Adults learning a second language (L2) (“late learners”) have difficulty achieving a native speaker’s level of accuracy in both perception and production of L2 phonetic segments. This difficulty often results in deviant production of L2 segments that is perceived as accented speech by native speakers of that language. It is generally agreed that this failure in non native segmental production and perception is caused by previous linguistic experience with the first (L1) language. Late learners are expected to show stronger L1 effects than learners who learnt their L2 in early childhood (“early learners”). However, not all L2 phonetic segments are equally difficult for late learners. The learnability of L2 phonetic segments is thought to be perceptual in nature and depends on the perceived phonetic distance between them and the acoustically, phonetically and/or articulatorily most similar segment(s) in the learner’s L1 phonetic inventory. It is generally assumed that specific L2 segments will be perceptually related or assimilated to the most similar L1 segment(s) even if there is a detectable acoustic difference between them. The studies reported in this thesis examined Serbian-English bilinguals’ perception and production difficulties with two Australian English vowel contrasts that are not contrastive in Serbian: /e/ - /æ/ and /i:/ - //. We compared participants who began learning English before 5 years (“early”) versus those who began after 15 years (“late”). In Study 1and Study 2 early learners discriminated and produced both contrasts equally well, whereas late learners had greater difficulties perceiving and producing /e/ - /æ/. In Study 3 a priming paradigm was applied to discrimination and perceptual assimilation tasks in which the prime and target contain phonologically identical, phonetically similar or phonologically and phonetically unrelated vowels under two interstimulus intervals (ISI) that tap phonological versus phonetic levels of processing, according to prior research. Early versus late group differences suggest that discrimination and production accuracy reflect how listeners assimilate Australian English vowels to native Serbian vowels. “Early” and “late” learners related L2 vowels to L1 differently, which reflects differences in establishment of the L1 phonetic system at the time of L2 onset. / Doctor of Philosophy (PhD)

This dissertation examines the role of the native dialect in non-native perception and production in the specific case of Cuban and Peninsular Spanish as native varieties and of English vowels as in hat, hut, hot as the target. In most second language studies, the learners’ native variety is assumed to be homogenous, regardless of their regional variety. Nevertheless, regional varieties differ in non-trivial ways and such differences need to be considered when analyzing second language acquisition. This dissertation takes shape around the main research question of whether potentially systematic differences between vowels in the native dialects of Cuban and Peninsular Spanish would produce matching phonetic differences in non-native perception and production of English vowels. This question was addressed in three experiments that compared native vowels in these Spanish varieties, as well as the perception and production of the target English vowels as in hat, hut, hot by Cuban and Peninsular Spanish learners. Significant cross-dialectal differences were identified in the production of native vowels, namely, locations of /i, o/, durations of /e, a, o/ as well as different variability patterns, which were predicted to influence mapping of sounds in L2 perception and production. In L2 perception, discrimination for advanced and naïve listeners from each dialect was tested with English contrasts as in hat-hut, hut-hot, and hat-hot. No clear native dialect effect could be identified; however, Cuban advanced listeners obtained high error rates with hut-hot. In L2 production, though, dialect-specific patterns were found: Peninsular learners produced vowels as in hat, hut, hot with significant spectral but no durational differences whereas Cuban learners produced vowels as in hut, hot with no significant spectral but with duration differences. I concluded that the native dialect was one of several factors generating the differences between Peninsular and Cuban participants’ perception and production of English vowels as in hat, hut, hot. The conjoint effect of the native dialect, input and learning experience were shown to have contributed to the distinctions. This finding contributes to second language acquisition research because it stresses the need to control for learners’ native dialect. This research contributed new acoustic data on Cuban Spanish and on L2 English. It uncovered specific patterns and interlanguage strategies of Spanish learners of English.

This dissertation examines the role of the native dialect in non-native perception and production in the specific case of Cuban and Peninsular Spanish as native varieties and of English vowels as in hat, hut, hot as the target. In most second language studies, the learners’ native variety is assumed to be homogenous, regardless of their regional variety. Nevertheless, regional varieties differ in non-trivial ways and such differences need to be considered when analyzing second language acquisition. This dissertation takes shape around the main research question of whether potentially systematic differences between vowels in the native dialects of Cuban and Peninsular Spanish would produce matching phonetic differences in non-native perception and production of English vowels. This question was addressed in three experiments that compared native vowels in these Spanish varieties, as well as the perception and production of the target English vowels as in hat, hut, hot by Cuban and Peninsular Spanish learners. Significant cross-dialectal differences were identified in the production of native vowels, namely, locations of /i, o/, durations of /e, a, o/ as well as different variability patterns, which were predicted to influence mapping of sounds in L2 perception and production. In L2 perception, discrimination for advanced and naïve listeners from each dialect was tested with English contrasts as in hat-hut, hut-hot, and hat-hot. No clear native dialect effect could be identified; however, Cuban advanced listeners obtained high error rates with hut-hot. In L2 production, though, dialect-specific patterns were found: Peninsular learners produced vowels as in hat, hut, hot with significant spectral but no durational differences whereas Cuban learners produced vowels as in hut, hot with no significant spectral but with duration differences. I concluded that the native dialect was one of several factors generating the differences between Peninsular and Cuban participants’ perception and production of English vowels as in hat, hut, hot. The conjoint effect of the native dialect, input and learning experience were shown to have contributed to the distinctions. This finding contributes to second language acquisition research because it stresses the need to control for learners’ native dialect. This research contributed new acoustic data on Cuban Spanish and on L2 English. It uncovered specific patterns and interlanguage strategies of Spanish learners of English.

