Main session
Thursday, 6 April 2006
Invited Speaker: Douglas Pulleyblank (University of British Columbia) - 5:00-6:00
External possessor constructions in Sidaama
Kazuhiro Kawachi (University at Buffalo, SUNY) - 9:00-9:40
This study shows that the notion of part-whole relationship, rather than that of inalienability, is necessary for describing the possessor-possessee relationships involving the two types of external possessor constructions in Sidaama, a Cushitic language of Ethiopia. One type requires the possessee (oblique) and the possessor to be in a part-whole relationship. The other type can basically be used regardless of the type of possessive relationship, but shows a semantic difference, depending on whether or not the possessee is part of the possessor (dative); when the possessee is not part of the possessor, it expresses the possessor’s adversity or beneficial experience.
The development of syntax-pragmatics connections in early child grammars: Evidence from Greek
Konstantia Kapetangianni (University of Michigan) - 9:45-10:25
This paper focuses on the development of aspects of Greek syntax that interface with pragmatic knowledge (A-bar elements: subjects, focused/topicalized constituents and wh-questions). Unlike Grinstead (2004) who argues for a grammar-discourse interface delay in the early stages of Spanish and Catalan, this study shows that the topic-focus field in child Greek is available before the age of two. In addition, providing evidence from wh-questions in English, it will be argued that the focus field is active in early English, thus a universal interface delay between syntax and discourse does not seem to hold. An alternative analysis based on interpretability of features at LF (Tsimpli 2005) will be considered and the theoretical and empirical benefits will be discussed.
Scalar Structure and the Semantic Representations of Gradable Adjectives in Child Language
Kristen Syrett (Northwestern University), Jeffrey Lidz (University of Maryland), and Christopher Kennedy (University of Chicago) - 10:30-11:10
Preschoolers distinguish between three subclasses of gradable adjectives based on scalar structure, allowing a contextually-determined standard for relative, but not absolute, GAs. Where their responses diverge from adults is due to accommodation of presupposition failure. Asked for the full one when given two non-full containers, adults reject the request, while children are willing to suspend their own evaluation of the standard to accommodate the request, paying a price in their RTs. In loosening the criterion for the endpoint of maximum, but minimum, GAs, children highlight an inherent semantic difference in the potential for vagueness around two types of scalar endpoints.
Stress and Islands in North Bizkaian Basque
Karlos Arregi (University of Illinois at Urbana-Champaign) - 1:30-2:10
Phonological Implications of Grammatical Markedness
Cristian Iscrulescu (University of Southern California) - 2:15-2:55
The present paper argues that there is a correlation between the ranking of a category on a grammatical markedness hierarchy (defined along the lines of Greenberg 1966) and its phonological properties in that under similar phonological conditions the marked grammatical category can sponsor marked phonological material to an extent that is equal or greater than in the unmarked category. Illustrations are provided from Old Saxon, Romanian and Mayak (Western Nilotic). I argue that the correlations observed are grounded in economy of language use in that restricting marked phonological material to forms that occur less frequently in discourse represents an advantage for the speaker-hearer in that respect.
Evidence for rhythmic shortening in American English as conditioned by prosodic phrase structure
Heejin Kim and Jennifer Cole (University of Illinois at Urbana-Champaign)
Studies of English rhythm find little evidence for the isochrony of interstress intervals, but demonstrate a tendency toward isochrony in shortening that occurs in the inter-stress interval (ISI, or “foot”). It is not known how this rhythmic shortening relates to the larger prosodic structure of an utterance. We investigate the prosodic domains of rhythmic shortening in data from the Boston Radio News corpus, and show that the stressed syllable shortens with increased syllable count in the ISI. This effect is found only for ISIs that are contained within the intermediate phrase (ip). Thus, the stress foot within the ip is a timing unit for English.
Cluster Buster: A Typology of Onset Clusters
Rina Kreitman (Cornell University)
There are four logical possibilities for O(bstruents) and S(onorants) to combine in bi-consonantal onset position with the following implicational relations: SO ⇒ SS ⇒ OO ⇒ OS. While SO is expected to be the more marked and OS the less marked, the relationship between OO and SS is less transparent. The greater markedness of SS with respect to OO clusters is less obvious. One possible motivation for this is the proportionally smaller number of sonorants (30%), which limits sonorant clustering possibilities. Another possible motivation is that sonorants are not sufficiently acoustically distinct and their sequencing is, therefore, perceptually disadvantageous.
