Alumni Dissertations

 

Alumni Dissertations

Filter Dissertations By:

 
 
  • Phoneme restoration methods reveal prosodic influences on syntactic parsing: data from Bulgarian

    Author:
    Iglika Stoyneshka-Raleva
    Year of Dissertation:
    2013
    Program:
    Linguistics
    Advisor:
    Janet Fodor
    Abstract:

    ABSTRACT Phoneme restoration methods reveal prosodic influences on syntactic parsing: Data from Bulgarian by Iglika Stoyneshka-Raleva Adviser: Professor Janet Dean Fodor This dissertation introduces and evaluates a new methodology for studying aspects of human language processing and the factors to which it is sensitive. It makes use of the phoneme restoration illusion (Warren, 1970). A small portion of a spoken sentence is replaced by a burst of noise. Listeners typically mentally restore the missing phoneme(s), reporting that they heard a complete utterance with the noise burst overlaid on it. Phoneme restoration is used in the present study to explore the role of prosodic phrasing, specifically the impact of a major prosodic break on the interpretation of a sentence where no other type of disambiguation remains. Materials are constructed so that the phoneme(s) that listeners indicate they had heard reveal which syntactic/semantic interpretation of the sentence they had computed. Two target constructions in Bulgarian are investigated in a series of experiments, using the same materials throughout: NP/S coordination as in (1) and RC attachment as in (2). The target items are originally disambiguated by morphological agreement: number agreement between subject and verb in (1), and gender agreement between the head noun and the relative pronoun in (2). The sentences in each pair differ with respect to one or more phonemes in the word on which the morphological disambiguation is encoded; these are the phonemes that are noise-replaced in the experimental materials. (Prosodic boundaries are indicated by ||.) (1) a. Nakraia sreshtnahme Ani || i Ivan i Mimi biaha vav vaztorg. In the end meet-past-1p.pl Ani and Ivan and Mimi were in ecstasy `In the end, we met Ani and Ivan and Mimi were in ecstasy.' b. Nakraia sreshtnahme Ani i Ivan || i Mimi bee vav vaztorg. In the end meet-past-1p.pl Ani and Ivan and Mimi was in ecstasy `In the end, we met Ani and Ivan and Mimi was in ecstasy.' (2) a. Podtseniha advokata || na pevitsata koiato kupi imenieto. underestimate-past lawyer-m of singer-f who-f buy-past estate-det `(They) underestimated the lawyer of the singer who=N2 bought the estate.' b. Podtseniha advokata na pevitsata || kojto kupi imenieto. underestimate-past lawyer-m of singer-f who-m buy-past estate-det `(They) underestimated the lawyer of the singer who=N1 bought the estate.' The alternative structures for both (1) and (2) are also disambiguated by the location of a major prosodic boundary, as shown. For the NP/S coordination a boundary after the first noun (1a) signals that only that noun (Ani) is the object of the first clause; a boundary after the second noun (1b) signals that the first two nouns (Ani and Ivan) form a coordinate object in the first clause. For the RC construction, a boundary after the first noun (2a) favors the interpretation that the RC modifies the second noun (the singer), whereas a boundary after the second noun (2b) favors RC modification of the entire complex NP (the lawyer of the singer). After noise-replacement of the disambiguating phonemes, only these prosodic boundaries differentiate the competing interpretations. Three different response tasks are employed to tap which phonemes listeners mentally restored. In a visual word choice task, participants indicate which of two words (e.g. biaha / bee in the case of (1)) they had `heard' in the stimulus sentence (Experiments 1a-c). In a sentence repetition task, participants repeat back the sentence after hearing it (Experiments 2a-b). In a sentence shadowing task, participants repeat back the sentence as they are listening to it (Experiments 3a-b). In all three tasks, responses confirmed that prosodic boundary location provided viable disambiguation of syntactic structure. However, the effect of prosody differed for the two constructions. Both prosodic contours disambiguated NP/S coordination equally well, but a break after the first noun was a stronger structural cue for RC attachment than a break after the second noun. These findings for Bulgarian comport well with data from more traditional methodologies on prosodic influences on coordination and RC attachment interpretation in other languages, thereby validating the sensitivity of the new methodology. A broader range of other ambiguities in Bulgarian were explored in a subsequent pilot experiment. Also, the original target constructions were tested in written form, with a simulated `ink blot' obscuring the morphological agreement. Two tasks were employed: silent reading with visual word choice (Experiments 4a-b) and silent reading with sentence repetition (Experiments 5a-b). Visual grouping of phrases disambiguated NP/S coordination quite effectively, but had little impact on RC attachment interpretation. The advantage of the phoneme restoration approach is that it is unobtrusive, resembling everyday listening in a partly noisy environment. It draws no attention to the presence of ambiguity or to the relevance of prosody, it does not interrupt the stimulus sentence, and no comprehension question intrudes between stimulus and response. It is equally well-suited to studies of parsing sensitivity to non-acoustic factors such as syntactic priming, discourse context or lexical frequencies.