Thesis (M.S.)--State University of New York at Binghamton, Department of Psychology, 2009. / Includes bibliographical references.

Recent studies have shown that even highly-proficient Spanish-English bilinguals, who acquired their second language (L2) in childhood and have little or no foreign accent in English, may require more acoustic information than monolinguals in order to identify English vowels and may have more difficulty than monolinguals in understanding speech in noise or reverberation (Mayo, Florentine, and Buus, 1997; Febo, 2003). One explanation that may account for this difference is that bilingual listeners use acoustic cues for vowel identification differently from monolinguals (Flege, 1995).In this study, we investigated this hypothesis by comparing bilingual listeners use of acoustic cues to vowel identification to that of monolinguals for six American English vowels presented under listening conditions created to manipulate the acoustic cues of vowel formant dynamics and duration. Three listener groups were tested: monolinguals, highly proficient bilinguals, and less proficient bilinguals. Stimulus creation included recording of six target vowels (/i, I, eI, E, ae, A/) in /bVd/ context, spoken in a carrier phrase by four American monolinguals (two females, two males). Six listening conditions were created: 1) whole word, 2) isolated vowel, 3) resynthesized with no change, 4) resynthesized with neutralized duration, 5) resynthesized with flattened formants, and 6) resynthesized with flattened formats and neutralized duration. The resynthesized stimuli were created using high-fidelity synthesis procedures (Straight; Kawahara, Masuda-Katsuse, and Cheveigne 1998) and digital manipulation. A six-alternative forced choice listening task was used. The main experiment was composed of 240 isolated vowel trials and 48 whole word trials.

published_or_final_version / abstract / toc / Speech and Hearing Sciences / Master / Master of Science in Audiology

This thesis presents an acoustic vowel space analysis of F1 and F2 frequencies from 10 speakers of Pijal Media Lengua (PML) and 10 speakers of Imbabura Quichua (IQ). This thesis also provides a brief grammatical discription of PML with insights into contrasts and similarities between Spanish, Quichua and other documented varieties of Media Lengua (ML). ML is typically described as a mixed language with a Quichua morphosyntactic structure where almost all content words are replaced by their Spanish-derived counterparts through the process of relexification. I use mixed effects models to test for statistical significance between PML Spanish-derived vowels and Quichua-derived vowels. The results provide suggestive data for (1) co-existing vowel systems in moderate contact situations and (2) moderate evidence for co-exsiting vowel systems in extreme contact situations. Results also show that PML may be manipulating as many as eight vowels and IQ as many as six.

This thesis presents an acoustic vowel space analysis of F1 and F2 frequencies from 10 speakers of Pijal Media Lengua (PML) and 10 speakers of Imbabura Quichua (IQ). This thesis also provides a brief grammatical discription of PML with insights into contrasts and similarities between Spanish, Quichua and other documented varieties of Media Lengua (ML). ML is typically described as a mixed language with a Quichua morphosyntactic structure where almost all content words are replaced by their Spanish-derived counterparts through the process of relexification. I use mixed effects models to test for statistical significance between PML Spanish-derived vowels and Quichua-derived vowels. The results provide suggestive data for (1) co-existing vowel systems in moderate contact situations and (2) moderate evidence for co-exsiting vowel systems in extreme contact situations. Results also show that PML may be manipulating as many as eight vowels and IQ as many as six.

Research indicates that adult Second Language (L2) learners typically do not become native-like in their perception or production of their L2, likely as a result of interference from their native language (L1). Research also indicates that L2 perception and production nonetheless improves with increased experience with the L2. Until recently, however, theories of L2 acquisition (e.g., SLM: Flege, 1995; PAM: Best, 1994) have been vague in their account of the processes underlying this improvement. The recent PAM-L2 (Best and Tyler, 2007), however, opens up new ways to understand experiential change in L2 perception and production. Centrally, PAM-L2 suggests that a large L2 vocabulary curtails change in L2 perception and production because it forces the learner to settle on an accented version of the L2 phonology. The present thesis introduces the Vocabulary-Tuning Model of L2 Rephonologisation (Vocab Model). This model extends PAM-L2 by highlighting the facilitating effect of L2 vocabulary expansion, in early L2 immersion when the L2 vocabulary is still small, on the perception and production of an L2. It is further argued that the processes underpinning this improvement are analogous to those that underpin L1 acquisition in infants and toddlers. The thesis tests the Vocab Model in a series of studies (cross-sectional as well as longitudinal) of the perception and production of Australian English vowels by native speakers of Japanese, who have recently arrived in Australia for the purpose of acquiring English. The results show that L2 vocabulary size is indeed associated with L2 vowel perception and production and thus support the predictions of the Vocab Model. The thesis examines the usefulness of different criteria for L2-L1 vowel assimilation and discusses the findings in relation to results from L1-vowel perception research. The research design also pioneers a ‘whole system’ approach to cross-language vowel perception research that allows the learners to use all native vowels and all native vowel combinations (all three thesis studies), and to apply them to the full inventory of L2 vowels (Study 1). It is argued that results from such an approach more appropriately reflect the actual perceptual flexibility of the learners in a natural L2-immersion context than would a smaller subset of L1 and/or L2 vowels. This ‘whole system’ approach further suggests that L1 phonotactics is worthwhile to consider in future studies of L2 segmental perception and production. / Doctor of Philosophy (PhD)

Proefschrift (Ph. D.)--Rijksuniversiteit te Leiden, 1983. / Vita. Text in English; foreword, summary and vita in Dutch. Includes bibliographical references (p. 145-153).