Distributional learning is typically understood as (unattended) tracking of stimulus probabilities. Distributional training with speech yields mixed results and the influencing factors have not yet been fully investigated. This study explored whether prior linguistic experience could have an effect on distributional learning outcomes. Czech and Greek adults, whose native languages contain and lack abstract length categories, respectively, were exposed to novel vowels falling into unimodal or bimodal distributions along the durational dimension. A trending interaction suggested that the Czechs and the Greeks might have been affected differently by the distributional exposure. Improved discrimination of the “trained” contrast was observed in bimodally exposed Czechs (whose prior expectations about length categories could guide learning) and, rather surprisingly, in unimodally exposed Greeks (who, lacking any expectations, might have listened in a noncategorical, auditory mode). Prior linguistic experience could thus affect whether and how experienced language users exploit new distributional speech statistics. This proposal needs to be assessed in future studies.

As a western Slavic language of the Indo-European family, Czech is closest to Slovak and Polish. It is spoken as a native language by nearly 10 million people in the Czech Republic (Czech Statistical Office n.d.). About two million people living abroad, mostly in the USA, Canada, Austria, Germany, Slovakia, and the UK, claim Czech heritage (Ministry of Foreign Affairs of the Czech Republic 2009). However, it is not known how many of them are native speakers of Czech.

Some (though not all) previous studies have documented the interlanguage speech intelligibility benefit (ISIB), i.e. the greater intelligibility of non-native (relative to native) speech to non-native listeners as compared to native listeners. Moreover, some studies (again not all) found that native listeners consider foreign-accented statements as less truthful than native-sounding ones. We join these two lines of research, asking whether foreign-accented statements sound more credible to non-native than to native listeners and whether difficult-to-process (less comprehensible) utterances are less credible. In two experiments we measure the intelligibility, comprehensibility and credibility of native and foreign-accented statements for native listeners and non-native listeners matched or mismatched in L1 with non-native talkers. We find an ISIB in both matched and mismatched non-native listeners, and an analogous matched comprehensibility benefit. However, we obtain no evidence of an interlanguage speech credibility benefit. Instead, both matched and mismatched non-native listeners tend to trust native statements more (i.e. statements produced by their target-language models). For native listeners, we do not confirm the tendency to mistrust non-native statements, but we do find a moderate correlation between the comprehensibility and credibility of foreign-accented utterances, giving limited support to the hypothesis that decreased perceptual fluency leads to decreased credibility.

Ultimate attainment in foreign-language sound learning is addressed via vowel production accuracy in English spoken by advanced Czech EFL learners. English FLEECE-KIT, DRESS-TRAP, and GOOSE-FOOT contrasts are examined in terms of length, height, and backness. Our data show that, while being constrained by phonemic category assimilation (new vowel height distinctions are not created), the learners' interlanguage combines phonological parsimony (reusing L1 length feature to contrast L2 vowels) with phonetic flexibility (within-category shifts reflecting L1-L2 phonetic dissimilarity). Although achieving nativelike phonological competence may not be possible learners who acquire L2 in the prevailingly L1 environment, the Czech learners' implementations of English vowels revealed their ability to adjust for phonetic detail of L2 sounds. Keywords:English as a foreign language, L2 phonology, ultimate attainment, vowels 1 AOL stands for the Age of Learning, i.e. age at the onset of acquisition. The most frequently cited AOL studies include e.g.

Seeing a person's mouth move for [ga] while hearing [ba] often results in the perception of "da." Such audiovisual integration of speech cues, known as the McGurk effect, is stable within but variable across individuals. When the visual or auditory cues are degraded, due to signal distortion or the perceiver's sensory impairment, reliance on cues via the impoverished modality decreases. This study tested whether cue-reliance adjustments due to exposure to reduced cue availability are persistent and transfer to subsequent perception of speech with all cues fully available. A McGurk experiment was administered at the beginning and after a month of mandatory face-mask wearing (enforced in Czechia during the 2020 pandemic). Responses to audio-visually incongruent stimuli were analyzed from 292 persons (ages 16-55), representing a cross-sectional sample, and 41 students (ages 19-27), representing a longitudinal sample. The extent to which the participants relied exclusively on visual cues was affected by testing time in interaction with age. After a month of reduced access to lipreading, reliance on visual cues (present at test) somewhat lowered for younger and increased for older persons. This implies that adults adapt their speech perception faculty to an altered environmental availability of multimodal cues, and that younger adults do so more efficiently. This finding demonstrates that besides sensory impairment or signal noise, which reduce cue availability and thus affect audio-visual cue reliance, having experienced a change in environmental conditions can modulate the perceiver's (otherwise relatively stable) general bias towards different modalities during speech communication.

The perceptual attunement to native vowel categories has been reported to occur at 6 months of age. However, some languages contrast vowels both in quality and in length, and whether and how the acquisition of spectral and duration-cued contrasts differs is uncertain. This study traced the development of infants' sensitivity to native (Czech) vowel-length and vowel-quality contrasts. The results suggest that in a vowel-length language, infants learn to categorize vowels in terms of length earlier and/or more robustly than in terms of quality, the representation of which may still be relatively underdeveloped at 10 months of age.

Previous research indicates that alternating between a bilingual's languages during speech production can lead to short-term increases in cross-language phonetic interaction. However, discrepancies exist between the reported L1-L2 effects in terms of direction and magnitude, and sometimes the effects are not found at all. The present study focused on L1 interference in L2, examining Voice Onset Time (VOT) of English voiceless stops produced by L1-dominant Czech-English bilinguals-interpreter trainees highly proficient in L2-English. We tested two hypotheses: (1) switching between languages induces an immediate increase in L1 interference during code-switching; and (2) due to global language co-activation, an increase in L1-to-L2 interference occurs when bilinguals interpret (translate) a message from L1 into L2 even if they do not produce L1 speech. Fourteen bilinguals uttered L2-English sentences under three conditions: L2-only, code-switching into L2, and interpreting into L2. Against expectation, the results showed that English VOT in the bilingual tasks tended to be longer and less Czech-like compared to the English-only task. This contradicts an earlier finding of L2 VOT converging temporarily towards L1 VOT values for comparable bilingual tasks performed by speakers from the same bilingual population. Participant-level inspection of our data suggests that besides language-background differences, individual language-switching strategies contribute to discrepancies between studies.

Whether late learners discern fine phonetic detail in second-language (L2) input, form new phonetic categories, and realize them accurately remains a relevant question in L2 phonology, especially for foreign-language (FL) learning characterized by limited exposure to interactional native input. Our study focuses on advanced Czech learners’ production of the L2 English vowels GOOSE and FOOT. While English /u/ and /ʊ/ have been undergoing fronting, their Czech equivalents, /uː/ and /u/, are fully back. We show that although the spectral differentiation of /u/-/ʊ/ is smaller in the learners’ than in native speech, the vowels being contrasted primarily in length, even FL learners can shift their L2 sound categories towards native-like targets, or in this case, produce English /u/-/ʊ/ as fronted.