  • Japanese -te iru and -te aru: The Aspectual Implications of the Stage-Level and Individual-Level Distinction

    Author:
    Mamori Sugita
    Year of Dissertation:
    2009
    Program:
    Linguistics
    Advisor:
    William McClure
    Abstract:

    This dissertation investigates semantic and syntactic properties of the forms –te iru and –te aru in Japanese, as well as pragmatic effects of statements with these forms. With an activity verb in the –te iru form, progressive, experiential, and habitual readings are available. With an achievement verb in the –te iru form, perfective, experiential, and habitual readings are available. I address specifically the difference between perfective and experiential readings. After reviewing the literature, where it seems that the distinction is not clear, I give a series of empirical tests and argue that experiential sentences exhibit properties of individual–level predicates, while perfective (as well as progressive) sentences exhibit properties of stage–level predicates. There are two types of –te aru sentences, intransitivizing and non–intransitivizing –te aru, both of which have been claimed to yield perfective readings. However, I argue that all –te aru sentences are experiential and exhibit properties that parallel individual–level predicates. Formally, I propose that progressive and perfective –te iru are represented as sets of events with a requirement that the event be realized. In contrast, I propose that experiential –te iru and –te aru are represented as sets of individuals with a requirement that the event be realized. The relative scope difference of the event and individual variables in the semantic representation reflects the stage–level and individual–level distinction. Progressive and perfective –te iru denote properties of events, while experiential –te iru and –te aru denote properties of individuals. The stage–level/individual–level distinction is also reflected in the proposed syntax. Progressive and perfective –te iru sentences have raising structures, while experiential –te iru and –te aru sentences have control structures. The scope of the event and individual arguments in the semantics of –te iru and –te aru is reflected in the position of their subjects in syntax. Lastly, I argue that habitual –te iru sentences parallel experiential–te iru sentences in that they also exhibit properties of individual–level predicates.

  • Morphological and Phonological Factors in the Production of Verbal Inflection in Adult L2 Learners and Patients with Agrammatic Aphasia

    Author:
    Malgorzata Szupica-Pyrzanowski
    Year of Dissertation:
    2009
    Program:
    Linguistics
    Advisor:
    Gita Martohardjono
    Abstract:

    Failure to supply inflection is common in adult L2 learners of English and agrammatic aphasics (AAs), who are known to resort to bare verb forms. Among attempts to explain the absence of inflection are competing morphological and phonological explanations. In the L2 acquisition literature, omission of inflection is explained in terms of: mapping (Epstein et al., 1996; Prévost & White, 2000), failed modular interaction (Lardiere, 1998), L1 morpho-syntactic constraints governing the activation of the L2 features (Hawkins & Liszka, 2003), L1 prosodic organization which differs from that of L2 (Goad, et al., 2003), and L1 phonological constraints on final consonant clusters (Lardiere, 2003). In agrammatism, inflectional omission has been linked to: processing (Thompson et al., 2002), accessing (Kehayia et al., 1990), impaired rule implementation (Lee & Thompson, 2005), productive (Bird et al., 2003; Mathews & Obler, 1997) and receptive phonology (de Mornay Davies, 2006). Most, if not all, investigations of inflectional omission have focused on populations whose L1 lacks the syntactic representation of inflection. In this study, we concentrate instead on two groups (L1 Polish learners of L2 English and AAs of L1 English) who can be assumed to have an underlying representation of inflectional material because the L1 already has it; yet they show difficulties in the use of inflection. With regard to production, we therefore asked: What contributes more to the problems encountered by these speakers, morphology or phonology? To test this, we administered an elicited production task varying either the morphological or the phonological complexity of the environment of the inflectional morpheme. We hypothesized that if non-target production of inflection is constrained by morphological factors, we would likely see the following: 1. Both groups would perform better on mono- than bi-morphemic homophones. 2. The participants would differentiate between homophonous morphemes, e.g., PLUR, AGR, POSS and show different degrees of omission for these. 3. There would be no significant difference between inflection of mono-syllabic existing verbs and mono-syllabic pseudo verbs because the rule attaching inflection is present and intact. On the phonological side, we made the following predictions: 1. Sonority of the final segment of stems would affect the production of inflection. 2. Syllabic suffixes would be produced more accurately than their non-syllabic counterparts because of their saliency. 3. Shorter (mono-syllabic) verbs would be affixed better than longer (bi- and tri-syllabic) ones. Results show that in the production of inflection, similar patterns were found in these two fundamentally different populations, L2 learners and AAs. Morphological constraints seem to play a greater role in the omission of inflection than phonological ones. On the other hand, phonology is used by both groups as a compensatory strategy to preserve inflection.

  • Understanding Doubly Center-Embedded Sentences Through Contrastive Focus

    Author:
    Ashley Thorne
    Year of Dissertation:
    2014
    Program:
    Linguistics
    Advisor:
    Janet Fodor
    Abstract:

    A production and comprehension-judgment experiment examined the relationship between prosody and comprehensibility in doubly center-embedded relative clause constructions in English. Lexically identical sentences with contrastive emphasis on NP1, NP2, or VP1, and a baseline version for comparison, were read aloud for recording and judged for comprehensibility. Contrast on NP2 and VP1 yielded higher comprehensibility judgments compared to baseline than contrast on NP1 compared to baseline. This was contrary to the prediction that contrast on NP1 would encourage the prosodic pattern NP1 || RC || VP3, which in previous work has been reported to be optimal for comprehension. Several explanations for this outcome are considered, including the possibility that contrastive focus does not reliably induce a following prosodic boundary if phrase lengths are not supportive of it.

  • Input Consistency in the Acquisition of Questions in Bulgarian and English: A Hypothesis Testing Model

    Author:
    Lidiya Tornyova
    Year of Dissertation:
    2011
    Program:
    Linguistics
    Advisor:
    Virginia Valian
    Abstract:

    The goal of this dissertation is to address several major empirical and theoretical issues related to English-speaking children's difficulties with auxiliary use and inversion in questions. The empirical data on English question acquisition are inconsistent due to differences in methods and techniques used. A range of proposals about the source of auxiliary omission and inversion errors have been presented but they remain with a limited cross-linguistic scope and cannot explain the data in full. This dissertation addresses the issues by investigating the influence of various target-specific properties on children's production of wh- and yes/no questions in a cross-linguistic perspective. We propose a Hypothesis Testing model of language acquisition in which target specific properties of the language determine the types of errors and how difficult they are to correct. The more contrasts the language supports, the greater the number of incorrect hypotheses the child can entertain, and the longer it takes for children to acquire the target. Cross-linguistic variation in language acquisition is determined by differences in the amount of relevant contrasts supported in various target grammars. Languages with uniform properties facilitate the hypothesis-testing process by providing a narrow range of hypotheses to be considered and are learned easily. Languages with non-uniform properties extend the course of acquisition by providing a broad range of hypotheses to be tested. To test the proposed hypothesis testing model we conducted four elicited imitation experiments investigating 2-year-old children's auxiliary use and inversion in Bulgarian and English wh- and yes/no questions. We predicted similar performance on auxiliary use for the two language groups because English and Bulgarian are both inconsistent and support relevant contrasts regarding auxiliary inclusion. With respect to auxiliary inversion, Bulgarian is consistent and English is not. English supports three relevant contrasts that Bulgarian does not. Thus, we predicted that Bulgarian-speaking children would perform better on auxiliary/verb inversion than English-speaking children. Our results support the Hypothesis Testing model.

