Linguistics Dissertations and Theses

Permanent URI for this collection

Browse

Recent Submissions

  • Publication
    Acoustic and perceptual evidence of complete neutralization of word-final tonal specification in Japanese
    (University of Kansas, 2002-05-31) Maniwa, Kazumi
    This study investigates the extent to which Japanese lexical pitch-accent distinction is neutralized in word-final position. Native speakers of Tokyo Japanese produced minimal word pairs differing in final accent status. Words were produced both in isolation and in a sentential context, where neutralization would not be expected due to following tonal specification. Examination of pitch patterns on relevant moras revealed a clear distinction between accent-opposed pairs produced in context but no such difference between items produced in isolation. Both the words produced in isolation and the words excised from sentential contexts were then presented to Japanese listeners in a lexical identification task. Participants could clearly distinguish items extracted from sentences but identified words uttered in isolation at chance level. These results suggest that phonological neutralization of final pitch accent is complete, showing no effects of underlying specification in either production or perception.
  • Publication
    Acoustics and Perception of Clear Fricatives
    (University of Kansas, 2007-05-31) Maniwa, Kazumi
    Everyday observation indicates that speakers can naturally and spontaneously adopt a speaking style that allows them to be understood more easily when confronted with difficult communicative situations. Previous studies have demonstrated that the resulting speaking style, known as clear speech, is more intelligible than casual, conversational speech for a variety of listener populations. However, few studies have examined the acoustic properties of clearly produced fricatives in detail. In addition, it is unknown whether clear speech improves the intelligibility of fricative consonants, or how its effects on fricative perception might differ depending on listener population. Since fricatives are the cause of a large number of recognition errors both for normal-hearing listeners in adverse conditions and for hearing-impaired listeners, it is of interest to explore these issues in detail focusing on fricatives. The current study attempts to characterize the type and magnitude of adaptations in the clear production of English fricatives and determine whether clear speech enhances fricative intelligibility for normal-hearing listeners and listeners with simulated impairment. In an acoustic experiment (Experiment I), ten female and ten male talkers produced nonsense syllables containing the fricatives /f, &thetas;, s, [special characters omitted], v, δ, z, and [y]/ in VCV contexts, in both a conversational style and a clear style that was elicited by means of simulated recognition errors in feedback received from an interactive computer program. Acoustic measurements were taken for spectral, amplitudinal, and temporal properties known to influence fricative recognition. Results illustrate that (1) there were consistent overall clear speech effects, several of which (consonant duration, spectral peak location, spectral moments) were consistent with previous findings and a few (notably consonant-to-vowel intensity ratio) which were not, (2) 'contrastive' differences related to acoustic inventory and eliciting prompts were observed in key comparisons, and (3) talkers differed widely in the types and magnitude of acoustic modifications. Two perception experiments using these same productions as stimuli (Experiments II and III) were conducted to address three major questions: (1) whether clearly produced fricatives are more intelligible than conversational fricatives, (2) what specific acoustic modifications are related to clear speech intelligibility advantages, and (3) how sloping, recruiting hearing impairment interacts with clear speech strategies. Both perception experiments used an adaptive procedure to estimate the signal to (multi-talker babble) noise ratio (SNR) threshold at which minimal pair fricative categorizations could be made with 75% accuracy. Data from fourteen normal-hearing listeners (Experiment II) and fourteen listeners with simulated sloping elevated thresholds and loudness recruitment (Experiment III) indicate that clear fricatives were more intelligible overall for both listener groups. However, for listeners with simulated hearing impairment, a reliable clear speech intelligibility advantage was not found for non-sibilant pairs. Correlation analyses comparing acoustic and perceptual style-related differences across the 20 speakers encountered in the experiments indicated that a shift of energy concentration toward higher frequency regions and greater source strength was a primary contributor to the "clear fricative effect" for normal-hearing listeners but not for listeners with simulated loss, for whom information in higher frequency regions was less audible.
  • Publication
    An ERP investigation of individual differences in the processing of wh-dependencies by native and non-native speakers
    (University of Kansas, 2018-08-31) Covey, Lauren
    This study examines the processing of wh-dependencies by native English speakers and Mandarin Chinese-speaking learners of English. Wh-dependencies involve a long-distance relationship between a fronted wh-word (e.g., who) and the position in the sentence where it originated, called a gap site. The examination of wh-dependency resolution presents an interesting test case for whether or not grammatical knowledge is used online because, in languages such as English, wh-movement is constrained such that extraction is only possible from certain positions and is barred from other positions, called islands (Ross, 1967). In examining whether native speakers and second language (L2) learners are sensitive to island constraints online, this study tests predictions of prominent L2 processing theories which argue that adult learners are unable to utilize abstract grammatical information during processing (e.g., Clahsen & Felser, 2006). In the native literature, the processing of wh-dependencies has been proposed to involve at least two distinct processes, a predictive process in which the parser searches for a potential gap site, and an integrative process, when the dependency is successfully resolved at the gap site. The broader electrophysiological literature has linked these qualitatively different processes to distinct event-related potential (ERP) components: the N400 for prediction (e.g., Federmeier, 2007; Lau et al., 2008; 2013; Michel, 2014; Van Berkum et al., 2005), and the P600 for syntactic integration (e.g., Gouvea et al., 2010; Kaan et al., 2000; Phillips et al., 2005). Although previous ERP studies have examined these components independently, few studies have tracked the dynamics of wh-dependency resolution across the sentence, examining both prediction and integration to investigate whether these processes are indexed by unique components. The present study takes this approach, focusing on the processing of wh-dependencies at three critical regions across the sentence, two of which are associated with prediction, and one with integration. This study additionally investigates the extent to which the use of grammatical knowledge during online processing and the ability to engage in predictive processing is modulated by proficiency in L2 learners, and performance on a range of cognitive measures in native speakers. Results show that both native speakers and highly proficient learners engage in gap prediction during processing, although this is limited to certain contexts for learners. In examining processing inside of an island, a position from which extraction is prohibited, the current study shows that native speakers and highly proficient learners are guided by grammatical knowledge. Finally, both natives and learners show evidence of successful dependency resolution at the actual gap site, even in sentences with islands. Overall, the results present a complex picture of processing wh-dependencies by native English speakers and Mandarin-speaking learners of English, showing that while both native speakers and learners with higher proficiency are able to use grammatical information during online processing, the contexts in which L2 learners are able to predict differ from native speakers.
