Allophonic Variation Research Articles

Allophonic Variation in Romance Languages: Categorical and Gradient Processes

The distinction between categorical phonological alternations and gradual phonetic processes is an important one in the study of sound systems. However, the correct classification of specific phenomena as belonging to phonology or to phonetics is often difficult. In recent years some researchers have argued that many processes traditionally understood as phonological alternations actually belong to the realm of gradient phonetics. Here I focus on consonant lenition phenomena in Spanish and other Romance languages and examine the potential arguments for their traditional analysis as alternations between phonological categories.

A Case for Old Frisian Vowel Balance as Prosodically-driven Vowel Reduction

Abstract Drawing on data from the Old Frisian Riustring manuscripts, modern Frisian dialects, as well as older and more modern Scandinavian dialects and languages, this article argues for two points: first, that Vowel Balance is ultimately a reduction process, and second, that a foot-based analysis can account for Vowel Balance. In Old Frisian, i and u appear after light stems while e and o are found after heavy and polysyllabic stems; in some other languages and dialects, vowels after heavy and polysyllabic stems are reduced supporting an interpretation of e and o as reduced allophonic variants of i and u respectively. Two caveats can be made for arguments in favour of interpreting Vowel Balance as a reduction process, namely the case of trimoraic Vowel Balance in disyllabic words, and the reinterpretation of the contrast between i~e in Old Frisian as a means of marking the prosodic contrast between stem size. These caveats help clarify how prosody can shape stems both in terms of additional manifestations of Vowel Balance, and extending the vowel alternations based on light-heavy stem contrast to words which would not have otherwise undergone Vowel Balance.

Effects of exposure to dialect-specific allophonic variation on perceptual similarity ratings of vowels

Dialect exposure affects listeners’ lexical processing. In this study, we examined whether listeners’ exposure to dialect-specific allophonic variation in two linguistic variables in American English (nasal split /æ/ system and pre-voiceless /ɑɪ/ raising) affects their perceptual similarity ratings of matching versus mismatching allophones. Listeners from the U.S. North were assumed to have more exposure to pre-voiceless /ɑɪ/ raising and less exposure to a nasal split /æ/ system, while listeners from the U.S. Midland were assumed to have more exposure to a nasal split /æ/ system and less exposure to pre-voiceless /ɑɪ/ raising. We analyzed Northern and Midland listeners’ perceptual similarity ratings on a 1–5 scale for three types of auditory vowel pairs: match (e.g., ban-bang or bike-bite), allophonic mismatch (e.g., ban-bat or bike-bide), and phonemic mismatch (e.g., ban-bike). We predict that listeners with more exposure to dialect-specific allophonic variation should rate allophonic mismatches as more different than allophonic matches compared to listeners with less exposure. Preliminary results suggest that listeners rate allophonic matches as more similar than allophonic mismatches, and both of these types are rated as more similar than phonemic mismatches, consistent with existing literature. There was no effect of dialect exposure to allophonic variation on perceptual similarity ratings.

Using allophonic variation in L2 word recognition: French listeners’ processing of English vowel nasalization

During spoken word processing, native (L1) listeners use allophonic variation to predictively rule out word competitors and speed up word recognition. There is some evidence that second language (L2) learners develop an awareness of allophonic distributions in their L2, but whether they use their knowledge to facilitate word recognition online, like native listeners do, is largely unknown. In an offline gating experiment and an online eye-tracking experiment in the visual world paradigm, we compare advanced French learners of English and a control group of L1 English listeners on their processing of English vowel nasalization during spoken word recognition. In the gating task, the French listeners’ performance did not differ from that of the English ones. The eye-tracking results show that French listeners used the allophonic distribution in the same way as English listeners, although they were not as fast. Together, these results reveal that L2 learners can develop novel processing strategies using sounds in allophonic distribution to facilitate spoken word recognition.

Consonantal decomposition in interlanguage mapping*

AbstractThis paper analyzes segmental decomposition processes in interlanguage mapping, showing that the major place property is the most important factor among all phonological properties to be preserved. For instance, the duality of the Japanese fricative /ɸ/ triggers segmental decomposition both in Japanese loan adaptation of the English /f/ and in Korean adaptation of the Japanese fricative. Moreover, the targets of decomposition may include affricates, fricatives, and even a velar nasal. Furthermore, allophonic variation in the source language may show up in the recipient language as variations, even in transcription. Finally, I discuss the role of VOT which contributes to the interlanguage mapping.