Friday, 7 April 2006
Invited Speaker: Marianne Mithun (University of California at Santa Barbara) - 5:00-6:00
Markedness Hierarchy in the Vocabulary in Distributed Morphology
Alexandra Galani (University of York) - 9:00-9:40
In this paper, I revisit the way vocabularies entries are organised in the repository of
Distributed Morphology, the Vocabulary (Halle and Marantz 1993). Evidence is brought
forward from the verbal morphological system in Modern Greek. It is proposed that
Vocabulary Items are no longer organised in blocks in terms of specificity. Instead, they
are organised in hierarchical tree structures based on the marked or unmarked value of
theme vowels, the markers of the conjugational classes. It is explicitly shown how this
alternative view to the organisation of the Vocabulary in Distributed Morphology
accounts for verb classification in Modern Greek.
Grammaticalization and Korean Datives
Chongwon Park and Sook-kyung Lee (University of Minnesota Duluth/University of Illinois at Urbana-Champaign) - 9:45-10:25
In our presentation, we report and claim that Korean Dative markers are formed through grammaticalization by examining a good deal of data from various sources (15th - 20th century documents containing 16.5 million syllables) extracted from Sejong Corpora. The four Korean Dative markers (-eykey, -hantey, -poko, and -tele) were first attested as phrases in 15th century documents. The phrases then began to undergo grammaticalization to play a role as function affixes. Although the phrases originally were all contentful, the meanings o f the phrases were obliterated over time to become full-fledged function affixes. After undergoing grammaticalization, the newly acquired affixes simply follow Korean morphotactics by affixing themselves to nominal stems. Our treatment of the Dative markers’ grammaticalization is based on the assumption that grammaticalization is a subset of linguistic changes.
Temporal reference in Inuktitut: a feature-geometric approach
Midori Hayashi (University of Toronto) - 10:30-11:10
The purpose of this paper is to offer a feature-geometric account for tense in Inuktitut within the framework of
Distributed Morphology. I will develop an analysis based on Cowper (2003, 2005) to demonstrate that tense in Inuktitut can be explained using the same set of features that accounts for the tense in English and Spanish with an additional feature [Local]. The analysis has a universal implication that features available to a language are a subset of a small number of universal features and that languages can vary in terms of kinds of features available and their possible combinations.
Past-Time Denoting Adverbs and the African-American English Preverbal done Construction: A Case of Variable Judgments
Michael Terry (University of North Carolina) - 1:30-2:10
This talk focuses on what I take to be the unclear grammatical status of African-
American English (AAE) preverbal done sentences when, as in (1), they are modified by
definite past-time denoting adverbs like yesterday.
(1) ?John done baked a cake yesterday. AAE
Speaker judgments are sharpened in two ways: First, preposing the adverb, as in (2),
generates clearly unacceptable sentences.
(2) *Yesterday, Mary done baked a cake. AAE
Second, leaving the adverb in place while providing a carefully constructed context
produces clearly acceptable sentences. These new data are used to develop a synchronic
analysis of the variable judgments of the out-of-the-blue cases.
Tongzhi, Ideologies, and Semantic Change
Andrew Wong (University of Hawaii) - 2:15-2:55
This study investigates the role of ideology in semantic change. It focuses on the semantic change of tongzhi from 'comrade' to 'sexual minorities' in Hong Kong. A common address term among Chinese revolutionaries, tongzhi was appropriated by gay rights activists in the late 1980s to refer to 'sexual minorities.' This study illuminates the actuation and the transmission of the semantic change by examining the labeling practices of gay rights activists and of gay and lesbian Hongkongers who are not involved in the gay rights movement.
The unexpected survival of German discourse markers in Texas German
Hans C. Boas and Hunter Weilbacher (UT Austin) - 3:00-3:40
This paper discusses the distribution of German discourse markers 'ja' (marker of emphasis) and 'mal' (just) in Texas German. Contrary to Salmons' (1990) and Fuller's (2001) claims that these discourse markers have been virtually lost in German American dialects, our large electronic corpus of spoken Texas German demonstrates that these discourse markers still occur frequently. Our results are compared with the distribution of bilingual TX German discourse markers such as 'you know' and 'weisst du' (you know) in order to show that Matras' (1998) pragmatic detachability hierarchy is not universally applicable.