  • Publication
    A lower Chehalis phonology
    (University of Kansas, 1969) Snow, Charles T.
  • Publication
    Sibilant Contrast: Perception, Production, and Sound Change
    (University of Kansas, 2017-08-31) Li, Mingxing
    This study examines sibilant place contrast in the [_i] context in terms of its typology across Chinese dialects and its role in the historical development of Mandarin sibilants. The typology across 170 Chinese dialects reveals that (i) for dialects that have sibilants at three places (dental, palatal, and retroflex), place contrasts in the [_i] context are generally avoided, e.g., */si-ɕi-i/; (ii) for dialects that have sibilants at two places, mostly dental vs. palatal, place contrasts in the [_i] context also tend to be avoided, e.g., */si-ɕi/; (iii) for dialects that do have contrastive dental vs. palatal sibilants in the [_i] context, the place contrast of affricates implies that of fricatives. The first two patterns are mirrored in the sound changes of Mandarin in that contrastive dental and palatal sibilants in the [_i] context that emerged from independent processes have always been enhanced or avoided. In addition, the sound changes also showed an avoidance of contrastive dental vs. palatal in the [_i] context with the shift of palatal sibilants into retroflex sibilants from the 11th to the 14th century. The connection between the synchronic typology and the diachronic changes raises a number of research questions: (i) Does the vowel context affect the perceptual distinctiveness of sibilant place contrasts, e.g., is [si-ɕi] less distinct than [sa-ɕa]? (ii) Do place contrasts differ in perceptual distinctiveness, e.g., is [si-ɕi] less distinct than [si-i]? (iii) Do different manners of articulation differ in perceptual distinctiveness, e.g., is [tsi-tɕi] less distinct than [si-ɕi]? These issues were investigated through a speeded AX-discrimination experiment, which has been shown to be able to evaluate the relative perceptual distinctiveness of sound pairs independent of the listener’s native phonology. Twenty-nine listeners were put under time pressure to judge if a CV pair is the same or not, where the sibilant onsets of the CV pairs contrast in place (e.g., [si-ɕi]) and the vowels were [i] vs. other vowels (e.g., [si-ɕi] vs. [sa-ɕa] vs. [sɹ̩-ɕi]). Assuming that a longer response time indicates less perceptual distinctiveness, the results showed that (i) the [_i] context reduces the perceptual distinctiveness of the place contrasts of dental vs. palatal sibilants; (ii) the introduction of the apical vowel enhances the perceptual distinctiveness between the contrastive sound pairs; (iii) the dental vs. retroflex contrasts are more distinct than the dental vs. palatal contrasts. These findings match the observations in the cross-linguistic typology and the historical development of Mandarin and support the claim that perceptual distinctiveness regulates the phonological system. The reduced distinctiveness of dental vs. palatal sibilants in the [_i] context suggests that contrastive dental vs. palatal sibilants are unstable and are likely to be avoided in sound change. A phonetic study was conducted on the sibilants in Xiangtan, a Chinese dialect reported to have the same sound system as 18th century Mandarin (i.e., [sɹ̩ si i ɻ̩]) with fully contrastive dental vs. palatal sibilants in the [_i] context. It is predicted that in Xiangtan, the pre-[i] dentals in /si tsi tsʰi/ may show signs of being palatalized and thus neutralized with the palatals. Natural productions of /si tsi tsʰi/, /ɕi tɕi tɕʰi/, and /sɹ̩ tsɹ̩ tsʰɹ̩/ syllables with matched tones were recorded from 11 native female speakers of Xiangtan. Center of gravity, energy dispersion, intensity, and duration were extracted for three types of sibilants: Canonical dentals as in /sɹ̩/, canonical palatals as in /ɕi/, and pre-[i] dentals as in /si/. A discriminant analysis was performed by first training a classifier on the canonical dentals and canonical palatals and then using the classifier to predict the place (dental vs. palatal) of the pre-[i] dentals. Native Mandarin listeners were also recruited to identify the isolated first half of the pre-[i] dental sibilants as being dental vs. palatal. The results from both studies showed that (i) some Xiangtan speakers have palatalized the dentals in /si tsi tsʰi/, and (ii) certain speakers variably produce dental and palatal sibilants for the same lexical item. Therefore, the results support the contention that dental and palatal contrasts are perceptually less distinct in the [_i] context and the variation in the realization of the pre-[i] dentals indicates that a merger replicating the development of Mandarin is in process. In general, the perceptual experiment reveals that dental vs. palatal sibilants in the [_i] context form weak contrasts, based on the psychoacoustic similarity of the contrastive elements. The avoidance of weak contrasts is observed in cross-linguistic typology, historical sound change, and speech production. This study thus establishes an empirical connection among the perceptual distinctiveness of sibilant place contrasts, the production of these contrasts, cross-linguistic typology, and historical sound changes.
