Linguistics

The frequency code (Ohala, 1994) posits that hearers associate higher frequencies of F0 and F2 with small size and lower frequencies with large size. There is some evidence that sound symbolism is at work in variation; Eckert (2010), e.g., found that high F2 is linked with positive affect and low F2 with negative affect in preadolescent girls’ speech. This paper compares two kinds of Baby Talk in order to probe the robust potential for phonetic symbolism in adult speech.
Twelve 20-30 year olds talked to a baby pig plush toy in American English for two tasks. In Task 1 (Cute Baby Task), they were instructed to tell the pig how cuddly and wonderful she is. In Task 2 (Poor Baby Task), they were instructed to comfort the pig because she got a boo-boo. Participants were then interviewed to assess their baseline F0s and vowel spaces. In total, 360 minutes of speech were analyzed.
Cute Baby Talk and Poor Baby Talk differ significantly (from each other) in vowel placement, mean F0, and Intonational Phrase (IP) contours. The Baby Talk samples also diverge significantly from the interview speech.
Cute Baby Talk realizes high frequencies of F0 and F2 while Poor Baby Talk realizes low frequencies of F0 and F2. Poor Baby Talk also realizes significant IP-initial and IP-final lengthening. This empathetic lengthening mirrors the hearer’s imagined pain; the participants’ long, backed /o/s, /a/s, and /u/s create a full, wailing sound.
These findings show how the frequency code operates in Baby Talk, introduce a new social meaning of IP-initial and -final lengthening, and hint at how child-directed and childlike speech features may be used in adult-adult conversation. Inter-adult uses of Baby Talk features—e.g., in dialogue between intimates—are likely to carry some of the meanings uncovered here.

References
Eckert, P. (2010). Affect, Sound Symbolism, and Variation. University of Pennsylvania Working Papers in Linguistics, 15:2, Article 9, 68-80.
Ohala, J. (1994). The biological bases of sound symbolism. In L. Hinton, J. Nichols & J. J. Ohala (Eds.), Sound Symbolism. Cambridge: Cambridge University Press, 222-36.

Traditionally, within-turn pauses have been analyzed as silences that evince the hesitancy, disfluency, or ‘social ineptitude’ of the speaker (Goldman-Eisler 1968). Yet, interlocutors’ between-turn pauses have been analyzed as meaning-potent variables that can promote or undermine those who have just spoken, as in Mendoza-Denton (1994)’s investigation of the Hill-Thomas hearings. Though ethnographic analyses have contributed to our understanding of the social meaning of linguistic features across contexts and topics (e.g., Podesva 2007), the assumption that pauses are primarily cognitively driven has hindered further study on pausing strategies and how they are received by listeners. This study tests the implications of Mendoza-Denton’s between-turn analysis on the within-turn speech of undergraduates, asking, “Do long/short phrase-final pauses affect how we perceive students’ personal qualities?”
This experiment utilized speech samples collected in a Cornell University classroom that were digitally manipulated to include long (1.6s) and short (.4s) pauses. Indeed, the results show that listeners hear short pausers as more “confident” and more “informed,” while long pausers are judged as less “confident” and less “informed.” These effects become most robust when the long pauses align with slower speech rate and the short pauses with faster speech rate such that the indexical meanings of the individual variables converge towards one coherent style. The study also shows effects for perceived gender (more “feminine” speakers of both sexes were rated as less “intelligent”) and High Rising Terminal intonation (HRT users were rated as less “intelligent” and less “confident”).
Overall, this paper unites an interest in style in Third Wave Sociolinguistics (Eckert 2005) with sociophonetic experimentation (Campbell-Kibler 2007) in order to disentangle the linguistic resources speakers draw upon in creating style. By testing naturally-occurring variants in a controlled environment, this study shows that pauses, intonation, and perceived gender are salient to listeners in the co-construction of academic styles.

References
Campbell-Kibler, K. (2007). Accent, (ING), and the social logic of listener perceptions. American Speech 82.1: 32-64.
Eckert, P. (2005). Variation, convention, and social meaning. Paper presented at the annual meeting of the Linguistic Society of America. Oakland, California: January 7, 2005.
Goldman-Eisler, F. (1968). Psycholinguistics: Experiments in spontaneous speech. London, New York: Academic Press.
Mendoza-Denton, N. (1994). Pregnant Pauses: Silence and Authority in the Hill- Thomas Hearings. In K. Hall & M. Bucholtz (Eds.) Gender Articulated: Language and the Socially Constructed Self. London: Routledge, 51-68.
Podesva, R. (2007). Phonation type as a stylistic variable: The use of falsetto in constructing a persona. Journal of Sociolinguistics 11.4: 478-504.