  • ADICITY AND REFERENCE: MIDDLE VOICE AND ITS COMPONENTS

    Author:
    Erika Troseth
    Year of Dissertation:
    2009
    Program:
    Linguistics
    Advisor:
    Robert Fiengo
    Abstract:

    ADICITY AND REFERENCE: MIDDLE VOICE AND ITS COMPONENTS In this thesis I provide an analysis of middle voice sentences (as in The book reads well, El libro se lee bien, Das Buch liest sich leicht) in which the characterizing feature of middles is a mismatch with respect to predicate adicity and the number of argument expression occurrences in the syntactic structure. Throughout the thesis I rely on the distinction between linguistic types and linguistic tokens. Thus, although it might rightly be said, when considering orthography or phonology, that in the sentence Lolita si legge facilemente, there are two items: si and Lolita, we can also rightly say, when considering syntax or semantics, that together si and Lolita constitute a single abstract object. A significant feature of the analysis is indeed the proposal that the syntactic subject of middles and the weak reflexive together formally constitute a single syntactic object. The analysis predicts the various properties of the weak reflexive that appears in many languages' middle voice sentences, including their Case, referential, and agreement properties. Taking the aforementioned mismatch to be the core characterization of middles predicts that they are morphologically and semantically less restricted than previously thought. Data presented in the thesis support this conclusion.

  • The Social Correlates of Lexical Borrowing in Spanish in New York CIty

    Author:
    Rachel Varra
    Year of Dissertation:
    2013
    Program:
    Linguistics
    Advisor:
    Ricardo Otheguy
    Abstract:

    This dissertation investigates lexical borrowing in Spanish in New York. English-origin lexical material was extracted from a stratified sample of 146 Spanish-speaking informants of different ages, national origins, classes, etc., living in New York City. ANOVAs and Pearson correlations determined whether lexical borrowing frequency and the type of vocabulary used (i.e. idiosyncratic and shared) were correlated with speaker traits. Results show that all speakers, regardless of their English proficiency or how long they have been in the U.S., borrow. However, borrowing frequency is best predicted by immigrant generation. Furthermore, borrowing rate may play a distinct role for first and second generation Spanish speakers. In the first generation, the middle class, those with more education, better proficiency in English and more Spanish confidence borrow most often. In the second generation, Puerto Ricans and those with more English confidence borrow most. Results for borrowing vocabulary suggest that the middle class is more responsible for introducing novel vocabulary to the Spanish-speaking community than the working class. Overall, though, both novel and shared vocabulary are integral components of speakers' borrowing inventories. Finally, this dissertation examines flagging (e.g. pauses, fillers, metalinguistic commentary) near other-language strings to determine whether flagging is indicative of language awareness or linguistic disfluency. Results showed no support for the latter; but evidence for the former interpretation suggests that a simultaneous, albeit slight, process of deborrowing accompanies lexical borrowing in New York. The findings and the methodology of this dissertation contribute to several fields of language study. First, the definition of lexical borrowing used makes it germane to studying borrowing in situations of on-going, face-to-face contact in multidialectal communities. Furthermore, a corpus-based approach to differentiating between two types of bilingual speech phenomena, lexical borrowing and codeswitching, is offered. Second, findings show that borrowing in New York is not a `deficit' behavior and that some aspects of identity (such as arrival age and class) cross-cut traditional (i.e. regional) characterizations of U.S. Latinos. Finally, findings for bilingual speech partially corroborate models describing the long-term outcomes of contact that predict that the quantity of contact features observable in a language is a function of the intensity of contact.