  • Publication
    How Native Chinese Listeners and Second-Language Chinese Learners Process Tones in Word Recognition: An Eye-tracking Study
    (University of Kansas, 2017-08-31) Qin, Zhen
    Chinese and English differ in the types of information they use to convey meaning in words: unlike English, Chinese uses lexical tones (i.e., pitch movement) to contrast word meanings (e.g., in Chinese, the word “ma” can mean mother, hemp, horse, or scold depending on its lexical tone). This difference between Chinese and English poses word recognition difficulties for English-speaking learners of Chinese in using lexical tones to recognize Chinese words. Existing research on L2 learners’ perception and processing of lexical tones has focused on whether native listeners of languages that do not have lexical tones can discriminate and identify lexical tones. To date, no study to our knowledge has examined how L2 learners use the fine-grained phonetic details of tonal information in the time course of spoken word recognition — that is, as the speech signal unfolds over time. In fact, little research has looked into the time course with which native listeners use the fine-grained phonetic details of tonal information in spoken word recognition. This doctoral dissertation examines how native Chinese listeners and highly proficient adult English-speaking learners of Chinese use tonal information in spoken word recognition as the speech signal unfolds in time. More specifically, this research uses the visual-world eye-tracking paradigm to shed light on the precise time course with which native and non-native listeners use tonal information in online word recognition. The proposed research aims to investigate two potential differences between native listeners and highly proficient English-speaking L2 learners of Chinese in their use of tonal information as the speech signal unfolds: (i) their potentially different incremental use of the early pitch height before pitch contour information of the tone is available; (ii) their potentially different sensitivities to fine-grained within-category gradience of level and contour tones in the word recognition process. Experiment 1 investigates whether or not native and non-native listeners make similar use of early between-category pitch height (T1-T4 with similar early pitch height vs. T1-T2 with different pitch height) before pitch contour information is available. A visual-world eye-tracking experiment in Chinese was conducted with two groups of participants: 36 native Chinese listeners and 26 highly proficient English-speaking L2 learners of Chinese. The target was either T1 or T2 word in T1-T2 condition whereas the target was either T1 or T4 word in T1-T4 condition. The auditory stimuli were natural tonal tokens. The time-window analyses on fixations showed that early pitch height constrained both Chinese and English listeners’ lexical access. While Chinese listeners started using early pitch height in the time window in which pitch contour information was not available, English listeners started using early pitch height in the time windows in which pitch contour information had been available, and showed more tonal competition than Chinese listeners. The findings suggest that whether or not prosodic cues contribute to distinguishing among words in the L1, and how they do so, influence listeners’ use of these cues in spoken word recognition. Experiment 2 investigates whether native Chinese listeners and English-speaking L2 learners of Chinese differ in using the within-category gradience of level and contour tones to recognize spoken words. Another visual-world eye-tracking experiment in Chinese was conducted with the same participants. The target was a level tone (i.e., T1) and the competitor was a high-rising tone (i.e., T2), or vice versa. The auditory stimuli were manipulated such that the target tone was either canonical in the standard condition, acoustically more distant from the competitor in the distant condition, or acoustically closer to the competitor in the close condition. Growth curve analysis on fixations suggested that Chinese listeners showed a gradient pattern of lexical competition, with decreased competition in the distant condition and increased competition in the close condition than in the standard condition for the contour tone; English listeners, on the other hand, showed increased competition in both the distant and close conditions relative to the standard condition for the level tone. These findings suggest that Chinese listeners may show sensitivity to fine-grained tonal variability when this variability is along a dimension (i.e., pitch contour) that is meaningful for distinguishing tones whereas English listeners might show sensitivity to the fine-grained tonal variability along a dimension (i.e., pitch height) encoded in their L1 lexical representations. Moreover, native and non-native listeners, who potentially differ in the robustness of their representations of lexical tones, may adopt different strategies to deal with fined-grained tonal information to resolve the lexical competition. The findings of this doctoral dissertation make a contribution to the understanding of how tonal information modulates lexical activation in native and non-native Chinese listeners. This research also has pedagogical implications for Chinese language teaching.
  • Publication
    Explicit teaching of Japanese mimetic words using voicing, gemintion, and reduplication rules
    (University of Kansas, 2017-05-31) Nakata, Kotoko
    Mimetics are commonly used by Japanese native speakers to express the manner of actions and sensations. However, they are often not taught explicitly in many Japanese language classrooms. The current study tested a novel teaching methodology to help English-speaking learners of Japanese learn Japanese mimetics. Second language learners were explicitly taught three phonological/morphological rules during learning. The three rules are: (i) voicing, (ii) gemination, and (iii) reduplication. In Japanese mimetics, these phonological/morphological factors systematically affect the meaning of mimetics. The current study examined whether explicitly teaching these three rules helps English-speaking learners of Japanese, who vary in Japanese proficiency, acquire mimetics as well as help them generalize these rules to newly encountered mimetics. The procedure used a Pretest-Learning-Posttest design. First, all participants took a Pretest. Approximately one week later, all participants learned mimetics during a Learning Session. In the Learning Session, all participants were taught 32 mimetic words with a verbal description and a static picture along with a sentence that contained the mimetic word. There were two different participant groups in the Learning Session: an Experimental Group and a Control Group. The Experimental group explicitly learned the three phonological/morphological rules while the Control group did not. Finally, all learners participated in a Posttest and a Delayed posttest (approximately 4 weeks later) to assess their retention of the mimetic vocabulary. We found that the novel teaching methodology (teaching mimetics with a picture and a context along with a verbal description) is effective in acquiring and remembering mimetics. Participants showed a great improvement after the Learning Session for both the trained mimetics and newly introduced mimetics, suggesting that participants successfully learned the mimetics and the sound regularities both with and without the explicit introduction of the three phonological/morphological rules. Additionally, we also found that learners who were explicitly taught the three phonological/morphological rules showed a greater improvement than those who were not. Therefore, the explicit introduction of the sound regularities is more effective in the current methodology. We also found that the proposed methodology is effective regardless of learners’ proficiency in Japanese. While advanced learners overall acquired more mimetics than beginning learners, beginning learners showed a greater improvement than advanced learners. These results suggest that teaching mimetics does not need to be limited to advanced learners (as it often is in Japanese language classrooms) but it should be encouraged for learners at all levels.
  • Publication
    Nominalization in Pulaar
    (University of Kansas, 2017-05-31) Ba, Ibrahima
    There are a few strategies to construct nominal structures and phrases. The Pulaar infinitive, for instance is used genitive nominalization and other non-finite clauses with nominal properties. Relative clause (RC) nominalization (headed relative clause and factive clauses) is also another nominalization process in the language. This dissertation lays out a description of these nominalization processes with a main focus on constructions involving the infinitive and relative clause nominalizations. The infinitive in Pulaar can have both nominal and verbal properties and this is usually indicated by an agreeing determiner but also through adjectival or adverbial modification. RC nominalization is a very prominent and productive in the language. The verb within the relative clause can occur with functional morphemes such as tense, aspect and negation. In addition, a relativized verb can have derivational morphemes like valency-changing affixes attached to it. The factive RC nominals can have an event, factive and manner interpretation whereas the GN nominal can have a generic, event and factive interpretation. This dissertation also places Pulaar within the typology with respect to this type of nominalization with similar nominalization types found in related African languages like Wolof, Ewe, Gungbe, Yoruba and other related languages, thus showing that Pulaar nominalization types fits within the typology of nominalization. Also, a brief parallel is drawn with English nominal constructions built on the gerund.
  • Publication
    The role of argument structure in Meꞌphaa verbal agreement
    (University of Kansas, 2017-08-31) Duncan, Philip Travis
    This dissertation explores aspects of Me\textcharis{ꞌ}phaa morphosyntax, from verb roots to verb-initial word orders. I argue that patterns of agreement map directly onto the syntax of argument structure, which in turn feed the language's unique manifestation of ergativity. Me\textcharis{ꞌ}phaa agreement morphology is richly complex, and I show that this is due, in part, to three core ``ergativity properties'' \citep{deal2015ergativity} coalescing in the language: transitive subjects and intransitive subjects are encoded differently for a subset of verbs (the ergative property), intransitive subjects are sometimes marked with the same morphology as transitive objects (the absolutive property), and split-intransitivity in the language yields differential marking for intransitives (the argument-structural property). This produces a system that is \textit{consistently} ergative, with a particular ergativity property being visible depending on what verbs are under comparison, and what clause types are involved. I propose an explanation of this rich agreement morphology by appealing to the syntax of argument structure. I argue that Me\textcharis{ꞌ}phaa's way of being ergative is not about verbs \textit{per se}, but verbal structures \citep{marantz2013verbal} with distinct functional components and configurations. Language-specific unaccusativity diagnostics and other tests point to the existence of a constellation of verbal structures. Orienting to structural diversity reveals how Me\textcharis{ꞌ}phaa's patterns of verbal agreement reflect a high degree of sensitivity to underlying geometries. Me\textcharis{ꞌ}phaa agreement exponents reflect their probe, and an array of functional heads in the verbal domain participate in Agree(ment). This means that higher functional heads (e.g., T) are not required for such operations in the language. Instead, the very pieces involved in building verbal structures are the ones responsible for determining verbal agreement. In addition to laying the foundation for building the clause from the verb up, showing how the verb itself offers a snapshot of the core clausal architecture, this work further accounts for how the verbal and inflectional domains interact to produce verb-initial orders. I propose a VP-remnant raising account for Me\textcharis{ꞌ}phaa, attending to aspects of the derivation that successfully account for both morpheme order in the verb stem as well as constituent order at the larger clausal level.
  • Publication
    THE USE OF SEGMENTAL AND SUPRASEGMENTAL INFORMATION IN LEXICAL ACCESS: A FIRST- AND SECOND-LANGUAGE CHINESE INVESTIGATION
    (University of Kansas, 2017-08-31) Connell, Katrina S.
    The present study investigated first language (L1) and second language (L2) Chinese categorization of tones and segments and use of tones and segments in lexical access. Previous research has shown that English listeners rely more on pitch height than pitch direction when perceiving lexical tones; however, it remains unclear if this superior use of pitch height aids English-speaking learners of Chinese in identifying the tones of Chinese that differ in initial pitch height. The present study aimed to investigate this issue to determine whether this pitch height advantage aids English-speaking Chinese learners in identifying the tones of Chinese by looking at the time course of categorization of Chinese tones that differed in initial pitch as well as segments. A norming study was first conducted to investigate the duration of acoustic input needed to hear tone and segment (rime) distinctions. In a gated AX discrimination task, native Chinese listeners and naïve English listeners heard increasingly large fragments of tonal pairs and segmental pairs that varied in the expected disambiguation point. The results of this norming study were used to select tonal and segmental stimulus pairs were controlled (as best as is possible) for the disambiguation timing in the next two experiments. Experiment 1 investigated the time course of categorization of tones and segments using a forced-choice gating task designed to tap into listeners’ identification of fragment categories taken from syllables that differ only in tones or only in segments. Native Chinese listeners and L1-English L2-Chinese listeners heard a single fragment of a Chinese word and identified either the tone or the rime of the heard fragment from two presented options. The results showed that the segmental contrasts had higher accuracy than tonal contrasts for both groups. The L2-Chinese listeners performed comparably to the native listeners on both tonal and segmental contrasts, and L2 Chinese listeners showed no advantage over native listeners. The second goal of this study was to investigate the time course of the use of tones and segments in lexical access. Previous work has shown that native Chinese listeners use tones and segments simultaneously in lexical access. Previous work on how second language learners of Chinese use tones in lexical access compared to segments showed that tones and segments are used at the same time; however, work in the segmental domain suggest that this should not be the case, and learners should struggle to use the new tones in online lexical access. As such, this work aimed to reinvestigate the timing of use of tones and segments in second language Chinese, as well as to compare learners’ use of tones and segments to native listeners with a highly time-sensitive measure: visual-world eye-tracking. Experiment 2 investigated the time course of use of tones and segments in online spoken word recognition for L1 and L2 groups. The same segmental and tonal pairs used in Experiment 1 were used in a visual-world eye-tracking experiment. Native Chinese listeners and L1-English L2-Chinese listeners saw two pairs of words displayed as corresponding images: one tonal pair and one segmental pair. Eye movements were recorded as participants heard a single target word in isolation and clicked on the corresponding picture. The eye movement data revealed that native Chinese listeners use tones and segments to begin constraining the lexical search at approximately the same time, and tonal information constrained the search more rapidly than did segments. The L2-Chinese learners showed segmental use comparable to that of native listeners; however, their tonal use was delayed by approximately 100 ms. In terms of speed, learners also showed more rapid use of tones in constraining the lexical search, although tones and segments were used to constrain the lexical search more slowly than they did for native listeners. These results are discussed in relation to recent L1 studies on lexical access of tones and segments and computational modeling of suprasegmental information. The results of this research is in line with previous work that showed tones and segments are used to constrain lexical access simultaneously; however, the current work does not support the conclusion that tones and segments are used in the same way, with tones constraining the lexical search faster than segments. It is suggested that the cause of this tone speed advantage is the number of competitors removed from competition when the processor is certain of a tone as opposed to certain of a segment or even rime. The present results also speak to the literature on the use of segmental and suprasegmental information in a second language and suggest that the timing of use of different cues to lexical identity is dependent on if that cue is used in the L1, since segments were processed at the same time as native speakers while tones were delayed. Speed of use seems to be independent of whether or not it is used in the L1, with both tones and segments being processed slower overall compared to native listeners.
  • Publication
    Second language lexical processing: influence of teaching method and word characteristics
    (University of Kansas, 2004) Pastoriza Espasandín, Diana
    Word learning was investigated in two experiments: a word translation experiment and a picture naming experiment. Two groups of bilinguals, differing in second language proficiency, were taught 40 Spanish words using one of these two tasks. One group of participants translated a set of words from L1 (English) to L2 (Spanish) Another group of participants named pictures. For each task, the training involved two presentations of the same 40 Spanish words, coupled either with the translation in English or with a picture. In both experiments subjects heard each Spanish words repeated 3 times in each presentation. Subjects' task was to name the Spanish word either given an English word prompt (word translation) or a picture prompt (picture naming). The stimulus materials were manipulated on word frequency and cognate status. The results show that cognate and high frequency words were easier to learn (fewer errors and shorter response times). Proficiency and task hardly affected error rates. Overall, picture naming showed better recall for beginner learners whereas word translation showed better recall for learners with a good proficiency level. This suggests that conceptual memory appeared to operate as much in the translation task as in the picture naming task.