Research on dialect- and identity-based Arabic variation allows us to explore the sociolinguistic dynamics of Arabic speech communities beyond the influences of Classical Arabic (Haeri 1997). Four Cairene Arabic variants not present in Classical Arabic are the use of [tʃ] or [tj] for /t/ and /tʕ/and [dʒ] or [dj] for /d/ and /dʕ/. The affricate variants are referred to as "strong palatalization," while the palatalized variants are referred to as "weak palatalization" (Haeri 1997; Youssef 2010). This paper presents the results of an experiment testing a variety of social meaning potentials for strong palatalization, the more socially and perceptually salient of the two.

Haeri found strong palatalization to be prevalent in the interview speech of lower-class women under 30, but less frequent among older and upper-class women (70-3), and infrequent among all men (68). Haeri thus hypothesizes that the social meanings of women's strong palatalization include "blue collar", informal, independent, and tough (97-8). Haeri then explains that, though the meanings of men's palatalization are less clear, they are likely related to those of women's palatalization (96-7).

To test these social meanings, eight Egyptian Arabic words were recorded with and without strong palatalization by two Cairene women and two Cairene men aged 18 to 29. Four of the words realize the /t/, /tʕ/ ↔ [tʃ] alternation while the other four realize the /d/, /dʕ/ ↔ [dʒ] alternation. Each speaker also recorded eight filler words.

A free-response pre-test was completed to determine which adjectives were salient to listeners in describing the stimuli. The experiment was presented in Arabic script to 40 native Cairene Arabic speakers currently living in Egypt. I then chose eight adjectives from the set of pre-test responses whose token frequency was in the top 25% for use in a 6-point Likert scale Matched Guise-type Test. The adjectives were comprised of two groups: five adjectives Haeri's findings would predict to vary uniformly across the palatalized and non-palatalized forms (عامي "colloquial", مثقف"educated", غني"rich", لبق"refined", وﺍاثق "confident") and three adjectives Haeri’s findings would make no clear predictions about (غزلي "flirtatious", جمﻴﯿل"attractive", جاﺩد"earnest").

Significant results emerged for women and men across both groups of adjectives. Both the palatalizing men and the palatalizing women were judged to sound significantly less attractive, confident, earnest, educated, refined and so on than the non-palatalizing men and women. The men, however, are penalized more for palatalizing than the women. These findings suggest that, over the 20+ years since Haeri (1997) completed her ethnography, strong palatalization has become stereotyped to lower-class women's speech. More broadly, this paper illustrates the benefit of completing production and perception studies on the same sociolinguistic variable.

References:
Haeri, Niloofar. 1997. The Sociolinguistic Market of Cairo: Gender, Class and Education. London & New York: Kegan Paul International.
Youssef, Islam. 2010. Palatalization in Cairene Arabic: From phonetics to phonology. CASTL Colloquium Series, University of Tromsø.

Psycholinguistic experiments show that pronouns tend to be resolved differently depending on whether they occur in main or subordinate clauses. If a pronoun in a subordinate clause has more than one potential antecedent in the main clause, then the pronoun tends to refer to the antecedent which has a certain thematic role (depending on the verb and on the subordinating conjunction). In contrast, pronouns in main clauses tend to refer back to the subject of the previous main clause, and this tendency is not affected by any verbs or conjunctions. In natural language processing, these findings have recently led to a proposal that pronoun resolution systems should have a split architecture, i.e. that they should use different mechanisms for pronoun resolution in the two cases.
With the help of two parsed and coreference-annotated corpora, this paper estimates the impact of the split-architecture proposal. The findings of this work are as follows: (1) Subject pronouns in authentic texts behave the same way in main and subordinate clauses. (2) The number of sentences in which a split architecture would behave differently than a system that treats both cases the same way is close to zero. Therefore, a separate treatment of resolution within and across units is unlikely to improve the performance of any system. This result casts a doubt on the split-architecture proposal, and more generally on approaches that directly incorporate psycholinguistic results into performance-oriented algorithms for anaphora resolution without assessing the relative importance of the phenomena that underlie them.

This paper provides a unification-based implementation of Binding Theory (BT) for the English language in the framework of feature-based lexicalized tree-adjoining grammar (LTAG). The grammar presented here does not actually coindex any noun phrases, it merely outputs a set of constraints on co- and contraindexation that may later be processed by a separate anaphora resolution module. It improves on previous work by implementing the full BT rather than just Condition A. The main technical innovation consists in allowing lists to appear as values of semantic features.