  • Prosody and Parsing in a Double PP Construction in Hebrew

    Author:
    Ronit Webman Shafran
    Year of Dissertation:
    2011
    Program:
    Linguistics
    Advisor:
    Janet Fodor
    Abstract:

    It is a standard finding that speakers reliably produce prosodic cues to clause boundaries and listeners use these cues in parsing sentences. However, considerable uncertainty remains regarding whether the same applies to syntactic phraseboundaries. A long series of studies in the parsing literature on ambiguous PP attachment constructions, such as Susan hit the man with the umbrella,has yielded mixed results. This dissertation investigates the prosody-syntax interface in the processing of a double PP construction in Hebrew. Selection restrictions force the first prepositional phrase (PP1) to attach low, but attachment of the second one (PP2) is ambiguous: it can attach maximally high to VP (as an argument of described) or maximally low to the NP inside PP1 (modifyingmarriage ). Dana VP[ te&lsquoara &lsquoet NP[ ha-kšayim PP1[ be-nisu&lsquoe-ha PP2[ la-šadran Dana described ACC the-difficulties in-marriage-her to the-broadcaster A length contrast in PP2 was also examined. PP2 was either short (as here) or long (with addition of a modifier to the short version). This double PP construction exhibits a sharper structural contrast between the two potential attachment sites than the long-studied single-PP construction: there is a greater discontinuity in the syntactic tree for the high attachment analysis, which could encourage a stronger prosodic break before the ambiguous PP, yielding more reliable results than for the single-PP construction. An advantage of conducting the experiment in Hebrew is that the acoustic markers of prosodic phrase boundaries, which include a final high boundary tone, are clearer than in English. Two experiments were conducted. The first was a combined production-comprehension study examining the relationship between preferred interpretation and preferred prosodic phrasing in reading aloud. The results showed a reliable association between high attachment of PP2 and the presence of a prosodic break immediately preceding it, though as predicted there were significantly more instances of this pattern (prosodic break + high attachment) for long than for short PP2. The second experiment tested comprehension of the same items in silent reading. PP2 length effects on attachment were very similar in silent reading and reading aloud, providing new insight into the Implicit Prosody Hypothesis (Fodor 2002).

  • Understanding and interpreting Japanese NP1 wa NP2 da sentences: Mechanism and contextual factors

    Author:
    Megumi Yoshida
    Year of Dissertation:
    2013
    Program:
    Linguistics
    Advisor:
    William McClure
    Abstract:

    This dissertation investigates the contextual factors that affect the understanding and interpretation of one Japanese topicalized construction, NP1 wa NP2 da sentences, by native speakers of Japanese. The construction allows two possibilities in the relation between the NP1 and the NP2. When the two NPs are not syntactically connected (Type I), the sentence is generally vague, and a particular context is required to specify the meaning. When they are syntactically connected (Type II), they can refer to a semantically identical referent, and the sentence is naturally interpreted as an identity sentence. The aim of the study is to examine how context determines the meaning of Type I and Type II NP1 wa NP2 da sentences. These sentences were examined in a set of controlled experimental contexts by two kinds of test: Understandability and Interpretation. Results showed that readers generally tried to connect the NP 1 wa NP2 da sentences to the context syntactically, semantically, or pragmatically when the sentences were presented in a context. Specifically, a syntactic and semantic relation with a particular verb in the preceding context sentence and the NP2 or information about a particular place presented by a locative frame enhanced the comprehension of Type I NP1 wa NP2 da sentences. When these contextual factors were presented consistently and appropriately, Type II NP1 wa NP2 da sentences could be interpreted as non-identity sentences. When such context was not available, the interpretations tended to depend on the sentence-internal conceptual connection between the NP1 and the NP2 in both types of NP1 wa NP2 da sentences. These results suggest the reader's use of their linguistic and pragmatic knowledge differs according to the context and sentence type. The results also reveal a new understanding of the `aboutness' relation, a notion that accounts for the non-syntactic connections between the topic and the predicate. Specifically, in the process of understanding NP1 wa NPsub>2 da sentences, establishing an `aboutness' relation refers to the process of finding an appropriate predicate in the context to create a proposition to connect the predicate (NP2) to the topic NP (NP1).