On the role of generalized conversational implicature in semantic change
Maj-Britt Mosegaard Hansen (Copenhagen) & Richard Waltereit (Newcastle-upon-Tyne) - 4:00-4:40
Levinson (1995, 2000) and Traugott & Dasher (2002) suggest that semantic change proceeds from particularized conversational implicature via generalized conversational implicature to coded meaning. However, this model is ultimately neither theoretically nor empirically tenable.
Our alternative proposal builds on the assumption that PCI are in the communicative foreground of a message while GCI are in its background. The following sequences therefore seem to be possible:
1. A PCI semanticizes directly:
PCI (* → GCI) → coded meaning.
2. A PCI turns into a GCI, but is not fully semanticized:
PCI → GCI (*→ coded meaning).
3. A GCI semanticizes, but only after being foregrounded as a PCI:
GCI → PCI → coded meaning.
Saturday, 8 April 2006
Invited Speaker: John Ohala (University of California at Berkeley) - 5:00-6:00
Degree Phrase Structure
Scott Fults (University of Maryland) - 9:00-9:40
We provide evidence that the degree phrase projection of gradable adjectives is richer than previously assumed. Gradable adjectives can appear in positive form with the standard of comparison expressed by for-a-NP (John is tall for a wrestler) or compared-to-DP (John is tall compared to Bill). We argue that both phrases behave like syntactic arguments; but they are introduced by different degree projections of the adjective; and, they are structurally ordered
with respect to an argument PP (as in patient with Mary) such that the compared-to-DP phrase is higher than the PP and the for-a-NP phrase is lower than the PP.
Why four kinds of V (DP) PP verbs? A non-lexicalist approach
Eva Juarros-Daussà (SUNY Buffalo) - 9:45-10:25
My analysis is exemplified by the locative alternation and framed within the theory of Hale and Keyser (2002). (Examples from Spanish.) H&K distinguish between object-oriented/proximate and subject-oriented/obviative indexes, which condition the possible combinations of the lexical categories building the pertinent argument structures. My innovation is to allow for these two factors to be independent of each other, producing four, instead of two, possible structures (a typologically correct prediction). My expansion of H&K also predicts the existence of two other alternations that were extremely problematic for their theory, mainly, the unergative-unaccusative alternation in Italian, and the unergative-transitive alternation in English.
Multiple Dominance CAN’T, but PF-deletion CAN account for Right Node Raising
Seungwan Ha (Boston University) - 10:30-11:10
This paper concerns syntactic structures and licensing conditions of Right Node Raising. First, I will argue for a PF-deletion analysis (Wexler & Culicover 1980, Abels 2004, Harmann 2000), according to which the RNR elements are deleted at PF, but remain at LF, and argues against current analyses assuming Multiple Dominance (Wilder 1999), according to which the RNR element is dominated by both the VP in the first conjunct and the VP in the second. Second, I will propose semantic licensing conditions for RNR, relying on Schwarzschild’s (1999) and Merchant’s (2001) e-givenness, and discuss similarities and differences between ellipsis and RNR.
Comp-trace Effects and the Syntax-Phonology Interface
Jason Kandybowicz (UCLA) - 1:30-2:10
The Comp-trace effect is standardly understood as describing the outcome of movement operations that displace subjects across overtly headed clause boundaries. In this talk, we develop a theory of Comp-trace effects that reduces the effect to a failure of convergence at PF rather than to an illicit output of the narrow syntax. Drawing on data from English and Nupe, we argue that Comp-trace effects derive from the existence of a PF Bare Output Condition that is violated whenever a fully propositional phase fails to spell-out a discourse prominent element, essentially, a subject, topic, or focal element.
Stress is independent of vowel reduction and accent
Marta Ortega-Llebaria and Pilar Prieto (University of Texas at Austin) - 2:15-2:55
Recent investigations have lead to two main views about the phonetic properties that
characterize stress in stress-accent languages: (1) duration and spectral balance are cues
to stress, and (2) there are no ‘direct’ phonetic properties associated with stress at lower
levels of the hierarchy since these properties are “parasitic” on vowel reduction
differences. In order to disentangle this issue, we recorded 10 native speakers of Catalan
(a language with vowel reduction in unstressed positions) and 10 native speakers of
Castilian Spanish (a language with no vowel reduction). Our data offers clear support for
the idea that stress at lower levels of the hierarchy has direct duration and spectral
balance correlates in the absence of vowel reduction and thus that stress is something
more than a structural marking for a syllable to bear a pitch accent.