A Contrastive Study of The Effect of Initial /s/ on the Voice Onset Time (VOT) of Following Stops in English, Kurdish and Kurdish EFL Learners

When an initial /s/ precedes a voiceless stop in English, the aspiration is lost, and the voice onset time (VOT) is significantly shortened. This causes an allophonic variation of English stops based on context. This paper intends to comparatively investigate this phenomenon in English, Kurdish and Kurdish EFL learners from two proficiency levels. (40) participants which are divided among four groups (Kurdish, English, learners G1 and learners G2) are asked to read lists of words including voiceless stops preceded by /s/, constructed in English and Kurdish. Each stop in the list has two examples followed by a high and a low vowel. All the participants are asked to repeat each word three times to provide three tokens. These productions are recorded and VOT values are acoustically measured. Results are then statistically analyzed to check the significance of variable interactions. Results have shown that Kurdish voiceless stops preceded by /s/ are more aspirated than those of English in similar contexts and thus produced with longer VOT values. This indicates that Kurdish learners need to modify their voiceless stop production when speaking English. Results also show that advanced Kurdish learners have improved at pronouncing the aspirated English stop variety, but they still produce it with a noticeably extended VOT.

The Influence of Native Phonology, Allophony, and Phonotactics on Nonnative Lexical Encoding: A Vocabulary Training Study

AbstractSecond language (L2) speakers often experience difficulties in learning words with L2‐specific phonemes due to the unfaithful lexical encoding predicted by the fuzzy lexical representations hypothesis. Currently, there is limited understanding of how allophonic variation in the first language (L1) influences L2 phonological and lexical encoding. We report how the Mandarin Chinese L1 phonemic inventory and allophonic variation subject to phonotactic constraints predict phonological encoding problems for novel L2 English words with the /v/–/w/ contrast. L1 English and L1 Chinese participants speaking two varieties of Mandarin Chinese differing as to the presence of [ʋ]–[w] allophonic variation for the /w/ phoneme participated in a vocabulary learning task. The novel L2 words with the /v/–/w/ contrast were systematically less robustly encoded than the control words on the day of training and 24 hours later. The degree of fuzziness in lexical representations was jointly predicted by L1 allophonic variation subject to phonotactic constraints and L2 phonological categorization.

An instrumental investigation of the vowel inventory of China English

ABSTRACT In today’s globalized world where English is increasingly used as a lingua franca , an increasing number of new English varieties are emerging in different parts of the world. China English (CE), as one of the fastest-growing varieties of English with the largest number of speakers, has been attracting a great deal of attention in recent years. This study aims to explore the vowel inventory of CE using an exploratory approach. Data were collected from 20 CE speakers who were recorded reading a standard elicitation list of words containing the 11 English monophthongs. Phonetic properties, phonemic contrasts, and allophonic variations of CE vowels were examined acoustically. Results provide some empirical evidence for the existence of a CE vowel system in which the vowel length contrasts are maintained, whereas the vowel quality contrasts tend to be neutralized to different extents, as compared to Inner Circle English varieties.

The effect of allophonic variability on L2 contrast perception: Evidence from perception of English vowels.

Current frameworks of L2 phonetic acquisition remain largely underspecified with respect to the role of L1 allophonic variability in acquisition. Examining the role of L1 allophonic variability, the current study compared the perceptual discrimination of English /i-ɪ/ and /ɛ-æ/ by L1 Korean and L1 Mandarin speakers. Korean and Mandarin vowel inventories differ in that Mandarin employs significantly greater allophonic variation of the mid-region /E/ vowel. Results demonstrated worse perceptual accuracy by L1 Mandarin speakers for the /ɛ-æ/ contrast than L1 Korean speakers. These results suggest that both L1 phonemic inventories and allophonic variation play a role in L2 phonetic acquisition.