In this presentation, I provide an account for the weak/strong ambiguity of donkey anaphora. I argue that the way this ambiguity is resolved depends on the goals of the agents and on the background assumption that the individuals in the model behave rationally.

An argument by Kratzer (2000) based on Schein (1986, 1993) does not conclusively show that events and thematic roles are necessary ingredients of the logical representation of natural language sentences. The argument claims that cumulative readings of "every" can be represented only with these ingredients. But scope-splitting accounts make it possible to represent cumulative readings of "every" in an eventless framework. Such accounts are motivated by obligatory reconstruction  effects of "every" and by crosslinguistic considerations. Kratzer proposes that "agent" but not "theme" occurs in the logical representation of sentences because this allows her to model subject-object asymmetries in the distribution of cumulative every. But the reason for these asymmetries seems to be that every must be c-commanded by another quantifier in order to cumulate with it, no matter what its thematic role is. So the distribution of cumulative "every" does not provide support for Kratzer’s proposal.

Constructions analyzed in this talk:

- Distributively interpreted NPs: "Three boys hiccupped." "Each man wore a green tie."
- for-adverbials: "run for fifty minutes" vs. "*run to the store for fifty minutes"; "run for five miles" vs. *run to the store for five miles"
- Measure constructions: "five pounds of books" vs. "*five pounds of book"; "five inches of snow" vs. "*five degrees Celsius of snow"

Goals of this talk:

- Analyze all constructions as instances of distributivity
- Increase explanatory adequacy by reducing the overall size of the grammar
- Increase descriptive adequacy for each construction by capitalizing on insights gained from the other ones
- Increase empirical testing ground for any theory that explains aspects of one of these phenomena

We explore configurations in which "dou" in Chinese is compatible with cumulative readings involving numeral quantifiers. Such readings pose prima facie problems for a view that uniformly translates dou as as Link’s (1983) star operator similar to English "each".

An NP-hardness proof for non-local Multicomponent Tree Adjoining Grammar (MCTAG) by Rambow and Satta (1992), based on Dahlhaus and Warmuth (1986), is extended to some linguistically relevant restrictions of that formalism. It is found that there are NP-hard grammars among non-local MCTAGs even if any or all of the following restrictions are imposed: (i) lexicalization: every tree in the grammar contains a terminal; (ii) dominance links: every tree set contains at most two trees, and in every such tree set, there is a link between the foot node of one tree and the root node of the other tree, indicating that the former node must dominate the latter in the derived tree. This is the version of MCTAG proposed in Becker et al (1991) to account for German long-distance scrambling. This result restricts the field of possible candidates for an extension of Tree Adjoining Grammar that would be both mildly context-sensitive and linguistically adequate.

This paper proposes a solution to Haddock's puzzle: Why do embedded definites ("the rabbit in the hat") have different presuppositions from what would be expected on a strictly compositional semantics? For example, "the rabbit in the hat" does not presuppose that there is a unique hat, but only that there is a unique combination of a hat and a rabbit in it. This is unexpected if the presupposition of "the hat" is computed locally. The solution is given in terms of the independently attested mechanisms of intermediate presupposition accommodation and quantifier raising: the inner definite is raised above the outer one, whose uniqueness presupposition is then accommodated into the inner definite. Since quantifier raising is subject to island constraints, the account predicts that embedded definites should be sensitive to islands. This prediction is confirmed in an experiment with 800 participants conducted with Amazon Mechanical Turk.

Zimmermann (2002) identifies two kinds of distance-distributive items across languages. The first kind (e.g. each) is restricted to distribution over individuals; the second kind (e.g. German jeweils) can also be interpreted as distributing over salient occasions. I explain this behavior by formally relating this split to the two distributivity operators proposed in the work of Link (atomic operator) and Schwarzschild (cover-based operator), which I reformulate in a Neo-Davidsonian framework.

Temporal dependencies provide a useful testbed for syntactic and semantic theories of meaning/context interaction. In this poster, I consider two kinds of temporal dependencies: those that hold between temporal adjuncts as in (1a) and (1b), and those that hold between before and its complement, as in (2a).

(1) a. Last year, it rained in the afternoon.
b. Last year, it rained every day.
(2) a. John left before Mary claimed she would arrive.
b. John left before Mary made the claim that she would arrive.

I show that the former are anaphora-like and the latter are movement-like and I provide a variable-free implementation of the former.