  • Publication
    Production and perception of Korean and English word-level prominence by Korean speakers
    (University of Kansas, 2015-08-31) Lee, Goun
    Prominence refers to the relative emphasis that may be given to a syllable in a word (word-level prominence) or to one or more words in a phrase (phrase-level prominence). Korean has been claimed to have both word-level (Ko, 2013) and phrase-level (Jun, 1996) prominence, with the former realized mainly with duration and the latter with F0 height. However, given the claim that younger Korean speakers have lost duration as the main cue expressing word-level prominence (Kim & Han, 1998; Magen & Blumstein, 1993), it is not clear if and how younger Korean speakers produce word-level prominence. Thus, the primary goal of the current dissertation is to examine whether Korean still has word-level prominence. Two experiments investigated this research question in two domains – acoustically (Experiment 1) and perceptually (Experiment 2). Given the findings regarding the status of lexical stress in Korean, we further investigated which acoustic correlates/cues Korean learners of English are able to transfer from their L1 prosodic cues, and whether they can acquire a new cue that does not exist in their L1. Thus, the secondary purpose of this study is to investigate which acoustic correlates/cues Korean L2 learners of English utilize in producing and perceiving English lexical stress. These questions are addressed and examined in Experiments 3 and 4. In the acoustic study of the production of Korean word-level prominence (Experiment 1), measurements of duration, intensity, F0, F1, and F2 on (so-called) Korean stress minimal pairs by older and younger Korean speakers revealed that only at the sentence level, duration and intensity systematically distinguish stress pairs for the older speakers. A perception study on word-level prominence in Korean (Experiment 2) revealed that both older and younger Korean listeners weighted the duration cue most heavily in identifying minimal pairs of Korean word-level prominence when two of the suprasegmental cues were orthogonally manipulated in each syllable. Interestingly, this perceptual weighting was only observed in the first syllable: none of the listeners changed their perception when cues were signaling second-syllable stress. Based on these findings from an acoustic and a perception study, we conclude that Korean does not have word-level prominence, but only has a phonemic vowel length distinction. In the acoustic study on the production of English word-level prominence (Experiment 3), measurements of duration, intensity, F0, F1, and F2 on English stress pairs found that Korean learners were able to use not only all suprasegmental cues to indicate lexical stress in English, but also acquire a new cue (e.g., vowel reduction) that does not exist in their native language, although in a non-native like manner. The results of the perception study on word-level prominence in English (Experiment 4) revealed that when identifying English stress pairs, Korean learners weighted vowel reduction more heavily than any suprasegmental cues. Both intensity and F0 were weighted in Korean learners’ perception; however, duration was not weighted at all, although younger Korean speakers still retain the phonemic vowel length distinction in their L1. Taken together, the current dissertation increases our understanding of the status of lexical stress in Korean, as well as the extent to which L2 learners produce and perceive L2 lexical stress by transferring prosodic features from their native language.
  • Publication
    The nature of variation in tone sandhi patterns of Shanghai and Wuxi Wu
    (University of Kansas, 2016-08-31) Yan, Hanbo
    The primary goal of this dissertation is to understand the variation patterns in suprasegmental processes and what factors influence the patterns. To answer the questions, we investigated the variation patterns of tone sandhi in the Shanghai and Wuxi Wu dialects of Chinese. Shanghai disyllables and trisyllables have been documented to have two different sandhi patterns: tonal extension and tonal reduction. Some items can only undergo tonal extension, some items can only undergo tonal reduction, and some can variably undergo either type of sandhi. Previous works have indicated that the syntactic structure, semantic transparency, and lexical frequency of the items all play a role in the sandhi application. Additionally, the morpheme length of trisyllabic items (1+2, 2+1) is also expected to affect their sandhi application. A variant forms’ goodness rating experiment, together with a lexical frequency rating experiment and a semantic transparency rating experiment, showed that syntactic structure has a primary effect on sandhi application in general. It overrides the effect of semantic transparency, especially in modifier-noun items. The nature of the lexical frequency effect in Shanghai is related to the syntactic structure of the lexical item. Morpheme length effect is not found. Wuxi disyllables and trisyllables also have been observed to have two different sandhi patterns: tonal substitution and no sandhi. Some items can only apply tonal substitution, and some can apply either form variably. Syntactic structure and semantic transparency have been reported to affect Wuxi sandhi application, and morpheme length is also expected to have an effect in trisyllabic sandhi application. The three rating experiments conducted in Wuxi found that due to the lexical listedness of the opaque substitution pattern, frequency influences both modifier-noun and verb-noun positively, although modifier-noun prefers tonal substitution form more. Semantic transparency effect is only apparent for verb-noun disyllables. Moreover, morpheme length also distinguishes sandhi application between 1+2 and 2+1 modifier-noun items. In all, by using quantitative rating experiments, the present study shows that tone sandhi variation is regulated by both grammatical factors, such as syntactic structure, morpheme length, phonological opacity, and nongrammatical factors, such as lexical frequency.
  • Publication
    Tracking Bilingual Activation in the Processing and Production of Spanish Stress
    (University of Kansas, 2016-08-31) Martinez García, María Teresa
    Language bias and proficiency have been proposed to modulate cross-language activation, but it is unclear how they operate and whether they interact. This study sheds light on these questions by investigating whether stress differences between Spanish-English cognates (material, final-syllable stress in Spanish) affect how first-language(L1)-Spanish second-language-(L2)-English and L1-English L2-Spanish bilinguals recognize Spanish words (materia ‘subject/matter,’ second-syllable stress in Spanish). In two eye-tracking experiments and a production experiment, participants heard or produced trisyllabic Spanish targets with second-syllable stress (materia) and saw four orthographic words, including the target and a Spanish-English cognate competitor. Cross-language activation was examined by manipulating the English cognate stress; in comprehension, English cognates with the same stress as the Spanish target (target: materia, competitor: material) were predicted to cause more cross-language interference than non-cognates (e.g., target: asado ‘roasted,’ competitor: asador ‘rotisserie’) and than English cognates with a different stress (target: litera ‘bunk bed,’ competitor: literal;); in production, cognates with a different stress were instead predicted to cause cross-language interference. Experiment 1 was in Spanish; Experiment 2 was in Spanish and English, and participants were assigned to Spanish-bias or English-bias condition; Experiment 3 was a production version of Experiment 2. Second-language (L2) proficiency was assessed with cloze tests and LexTALEs. In Experiment 1, neither group showed interference from English stress. In Experiment 2, only the L1-English bilinguals showed interference from English stress, and they did so only in the English-bias condition, with this effect decreasing with increasing Spanish proficiency. In Experiment 3, both groups showed interference from English stress, but this effect was modulated by language bias only for the L1-English group. These findings indicate that cross-language interference is more likely to emerge when the unintended language is the L1, when listeners are biased towards hearing the L1, and when L2 proficiency in the intended language is lower.