Ingo’s Register Treadmill

Ingo's Register Treadmill Ingo Titze (bio) VOICE RESEARCH AND TECHNOLOGY A register treadmill was proposed eight years ago as a diagnostic exercise for exposing involuntary register shifts in the D4 to A4 pitch range.1 The basic concept was to create pitch and vowel progressions that would elicit maximum crossings of harmonics with supraglottal resonances (formants) of the vocal tract. It is well known that sudden increases or decreases in harmonic amplitudes in proximity of a resonance can cause a timbral change that can be considered an acoustic registration.2 In particular, strengthening or weakening the second harmonic with the first formant can be responsible for such a register shift. A pentavowel sequence, combined with a pentatonic scale, was proposed such that source frequencies move in opposite direction to resonance frequencies. This produces multiple harmonic-resonance crossovers. This article is a follow-up to expand the treadmill to include a wider pitch range and sex differences in formant frequencies. While allophonic variations of formants F1 and F2 are large across individuals and languages, a smoothed penta-vowel sequence has been chosen here. Each sequence begins with a wide mouth shape and ends with a narrow mouth shape. Formant frequencies for females have been chosen to be 20 % higher than for males. All formant frequencies are listed below for two vowel sequences. High Vowel Sequence [æ] [ɛ] [e] [ɪ] [i] Male F1 670 550 420 350 280 Female F1 860 660 504 420 320 Male F2 1800 1900 2000 2200 2300 Female F2 2160 2280 2400 2640 2760 Low Vowel Sequence [a] [Ɔ] [o] [Ʊ] [u] Male F1 850 500 360 300 300 Female F1 1020 600 431 360 370 Male F2 1610 800 640 610 595 Female F2 1932 960 768 732 714 In the treadmill, F1 is descending while fo is ascending to produce a maximum number of harmonic/formant crossings. Figure 1 shows the graphic display for a D4–A4 up and down pentatonic scale (solid lines), combined with the low vowel formant frequencies (dashed lines). Note that there are [End Page 197] Click for larger view View full resolution Figure 1. Treadmill for the D4–A4 pentatonic scale produced on low vowels; (a) male sequence and (b) female sequence. five crossings on the ascending scale for both males and females. For males, the crossings are: fo crosses F1 near the [o] vowel; 2fo crosses F1 near the [Ɔ] vowel 2fo crosses F2 near the [o] vowel 3fo crosses F2 near the [Ɔ] vowel 4fo crosses F2 near the [Ɔ] vowel. Note the importance of the [Ɔ] and [o] vowels in males in this region of the pitch range. For females, the same five crossings occur, but on slightly more closed-mouth vowels due to higher overall formant frequencies. For example, fo crosses F1 closer to the [Ʊ] vowel than the [o] vowel. Figure 2 shows the treadmill for the high vowel sequence. The pentatonic scale is kept the same, D4–A4. There are now only two major crossings: fo crosses F1 between [e] and [I] in males and between [I] and [i] in females; 2fo crosses F2 between [æ] and [ɛ] in males and between [ɛ] and [e] in females None of the first 4 harmonics cross the second formant F2. Thus, in this region of the pitch range (D4–A4), acoustic registration in high vowels is carried primarily by the first two harmonics. The fifth and higher harmonics are not plotted. They will interact with F2 and higher formants, particularly if fo is raised. Figures 3 and 4 show the treadmills when the pentatonic scale is raised to the G4–D5 range. Usually one progresses in half-steps to this higher level. Most of the high pitch range for males and female belters is then covered. Figure 3 shows the low vowel sequence. The bright [a] vowel is now included in the crossings, but the general picture has not changed much from the A4–D4 scale. Of special...

The acquisition of L2 allophonic variants: The role of phonological distribution and lexical cues

Adult learners acquire second language (L2) allophones with experience. We examine two mechanisms which may support the acquisition of allophonic variants in second language acquisition. One of the mechanisms is based on the distribution of phones with respect to their phonological context (i.e. phonological distribution). The other is based on the role the phones play in contrasting words (i.e. lexical contrast). Experiment 1 established adult native English speakers’ baseline sensitivity to the novel [b]–[β] auditory contrast. In Experiment 2 we examined whether adult native English speakers infer the phonological status of [b] and [β] in an artificial language based only on their distributions to phonological contexts. We observed no evidence that these participants were able to do so. In Experiment 3 we investigated whether learners infer the phonological status of [b] and [β] from the role they play in lexical contrast and observed both perceptual and lexical processing evidence to suggest that adults may use meaning-based cues to the lack of contrast to learn that two phones are allophones of the same phoneme. Together our findings suggest that adult L2 learners may prioritize information about function (in this case, lexical contrast) over the phonological distribution of phones as they determine the phonological status of L2 sounds.

Rethinking Gender Principles in El Salvador: Evidence from /s/ Weakening

Abstract Previous research on /s/ weakening in Spanish has consistently aligned with Labovian principles: women prefer the prestige variant, usually [s], while men favor nonstandard, lenited variants. However, in Salvadoran Spanish—a dialect that weakens /s/ across syllable positions and shows allophonic variation beyond the tripartite paradigm of [s]/[h]/[∅]—gender-based lenition patterns contradict this generalization. This study examines the production of phonological /s/ by 72 Salvadorans balanced for region, urbanicity, age, and gender who participated in sociolinguistic interviews in El Salvador in 2015. We find that women not only lenite /s/ at higher rates than men overall, but also produce significantly more of the variants that carry the most local stigma. We further find that, counterintuitively, women are significantly more likely than men to lenite /s/ in utterance- and word-initial prosodic positions, which are stronger and more perceptually salient than medial and final tokens. We argue that these discrepancies are best understood by taking El Salvador’s unique historical and sociopolitical context into account. Specifically, we propose that a culture of state-sanctioned violence against women and the unprecedented threat of gangs in El Salvador have led to the social segregation and linguistic isolation of women, affording them little access to standard linguistic forms even as globalization, urbanization, industrialization, and migration facilitate a shift toward linguistic standardization.

Allophonic familiarity differentiates word representations in the brain of native speakers of regional linguistic varieties

This study aims to shed light on the issue whether familiar allophonic variation is encoded in word representations. Both Italian speakers born in Trentino and speakers born in the Central-Southern regions of Italy took part in the experiment. We tested the MMN elicited by the same word encompassing two different allophones, one of which was more familiar to one group of participants than to the other, depending on their regional variety of Italian. The Trentino group showed an enhanced MMN for the word embedding the familiar variant while Central-Southern speakers showed no difference. The amplitude of the MMN for the unfamiliar word variant in Trentino speakers showed an inverse correlation with the passive exposure to the Trentino dialect. We conclude that words embedding familiar and unfamiliar allophones are differently represented in the brain of native speakers of regional language and the degree of differentiation is modulated by individual experience.

Free Allophonic Variation in Native and Second Language Spoken Word Recognition: The Case of the German Rhotic.

The question of how listeners deal with different phonetic variant forms for the same words in perception has sparked great interest over the past few decades, especially with regard to lenited and regional forms. However, the perception of free variant forms of allophones within the same syllable position remains surprisingly understudied. Because of this, in the present study, we investigate how free allophonic variation in the realization of the German rhotic (/r/) impacts spoken word recognition for native German listeners and two groups of non-native listeners (French and Italian learners of German). By means of a visual-world eye-tracking task, we tested the recognition of spoken German words starting with /r/ when the rhotic was produced either as the more canonical variant, the uvular fricative [] which is considered the German standard, or as an alveolar trill [r], a common realization in the south of Germany. Results showed that German listeners were more efficient at recognizing /r/-initial words when these were produced with the uvular fricative than with the alveolar trill. French listeners did not differ from German listeners in that respect, but Italian listeners showed exactly the opposite pattern: they showed an advantage when words were produced with the alveolar trill. These findings suggest that, for native listeners, the canonicity of the variant form is an important determiner of ease of recognition, even in the absence of orthographic or perceptual motivations for the primacy of canonical variants for this particular example of variation. For non-native listeners, by contrast, results are better explained by the match of the different allophones to the canonical realization of /r/ in their native language than by the status or frequency of the allophones in the non-native language itself.

Effects of Device Limitations on Acquisition of the /t/-/k/ Contrast in Children With Cochlear Implants.

The present study investigated how development of the /t/-/k/ contrast is affected by the unique perceptual constraints imposed on young children using cochlear implants (CIs). We hypothesized that children with CIs would demonstrate unique patterns of speech acquisition due to device limitations, rather than straightforward delays due to a lack of auditory input in the first year of life before implantation. This study focused on the contrast between /t/ and /k/ because it is acquired early in the sequence of development, requires less advanced motor control than later-acquired place contrasts, is differentiated by spectral cues (which are particularly degraded when processed by CIs), and is not easily differentiated by visual cues alone. Furthermore, perceptual confusability between /t/ and /k/ may be exacerbated in front-vowel contexts, where the spectral energy for /k/ is shifted to higher frequencies, creating more spectral overlap with /t/. Children with CIs (n = 26; ages 31 to 66 mo) who received implants around their first birthdays were matched to peers with normal hearing (NH). Children participated in a picture-prompted auditory word-repetition task that included over 30 tokens of word-initial /t/ and /k/ consonants. Tokens were balanced across front-vowel and back-vowel contexts to assess the effects of coarticulation. Productions were transcribed and coded for accuracy as well as the types of errors produced (manner of articulation, voicing, or place of articulation errors). Centroid frequency was also calculated for /t/ and /k/ tokens that were produced correctly. Mixed-effects models were used to compare accuracy, types of errors, and centroid frequencies across groups, target consonants, and vowel contexts. Children with CIs produced /t/ and /k/ less accurately than their peers in both front- and back-vowel contexts. Children with CIs produced /t/ and /k/ with equal accuracy, and /k/ was produced less accurately in front-vowel contexts than in back-vowel contexts. When they produced errors, children with CIs were more likely to produce manner errors and less likely to produce voicing errors than children with NH. Centroid frequencies for /t/ and /k/ were similar across groups, except for /k/ in front-vowel contexts: children with NH produced /k/ in front-vowel contexts with higher centroid frequency than children with CIs, and they produced /k/ and /t/ with equal centroid frequencies in front-vowel contexts. Children with CIs not only produced /t/ and /k/ less accurately than peers with NH, they also demonstrated idiosyncratic patterns of acquisition, likely resulting from receiving degraded and distorted spectral information critical for differentiating /t/ and /k/. Speech-language pathologists should consider perceptual confusability of consonants (and their allophonic variations) during their assessment and treatment of this unique population of children.

On the Difficulty of Defining “Difficult” in Second-Language Vowel Acquisition

Hierarchies of difficulty in second-language (L2) phonology have long played a role in the postulation and evaluation of learning models. In L2 pronunciation teaching, hierarchies are assumed to be helpful in the development of instructional strategies based on anticipated areas of difficulty. This investigation addressed the practicality of defining a pedagogically useful hierarchy of difficulty for English tense and lax close vowels (/i I u ʊ/) produced by Cantonese speakers. Unlike their English counterparts, Cantonese close tense-lax pairs are allophonic variants with [i u] occurring before alveolars and [I ʊ] before velars. Each tense-lax pair represents a “phonemic split” in which members of a single L1 category are realized contrastively in L2. Despite evidence that English tense-lax distinctions are challenging for Cantonese speakers, no previous empirical work has closely considered the problem from the standpoint of vowel intelligibility across multiple phonetic contexts and in different words sharing the same rhyme. In a picture-based word-elicitation task, 18 Cantonese-speaking participants produced 31 high-frequency CV and CVC words. Vowels were evaluated for intelligibility by phonetically-trained judges. A series of mixed-effects binary logistic models were fitted to the scores, with vowel quality, phonetic context (rhyme) and word as factors, and length of Canadian residence and daily use of English as co-variates. As expected, the general hierarchy of difficulty for vowels that emerged (/i/ > /u/ > /ʊ/ > /I/) was complicated by large differences across phonetic contexts. Results were not readily explicable in terms of transfer; moreover, different words with the same rhyme were not produced with equal intelligibility. The most serious modeling complication was the sizeable inter-speaker variability in difficulties, which could not be accounted for by model co-variates. Although some difficulties were roughly systematic at the group level, it is argued that establishing a pedagogically useful hierarchy on such data would prove intractable. Rather, L2 learners might be better served by assessment and instructional targeting of their individual problem areas than by a focus on errors predicted from hierarchies of difficulty.

The First Language Interference toward Students’ English Speaking as Foreign Language

This research aims at analyzing the first language interference toward students’ English speaking as foreign language made by sixth semester students in course design subject, such as phonological interference (pronunciation), grammatical (morphological and lexical), and lexical interference (vocabulary), and the factors that caused interference of the first language. This research is descriptive qualitative. The findings of the research explained that students made three types of interference categorized as phonological interference such as pronounce the word incorrectly (vowel, diphthong, consonan, and allophonic variation), grammatical interference such as in morphological (singular-plural agreement) and syntactical (subject-verb agreement, phrase, comparative adjective, possessive adjective, and parallel structure), and lexical interference such as in vocabulary (combining between Indonesian and the English language). The data showed that students often make error in phonological aspect. The factors that caused first language interference are lack of knowledge, Indonesian transfer, and lack of vocabularies of foreign language that mastered by students or respondents.

Evaluation of aspiration problems in L2 English pronunciation employing machine learning

The approach proposed in this study includes methods specifically dedicated to the detection of allophonic variation in English. This study aims to find an efficient method for automatic evaluation of aspiration in the case of Polish second-language (L2) English speakers' pronunciation when whole words are analyzed instead of particular allophones extracted from words. Sample words including aspirated and unaspirated allophones were prepared by experts in English phonetics and phonology. The datasets created include recordings of words pronounced by nine native English speakers of standard southern British accent and 20 Polish L2 English users. Complete unedited words are treated as input data for feature extraction and classification algorithms such as k-nearest neighbors, naive Bayes method, long-short term memory, and convolutional neural network (CNN). Various signal representations, including low-level audio features, the so-called mid-term and feature trajectory, and spectrograms, are tested in the context of their usability for the detection of aspiration. The results obtained show high potential for an automated evaluation of pronunciation focused on a particular phonological feature (aspiration) when classifiers analyze whole words. Additionally, CNN returns satisfying results for the automated classification of words containing aspirated and unaspirated allophones produced by Polish L2 speakers.

Phonetics and Phonology of Stop Consonants in the Guaymi Spoken in Costa Rica


 
 
 The present work consists of an analysis of the phonetic and phonological properties of the sequences /t-k/ and /b-d-ɡ/ from the Guaymi spoken in La Casona, Puntarenas, Costa Rica. The corpus was analyzed with Speech Analyzer and the Lisker and Abramson (1964) method was applied in order to obtain the values of the V.O.T. The results suggest that the sequence /t-k/ does not present systematic allophonic variations, while /b-d-g/ alternates with stop, fricative, and approximant allophones; neutralization among stop phonemes is not frequent. Lastly, the V.O.T. contributes to the differentiation between voiceless and voiced stops.
 
 

Die frühe voraltsächsische und voraltfriesische Runenüberlieferung

Abstract This paper deals with two runic inscriptions that are highly relevant to language history. 1. The runic sequences on the three Weser rune bones, which date to the first half of the 5th century, are not entirely clear. However, West Germanic (Pre-Old Saxon) linguistic features such as gemination by j (kunni ‘kin, clan’) and loss of -a < *-az (hari ‘army’) are obvious. By far the most interesting linguistic form is the preterite dede ‘did’ that reflects PGmc. *-dai (cf. PNorse talgi-dai ‘carved’ on the Nøvling fibula) as opposed to PGmc. *-dǣ(d) (PNorse -da, OHG -ta etc.). Apparently, we are dealing with two distinct endings, *-dai deriving from an PIE middle in *-(t)ói̯. 2. The legend ska 2 nomodu (a 2 = ᚪ) on a solidus of unknown provenance (ca. 600) renders the dithematic anthroponym Skānɔmōdə̣ (or *Skānɵmōdə̣), presumably the name of the moneyer. It seems that medial o stands for [ɔ] or [ɵ], an allophonic variant of the linking element /a/ before a labial consonant; parallels can be found in Old Germanic naming. Two linguistic features, viz. ā < WGmc. *au and the nominative ending -ə̣ < WGmc. -a < PGmc. *-az, indicate that the language of the inscription is Pre-Old Frisian.