00Jennifer Smith
"Accent and wh-questions in Fukuoka dialects" 14:00-15:00Mariko Sugahara
"Downstep and the formation of prosodic phrases in Japanese" Break 15:30-16:30Akihiko Uechi (Harvard University)
"Prosody-syntax/focus interface in Japanese"
The workshop will be held in a room at Campus Center. We would appreciate it if you let us know about your interest in attending the workshop.
Best wishes,
Hisao Tokizaki, toki@linguist.umass.edu
MIT Phonology Circle
Thursday, January 21, 1999, at 4:00 PM at MIT in E39-335Jay Rifkin
Massachusetts Institute of Technology"Ternarity is Prosodic Word Binarity" Abstract: Not available at this time.
For more information, contact Michael Kenstowicz.