  • Publication
    Processing morphologically complex words in native and non-native French
    (University of Kansas, 2016-08-31) Coughlin, Caitlin E.
    This dissertation investigates how individuals who learned French after childhood process inflected French verbs. Two experiments test the hypothesis that non-native speakers lack the grammatical representation responsible for processing inflection in the manner that native speakers are able to. Experiment 1 uses a masked priming lexical decision task to test if native and non-native French speakers are able to decompose inflected words into stem and affix, and access a morphological level of representation in the lexicon. Experiment 2 uses the same task as Experiment 1, but incorporates electroencephalography (EEG) to investigate the time-course of lexical access in native and non-native French speakers. The results of both Experiment 1 and 2 indicate that non-native French speakers process inflectional information in a qualitatively similar way as native speakers. Additionally, the ability to process inflection in a native-like way is not restricted to learners at higher levels of proficiency; morphological processing is found across a wide range of proficiency levels. The results of the two experiments suggest that the grammatical representations and brain mechanisms responsible for processing inflection are available to adult second language learners, and may be available even in the early stages of acquisition.
  • Publication
    Observing the contribution of both underlying and surface representations: Evidence from priming and event-related potentials
    (University of Kansas, 2016-08-31) Chien, Yu-Fu
    This dissertation aims to uncover the role of the acoustic input (the surface representation) and the abstract linguistic representation (the underlying representation) as listeners map the signal during spoken word recognition. To examine these issues, tone sandhi, a tonal alternation phenomenon in which a tone changes to a different tone in certain phonological environments, is investigated. This dissertation first examined how productive Mandarin tone 3 sandhi words (T3 → T2/___T3) are processed and represented. An auditory priming lexical decision experiment was conducted in which each disyllabic tone 3 sandhi target was preceded by a tone 2 monosyllable (surface-tone overlap), a tone 3 monosyllable (underlying-tone overlap), or an unrelated monosyllable (unrelated control). Lexical decision RTs showed a tone 3 (underlying-tone overlap) facilitation effect for both high and low frequency words. A second priming study investigated the processing and representation of the more complex and less productive Taiwanese tone sandhi. Lexical decision RTs, examining sandhi 24 → 33 and 51 → 55, showed that while both sandhi types exhibited facilitatory priming effects, underlying tone primes showed significantly more facilitation than surface primes for sandhi 24 → 33, while surface tone primes showed significantly more facilitation than underlying primes for sandhi 51 → 55, with both effects modulated by frequency. A third study used event-related potentials (ERPs) to examine Mandarin tone 3 sandhi. Using an oddball paradigm, participants passively listened to either Tone 2 standards ([tʂu2 je4] /tʂu2 je4/), Tone 3 standards ([tʂu3 je4] /tʂu3 je4/), Tone Sandhi standards ([tʂu2 jen3] /tʂu3 jen3/), or Mix standards (i.e., both tone 3 sandhi and tone 3 words), occasionally interspersed with a tone 2 word [tʂu2] (i.e., the deviant). Results showed a mismatch negativity (MMN) in the Tone 2 condition but not in the Sandhi condition, suggesting different neural processing mechanisms for Tone 2 and Sandhi words. Together, the current data suggest that the underlying tone contributes more to the processing of productive tone sandhi and the surface tone contributes more to the processing of less productive tone sandhi. In general, this dissertation provides evidence for the representation and processing of words that involve phonological alternation, both within the same language and across different languages.
  • Publication
    A Grammatical Sketch of Comox
    (University of Kansas, 1981) Harris, Herbert Raymond
    Comox is a Salish language spoken on both sides of the Gulf of Georgia on the west coast of British Columbia, Canada. Before white contact, it was spoken on Vancouver Island from the Salmon River in the north to the present day town of Comox in the south. It was also spoken on Quadra Island and other islands in the Strait and at sites on the mainland. It presently exists in two distinct dialects whose most striking difference is an s-theta correspondence and an aveolar-interdental affricate correspondence. The mainland dialect is spoken at three sites by about 500 people. The island dialect is spoken by two people at a reserve near the town of Comox half way up the east coast of Vancouver Island. The phonology of the mainland dialect was described in a master's thesis by John Davis at the University of Victoria, 1970. The only published description of island Comox is an article by Sapir on noun reduplication. Since the island dialect is about to die out and has not been extensively studied, this dissertation attempts it description before the dialect is lost. The data used were gathered in two summers of fieldwork in 1970 and 1971 at the Comox Reserve. The data presented in Sapir's article are also used as well as some data gathered by Boas in 1889 and presently in the Smithsonian Institution. The syntactic model of analysis is a transformational case grammer and the phonology, a generative model using the features of Chomsky and Halle. After a brief discussion of the model of analysis and the data available, an overview of the cultural and linguistic context of Comox is presented. The material for this overview is taken from secondary sources and does not represent original work. The majority of the dissertation is taken up with an exposition of the syntax and phonology of the language. Syntactically Comox is a VSO language. The order of cases following the verb is: Agent Dative Benefactive Object Instrument Locative and Time. Agreement transformations add suffixes to the predicate corresponding to the first two cases present. There are three object suffix paradigms: (1) a transitive, (2) a detransitive which indicates the agent caused the object to do the predicate. "He made me dig," where "dig" is the verb, "he" the subject and "me" the object. And finally (3) the unintentional transitive object. The three subject suffix paradigms are: (1) the intransitive, (2) the transitive and (3) the embedded. There are three tenses (present, past and future) and five aspects (perfect, imperfect, stative, inceptive and continuative). Nominals are reduplicated for the plural and diminutive as are verbal stems for the imperfect. A set of phrase structure rules is formulated and possible transformations discussed. The phonology consists of thirty-eight rules, seven of which are precyclic. Three of the pre-cyclic rules are spelling rules for reduplication. Among the more unusual phonological rules is the metathesis of glottal stops and resonants. Underlying "y" and "w" become "j" and "g" respectively. Nasals become voiced stops under circumstances that are not completely clear. The underlying segment inventory includes glottalized resonants, stops, and affricates. There are five vowels. Aspects of the historical development of Comox are discussed briefly in connection with various features of the syntax and phonology.
  • Publication
    The online use of markedness information in L1 and L2 Spanish gender agreement
    (University of Kansas, 2015-12-31) Lopez Prego, Beatriz
    ABSTRACT The present study focuses on the acquisition and processing of gender agreement by second language (L2) learners of Spanish, whose first language (L1; English) lacks gender. Some L2 theories argue that these learners will not be able to acquire gender, and will have to resort to different strategies to process it in their second language (Hawkins, 2009), particularly in long-distance agreement dependencies (Clahsen & Felser, 2006; Clahsen, Felser, Neubauer, & Silva, 2010). Other theories argue that it is possible for those learners to acquire gender, but they may experience difficulty accessing target gendered forms, due to the computational burden of using a second language (Haznedar & Schwartz, 1997; Prévost & White, 2000). The current study addresses these theories by investigating how native speakers and advanced L2 learners use the gender markedness information (masculine vs. feminine) conveyed by the first element in a long-distance agreement dependency in particular, to process the second agreeing element in the dependency. In addition, it is investigated whether native speakers performing a task under processing burden show similar patterns to L2 learners in their processing of gender agreement (Hopp, 2010; McDonald, 2006; López Prego & Gabriele, 2014). This latter approach attempts to test whether specific error patterns in L2 learners emerge due to processing difficulty, or to a flawed representation of the gender feature. Thus, the study contributes unique data to answer the following questions: whether advanced L2 learners can establish long-distance agreement dependencies; whether they can develop a native-like representation of the gender feature in their L2, when they lack gender in their L1; and whether they can use gender information in a native-like manner in their online processing of agreement. These questions were tested in a self-paced reading task in which a grammaticality judgment was provided after each sentence. The group of native speakers performing under processing burden was additionally asked to decide whether a string of numbers presented before each sentence was the same or different from a string presented after the grammaticality judgment was supplied. The main results of the study showed that the advanced L2 learners tested, like the native speaker control group, were sensitive to gender agreement violations in long-distance agreement dependencies. In addition, both groups revealed a significant facilitation effect from the marked (feminine) feature in their processing of long-distance agreement dependencies, crucially, in grammatical sentences. Finally, the native speakers performing under processing burden showed some weak patterns that nevertheless resembled those in the L2 learner group. Thus, the findings from the present study support theories that posit computational difficulty as the source of agreement variability in L2 learners, and run counter to theories proposing a grammatical deficit in the L2 grammar as the cause of agreement errors in learners.
  • Publication
    An analysis of the Arabic Pidgin spoken by Indian workers in Saudi Arabia
    (University of Kansas, 1996) Hobrom, Anwar
  • Publication
    INDIVIDUAL DIFFERENCES IN PREDICTIVE PROCESSING: EVIDENCE FROM SUBJECT FILLED-GAP EFFECTS IN NATIVE AND NONNATIVE SPEAKERS OF ENGLISH
    (University of Kansas, 2015-08-31) Johnson, Adrienne Marie
    This study examined whether native and non-native speakers of English show evidence of predictive processing in the domain of syntax, and investigated the extent to which linguistic and non-linguistic factors modulate prediction (e.g., Grüter et al., 2014; Hopp, 2013; Kaan, 2014). Using self-paced reading, this study examined wh- dependencies, focusing on filled-gap effects in the pre-verbal, subject position. Specifically, this study manipulated the distance between the filler and potential gap position to investigate the proposal that increasing the distance between the filler and potential gap may force the parser to strongly commit to the subject gap analysis (e.g., Clifton & Frazier, 1989; Lee, 2004). A clear subject filled-gap effect was found in the condition in which the filler was immediately adjacent to the potential subject gap position for both English native speakers and Korean learners of English. Thus, both native and non-native speakers of English are able to immediately predict subject gaps and do not need additional distance. No effects were found in the long-distance condition for either natives or non-natives, suggesting that the increased processing burden may have hindered rather than facilitated the generation of a strong prediction for a subject gap. Greater attentional control, as measured by the Stroop task, was associated with larger subject filled-gap effects for both populations. Thus, these results provide evidence that non-native speakers are able to predict syntactic structure and actively resolve wh- dependencies and that the cognitive abilities that impact prediction are qualitatively similar in both native and non-native speakers (Kaan, 2014).