Acoustic Effects of Prosodic Boundary on Vowels in American English
Eun-Kyung Lee and Jennifer Cole (University of Illinois at Urbana-Champaign) - 3:00-3:40
This study investigates the acoustic effect of prosodic boundaries on vowel quality in
American English, based on the Boston Radio News corpus. We examine the F1, F2 and
duration of 2 front high vowels [i, I] and 2 diphthongs [eI, oU] in the final rhyme of the
Prosodic word, intermediate phrase (ip), and the intonational phrase (IP), and find reliable effects
of strengthening (via enhancement of V-place contrasts) and durational lengthening mostly for
vowels at the IP boundary relative to vowels at the Wd boundary. We also find that these two
mechanisms are used independently for prominence-marking in speech production.
Statistical Regularities in the Distribution of Consonants and Vowels in Korean Lexicon – Their Implications for the Internal Structure of Korean Syllables
Yongeun Lee (Northwestern University) - 4:00-4:40
Behavioral results from syllable experiments with Korean speakers indicate that, given a C1VC2 syllable, users of the language show a strong tendency to group C1 and V into a unit, excluding C2. This pattern was interpreted as a direct function of two primitive sub-syllabic constituents, ‘body and coda’, for Korean syllables (Yoon & Derwing 1994; 2001). The current study, however, demonstrates that this seemingly atypical behavior of Korean users is not due to the ‘body-coda’ constituents per se, rather that it shows that Korean speakers are implicitly aware that the two-way dependencies are, on average, much greater for C1V than VC2 sequences in Korean.
The Biasing Effect of NPIs in Questions
Brian Reese (University of Texas at Austin) - 1:30-2:10
The biasing effect of NPIs in questions has long been noted. Questions like "Does Fred lift a finger around the house?" favor a negative answer, hence are said to be negatively biased. Previous research tends to derive bias from the semantics of the NPI itself, but an NPI is neither necessary nor sufficient to produce negative bias. The present paper argues that the intonational features of biased questions are crucial in explaining bias and couches an analysis in a framework for computing the rhetorical effect of utterances in discourse and dialogue.
Plural Events in Yurok
Tess Wood (University of California at Berkeley) - 2:15-2:55
This paper analyses plural-event semantics in the Yurok language of northwestern California. Yurok has two grammatical markers of event-plurality, which indicate plurality of phases within an event and plurality of events, respectively (Garrett 2001). These two categories have systematic differences in behavior and interact differently with verb arguments, as well as with Aktionsart and other features of verbal semantics. I analyse the semantic contrast between the categories as a distinction between grouped and ungrouped pluralities (cf. Landman 2000) and show how some significant characteristics of Yurok pluractionals arise naturally from this account.
The Representation of Object Sharing in Ditransitives
Youngmi Jeong (University of Maryland, College Park) - 3:00-3:40
In this talk I reconsider the nature of double object/low applicative constructions, andargue that a complete characterization of the thematic properties of the inner/lower object requires movement connecting thematic positions. The solution I end up developing in terms of object sharing extends to other constructions such as serial verbs and resultatives. Consequences for case-licensing in multiple object constructions are also explored.
Intervention Effects in Macedonian Wh-Questions
Slavica Kochovska (Rutgers University) - 4:00-4:40
This paper argues that D-linked wh-phrases (which N) and non-D-linked wh-phrases (what) in Macedonian, a multiple fronting language, move to different positions at the left periphery of the clause: SpecTopP and SpecFocP, respectively (following Rizzi 1997). Evidence for the analysis comes from subject and adverb intervention effects in wh-questions. It is argued that lexical subjects act as defective interveners for the purposes of agreement causing blocking effects in questions with non-D-linked wh-phrases, but not D-linked ones. It is argued that adverb intervention effects in questions with D-linked wh-phrases are due to the topicalization of both the fronted wh-phrases and the adverbs. Adverbs cannot intervene between two fronted non-D-linked wh-phrases because topics can only precede foci in Macedonian.
Questions or comments? Contact us here:
