Phonetic Distance Research Articles

The Case for a Quantitative Approach to the Study of Nonnative Accent Features.

Research with nonnative speech spans many different linguistic branches and topics. Most studies include one or a few well-known features of a particular accent. However, due to a lack of empirical studies, little is known about how common these features are among nonnative speakers or how uncommon they are among native speakers. Moreover, it remains to be seen whether findings from such studies generalize to lesser-known features. Here, we demonstrate a quantitative approach to study nonnative accent features using Dutch-accented English as an example. By analyzing the phonetic distances between transcriptions of speech samples, this approach can identify the features that best distinguish nonnative from native speech. In addition, we describe a method to test hypotheses about accent features by checking whether the prevalence of the features overall varies between native and nonnative speakers. Furthermore, we include English speakers from the United States and United Kingdom and native Dutch speakers from Belgium and The Netherlands to address the issue of regional accent variability in both the native and target language. We discuss the results concerning three observed features. Overall, the results provide empirical support for some well-known features of Dutch-accented English, but suggest that others may be infrequent among nonnatives or in fact frequent among natives. In addition, the findings reveal potentially new accent features, and factors that may modulate the expression of known features. Our study demonstrates a fruitful approach to study nonnative accent features that has the potential to expand our understanding of the phenomenon of accent.

Revisiting the automatic prediction of lexical errors in Mandarin

Abstract Speech errors provide cues for explaining the process of word retrieval. For example, speech errors are less likely to occur with high-frequency words since these words already receive a high level of activation. The current analysis further develops existing findings in two ways. First, instead of considering the overall frequency of the words in the entire corpora, we consider the gap in frequency between sequential pairs of words. We hypothesize that speech errors are more likely to occur if the target has a much lower frequency than its preceding word. Second, we use word embedding methods to quantify the semantic distance between sequential pairs of words. We hypothesize that speech errors are more likely to occur with words that have a large semantic distance from their preceding context. We also consider the potential effects of phonetic distance between sequential pairs of words and position-in-utterance of words in utterances. The results from a Mandarin corpus of speech errors show that word frequency and semantic distance between sequential pairs of words can be used to predict the occurrence of speech errors with an accuracy above the majority baseline.

"They sure aren't from around here": Children's perception of accent distance in L1 and L2 varieties of English.

Children exhibit preferences for familiar accents early in life. However, they frequently have more difficulty distinguishing between first language (L1) accents than second language (L2) accents in categorization tasks. Few studies have addressed children's perception of accent strength, or the relation between accent strength and objective measures of pronunciation distance. To address these gaps, 6- and 12-year-olds and adults ranked talkers' perceived distance from the local accent (i.e., Midland American English). Rankings were compared with objective distance measures. Acoustic and phonetic distance measures were significant predictors of ladder rankings, but there was no evidence that children and adults significantly differed in their sensitivity to accent strength. Levenshtein Distance, a phonetic distance metric, was the strongest predictor of perceptual rankings for both children and adults. As a percept, accent strength has critical implications for social judgments, which determine real world social outcomes for talkers with non-local accents.

Classification of Chubu-region dialects using random forest

The present study investigated how Japanese dialects in Chubu region are classified using random forest (RF) and an adaptive synthetic (ADASYN) sampling approach. We obtained the written-format pronunciations of 22 words from 884 Japanese speakers (average age = 71.1) who had been residing in their birthplaces. We calculated phonetic distance (ALINE distance) between the dialectal and standard Japanese pronunciations, and ran RF models with 1,000 bootstrap samples. The results of the RF models (ROC = 0.953, F1 = 0.68) demonstrated that speakers with a predicted probability greater than 50 percent (n = 415) were generally located in their residing prefectures, suggesting that each prefecture has its own dialect. Speakers with a predicted probability less than 50 percent were located in both their residing prefectures and the vicinities of the prefectures, particularly those in Aichi, Gifu, Shizuoka, Nagano, Gunma, and Niigata prefectures, suggesting that the speakers in the areas share some characteristics of the prefectural dialects. This potentially led to the poor predictions of the speakers. The expansion of these prefectural dialects seems to be limited by the Japanese Alps; dialects spoken on the eastern and western side of the mountain range are generally widespread only in the area, respectively.

Application of advanced language technologies in analysis of category naming fluency task in healthy participants

AbstractBackgroundCategory naming fluency tasks have been used frequently to measure patients’ working memory and executive functions. Previous studies show that differences in scores in this task could reflect features of neurodegenerative disease. To better understand decreased performance in patients, it is critical to establish performance in healthy aging populations. Here, we investigated differences in total scores and lexical/speech features of words produced during animal fluency tasks of healthy young and older speakers which can only be derived with novel, automated, computational technologies.MethodData were collected in 52 young participants (26.9% males, mean age = 23.5±3.17) and 46 elderly participants (63.0% males, mean age = 69.9±6.77). Participants were instructed to list as many animals as possible for 60 seconds. Speech was transcribed and automatically force‐aligned. Transcripts were run through our automated Python pipeline to calculate the total number of animals listed and rate each word for the following semantic parameters: age of acquisition (AoA), ambiguity, familiarity, frequency. We also calculated phonetic and semantic distances between each word and its antecedent. Linear regression models tested the association between these automated language measures and total scores. We also examined the differences of the young and elderly groups using t‐tests.ResultElderly participants produced significantly fewer animals than young participants (p<0.001). Higher overall scores were associated with higher mean AoA (beta = 0.22, p = 0.049) and longer word duration (beta = 1.62, p = 0.037), as well as less frequent (beta = ‐0.51, p = 0.021) and familiar (beta = ‐3.96, p<0.001) animal names. We found no interaction with the age groups in any of the models. Elderly participants produced less frequent animal names than young participants (p = 0.029), along with significantly reduced phonetic distance scores (p = 0.008). Elderly participants scored significantly higher articulation rates (p = 0.007) and had a faster speech rate (p<0.001), attributable to producing more non‐animal words.ConclusionThis study demonstrates the applicability of advanced language technologies in analyzing category naming fluency tasks in healthy adults. We developed an automated scoring system and identify novel semantic and phonetic measures from digitally recorded fluency tasks. We also quantified some confounding age effects which could be explained by lexical/speech differences of words produced by the groups.

Bairisch verstehen – auf Hochdeutsch reagieren. Strategien zum Aufbau einer rezeptiven Varietätenkompetenz für (tschechische) DaF-Lerner

Based on previous studies discussing the thematization of German dialects in lessons of German as a foreign language within and outside the German speaking countries, the article outlines a didactic concept for building receptive variety competence of Czech learners of German towards the diatopic varieties of German spoken in the Free State of Bavaria. Using the example of the North and Middle Bavarian dialects, it will be demonstrated how the Bavarian-Austrian loan vocabulary present in the Czech non-standard varieties can be used didactically to enable learners to overcome the phonetic distance between dialect and standard language, which is the main barrier to comprehension. In this respect, the learners’ inner multilingualism is used to differentiate their outer multilingualism. The current foreign German studies background to the concept is the establishment of interdisciplinary Bavarian Studies at the University of West Bohemia in Pilsen (2019) whose curriculum provides a more profound receptive study of the dialects of Bavaria.

Comparing Germanic, Romance and Slavic: Relationships among linguistic distances

Languages differ along multiple dimensions (lexis, phonology, morphology, syntax). Related languages descend from a common ancestor language but have diverged over time. This paper asks whether languages diverge equally along all dimensions, and, to the extent that they do not, which dimension reflects the traditional language family tree best. We computed measures of (i) lexical distance (ii) phonetic distance, and (iii) syntactic distance. The measures were computed on all words and sentences extracted from a corpus of translations of four relatively short English texts into another four Germanic languages (Danish, Dutch, German, Swedish), five Romance languages (French, Italian, Portuguese, Romanian, Spanish) and six Slavic languages (Bulgarian, Croatian, Czech, Polish, Slovakian, Slovenian). We examined the correlation structure of the distances for all pairs of Germanic (10), Romance (10) and Slavic (15) languages (i.e., within-family comparisons only). The results indicate that the linguistic dimensions are generally correlated (weakly but significantly), and that the correlations are stronger for pairs within families than when all 35 pairs are examined together. Cladistic family trees correlate best with the lexical distance (0.851 < r < 0.887). This confirms that the genealogical language trees are predominantly based on lexical rather than phonetic or syntactic considerations.

Assessing speech intelligibility of pathological speech in sentences and word lists: The contribution of phoneme-level measures.

Speech intelligibility is an important indicator of the degree of speech impairment in pathological speech. Articulation, as a key feature of dysarthria, has been found to be a stronger contributor to intelligibility of dysarthric speech compared to voice quality, nasality, and prosody. In fact, therapy addressing articulation is often used by speech-language pathologists. Since phoneme-level measures are more directly related to articulation, they may contribute to better evaluating articulation imprecision in speakers with dysarthria and to monitoring the effectiveness of therapy. We collected two types of phoneme-level measures: a) Accuracy of Phonemes, the percentage of correctly transcribed phonemes, and b) Phonetic Distance, from orthographic transcriptions obtained from expert raters in two types of speech materials (i.e., meaningful sentences and word lists). We first examined the measures' interrater reliability using Generalizability Theory. Then we studied the validity of the measures by correlating them to three criterion variables. Following this, we explored their ability in distinguishing speakers in two classification tasks according to speakers' types (i.e., healthy vs dysarthric) and their severity levels of dysarthria, respectively. The results showed that both types of phoneme-level measures are highly reliable and valid in two different speech materials. They also showed acceptable results for both classification tasks in different speech materials, with word lists performing better than meaningful sentences. The differences between the two speech materials may be largely caused by differences in word structures and contextual cues in the materials. The results indicate that both types of phoneme-level measures show largely similar reliability and validity in both speech materials. These measures perform better in word lists than in meaningful sentences, suggesting an advantage for using word lists in clinical practice and research. On the other hand, meaningful sentences can be used for classifying healthy and dysarthric speakers. Our results suggest that using different speech materials gives a better overview of the speakers' intelligibility at the segmental level and the implications of their articulation impairments.

Classification of Kansai and Chugoku dialects spoken by old Japanese speakers

The present study examined how Kansai and Chugoku dialects in Japanese are classified using hierarchical density-based spatial clustering (HDBSCAN) and Random Forest (RF). We obtained the written-format pronunciations of 48 words from 1450 Japanese speakers over age 50 who had been residing in their birthplaces. We calculated phonetic distance (ALINE distance) between the dialectal pronunciations and standard Japanese ones, and ran HDBSCAN and RF models with 1000 bootstrap samples. The optimal HDBSCAN model demonstrated that there are two groups of speakers in the northern and southern pastoral areas of Kansai region. The RF models demonstrated that speakers from each prefecture were classified with a wide range of accuracies (F1 = 0.73). Kansai-region speakers in the urban area where Osaka, Kyoto, and Nara prefectures share borderlines were poorly classified. Similarly, Chugoku-region speakers living near the borderlines of Hiroshima, Shimane, and Tottori prefectures were poorly classified. The rest of the participants were generally classified well. These results suggest that each prefecture generally has its own dialect, but its distribution goes beyond prefecture borderlines. This is the first study to reveal such classification patterns of Japanese dialects using machine learning approaches.

Unified Parametrization of Phonetic Features and Numerical Calculation of Phonetic Distances between Speech Sounds

Unified Parametrization of Phonetic Features and Numerical Calculation of Phonetic Distances between Speech Sounds

Unified Parametrization of Phonetic Features and Numerical Calculation of Phonetic Distances between Speech Sounds

ABSTRACT A metric method to numerically measure phonetic and phonemic distances or contrasts, between speech sounds, is put forward. The feature values of the compared phones taken from the standard IPA charts are treated as independent parameters that give rise to corresponding Euclidean distances. As an illustration, the general phone set is mapped to Ukrainian phonemes. The proposed model agrees well with the historical linguistic facts and experimental phonetic data. The described approach may find its due applications in various fields of linguistics and speech technologies, including historical and typological linguistics, language acquisition, phonetic studies, computational phonology, machine translation, information retrieval, and text-to-speech conversion.

Effects of native language and habituation in phonetic accommodation

Effects of vocal accommodation have been reported in a wide range of contexts, but they have typically been small. The absence of effects in some cases has proven perplexing. In the present investigation I present innovative methods for the representation of phonetic distance between phonetic tokens and the analysis of phonetic accommodation. I take a broad crosslinguistic perspective and report effects of linguistic background (L1) on patterns of phonetic convergence toward typical monolingual English voiceless stop voice-onset-times (VOTs).I propose that patterns of accommodation in laryngeal-oral coordination, as instantiated by voiceless stop VOTs, will reflect general principles of motor coordination (preferences for stable/in-phase coordination, cf. Browman & Goldstein, 1988; Haken, Kelso, & Bunz, 1985). Thus, stable, near-zero VOTs (cf. Spanish) will be less likely to show convergence toward intermediate English VOT, whereas less stable, long VOTs (cf. Korean) will be more likely to converge. Monolingual English and bilingual (Korean-English, Spanish-English) participants completed word shadowing and reading tasks. Their vowel-normalized voiceless stop VOTs were submitted to two analyses which confirm the articulatory stability hypothesis and reveal group-specific changes in vocal accommodation over time. The first involves a general baseline-to-test comparison, while in the second, a trial-specific difference from baseline is used as a dependent measure.The results offer new insights into the effects of language background on vocal accommodation, and the analytical approach offers a means to more cleanly isolate subtle effects of accommodation in speech among a multitude of competing factors.

Learning to Perceive Non-Native Tones via Distributional Training: Effects of Task and Acoustic Cue Weighting.

As many distributional learning (DL) studies have shown, adult listeners can achieve discrimination of a difficult non-native contrast after a short repetitive exposure to tokens falling at the extremes of that contrast. Such studies have shown using behavioural methods that a short distributional training can induce perceptual learning of vowel and consonant contrasts. However, much less is known about the neurological correlates of DL, and few studies have examined non-native lexical tone contrasts. Here, Australian-English speakers underwent DL training on a Mandarin tone contrast using behavioural (discrimination, identification) and neural (oddball-EEG) tasks, with listeners hearing either a bimodal or a unimodal distribution. Behavioural results show that listeners learned to discriminate tones after both unimodal and bimodal training; while EEG responses revealed more learning for listeners exposed to the bimodal distribution. Thus, perceptual learning through exposure to brief sound distributions (a) extends to non-native tonal contrasts, and (b) is sensitive to task, phonetic distance, and acoustic cue-weighting. Our findings have implications for models of how auditory and phonetic constraints influence speech learning.

Auditory predictions are phonological when phonetic information is variable

ABSTRACT An emerging body of studies has claimed that exposure to phonetically varying speech sounds results in strictly phonological representations of speech sounds by the brain’s auditory prediction system (Cornell et al., 2011; Eulitz & Lahiri, 2004; Hestvik et al., 2020; Hestvik & Durvasula, 2016; Phillips et al., 2000). We test this claim by measuring mismatch negativity (MMN) to a subcategorical contrast with two sets of phonetically varying standards. When controlling for non-prediction related contributors to the MMN, we find a mismatch in a late time window. However, the mismatch effect is only significant for participants who perceived the contrast as an across-category contrast (/t/ vs /d/). Additionally, we find no modulation of mismatch amplitude predicated on phonetic distance between standards and deviant. Taken together, this indicates that the prediction generated by the auditory system in response to phonetically varying speech sounds is indeed phonological and lacks any fine-grained phonetic content.

Arab EFL Learner Perceptions of English Phonemes: A Cross-Language Phonetic Interference

The extant studies addressing second language phonetic perception assume that second language phonemes are perceived to be similar to first language phonemes, and tend to be substituted by learners of English as a foreign language. This study aimed to assess the perceptual relationship between the phonemes of English and the sound units of Omani Arabic of the Al-Dakhiliyah region to discover the extent of the influence exerted by Omani Arabic on the perception of English phonemes by Omani learners. Two instruments were constructed to achieve this goal: first, the Omani phonetic system was elucidated via a questionnaire and the consonants and vowels uttered by the Omani people were archived as a phonetic dialect; second, a test was created to examine the perceptual phonetic distance between Omani Arabic and English phonemes. This investigation was designed to test the hypothesis of the revised speech learning model. The results confirmed the model&amp;rsquo;s postulation that learners instinctively and reflexively associate the second language sounds to the phonetic groups of their first language.

The language variation spectrum in Oldenburg from production and perception perspectives

This contribution deals with the vertical spectrum in Oldenburg. Using different methods (phonetic distance measurement, analysis of linguistic variables, cluster analysis, listener judgement tests), four speakers from three generations are analyzed. The results prove a vertical spectrum structured in three varieties (standard, regiolect, dialect), strong intergenerational differences and a large phonetic difference between dialect and regiolect.

Zur Stabilität dialektaler Formen.

Stability is an essential but frequently underrepresented fact in the study of language variation and change. Although in dialectology the distinction between “primary” and “secondary” dialect features initially led to a certain consideration of stable features as opposed to more dynamic ones, more recent empirical studies have not taken any notice of the former. The question of stability is nevertheless crucial for dialectology. For example, the extent of dialect loss cannot be assessed without establishing which forms are stable. The aim of this study is therefore to address this desideratum by investigating ten dialectal features in Austria’s rural dialects. All ten examples concern variables of the vowel system which show large phonetic differences between dialect and standard variants. On the basis of a comprehensive dialect survey, both a real- and an apparent-time study is conducted to examine the extent and the factors of dialectal stability and change. The study provides evidence for a surprisingly high degree of stability and continuity in Austria’s rural dialects. Changes are not only restricted to a few variables, but also limited (a) to a lexical factor – only a small group of lexemes is not completely stable –, and (b) to a regional factor – only some eastern parts of Austria show a higher degree of change. Other factors, such as the extent of the spatial distribution of a form or the phonetic distance to the standard, do not prove to be relevant for stability and change of the variables.

The Leipzig-Jakarta list as a means to test Old English / Old Norse mutual intelligibility

Abstract The use of basic word lists has long been common in the fields of second language acquisition and language typology. The application to the study of mutual intelligibility between closely related languages on the other hand has never gained much traction. This article will analyse the degree of mutual intelligibility between the vocabularies of Old English (Anglian) and Old Norse (Old Icelandic) with the use of the Leipzig-Jakarta List which ranks vocabulary by their resistance to borrowing. The entries were transliterated to the International Phonetic Alphabet and truncated so that only the word-roots remained. The entries were then compared using a rule-set based on phonetic deviations, the so-called Levenshtein Distance and a method derived from it called ALINE. The study finds a relatively low phonetic distance between the lists and concludes that they are overall close enough to be mutually intelligible.

Mapping the dimensions of linguistic distance: A study on South Ethiosemitic languages

We measured the distance among selected South Ethiosemitic languages from three dimensions: structural, functional and perceptual. The main objectives of our study was to determine the relationship among these three dimensions of linguistic distances, to re-examine previous classifications of the languages, and to measure the degree of mutual intelligibility among the language varieties. We determined the structural distance by computing the lexical and phonetic differences. The phonetic distance was computed using the Levenshtein algorithm. A Word Categorization test was adopted from Tang and van Heuven (2009) to measure the functional distance or the degree of intelligibility. A self-rating test, based on the recordings of ‘The North Wind and the Sun’ was administered to measure the perceptual distance among the languages. Then we performed a cluster analysis using Gabmap. Multidimensional scaling was employed for the cluster validation. The results of our study show that the selected South Ethiosemitic languages can be classified into five groups: {Chaha, Ezha, Gura, Gumer}, {Endegagn, Inor}, {Muher, Mesqan}, {Kistane} and {Silt’e}. Moreover, the language taxonomies obtained from the measures of the three dimensions of distance are very similar, and they are generally comparable to the classifications previously proposed by historical linguists. There is also a very strong correlation among the three dimensions of distance. Furthermore, the Word Categorization test results show that the majority of these languages are mutually intelligible with the exception of Silt’e.

Towards Modeling Second Dialect Speech Learning: The Production of Bogota [s] in Ciudad Bolivar by Speakers of Three Different Varieties of Colombian Spanish

This study investigates the second dialect production of Bogota Spanish /s/ in coda position by speakers of three different varieties of Colombian Spanish, who have been in contact in Ciudad Bolivar, a community located in Bogota, Colombia. The study has three aims. First, it will examine the role of phonetic distance in the acquisition of /s/ production. Second, it will determine the linguistic factors that constrain the realization of /s/ sound by the speakers of the three varieties studied. Third, it will look into the role of extralinguistic factors in the production of /s/. A total of 2322 tokens extracted from sociolinguistic interviews with 50 participants were acoustically analyzed in PRAAT. Statistical analyses were conducted using GoldVarb. The results showed the highest rate of [s] was produced by the speakers of the Eastern Andean variety, followed by the Western Andean, and then by the Coastal variety, suggesting that first dialect phonological processes may affect the acquisition of second dialect sounds. Consistent with previous studies that have examined /s/ variation and change, the linguistic factors position in the word, following segment, and syllable stress were also predictors of /s/ in second dialect production. The extralinguistic factors of age of arrival, age, and gender also had a significant effect on /s/ production in this study. Implications are discussed for models of second dialect speech learning.