Search for: All records

It has long been assumed that infants' ability to discriminate between languages stems from their sensitivity to speech rhythm, i.e., organized temporal structure of vowels and consonants in a language. However, the relationship between speech rhythm and language discrimination has not been directly demonstrated. Here, we use computational modeling and train models of speech perception with and without access to information about rhythm. We test these models on language discrimination, and find that access to rhythm does not affect the success of the model in replicating infant language discrimination results. Our findings challenge the relationship between rhythm and language discrimination, and have implications for theories of language acquisition. 

Rhythm plays an important role in language perception and learning, with infants perceiving rhythmic differences across languages at birth. While the mechanisms underlying rhythm perception in speech remain unclear, one interesting possibility is that these mechanisms are similar to those involved in the perception of musical rhythm. In this work, we adopt a model originally designed for musical rhythm to simulate speech rhythm perception. We show that this model replicates the behavioral results of language discrimination in newborns, and outperforms an existing model of infant language discrimination. We also find that percussives — fast-changing components in the acoustics — are necessary for distinguishing languages of different rhythms, which suggests that percussives are essential for rhythm perception. Our music-inspired model of speech rhythm may be seen as a first step towards a unified theory of how rhythm is represented in speech and music. 

Non-native speakers show difficulties with spoken word processing. Many studies attribute these difficulties to imprecise phonological encoding of words in the lexical memory. We test an alternative hypothesis: that some of these difficulties can arise from the non-native speakers' phonetic perception. We train a computational model of phonetic learning, which has no access to phonology, on either one or two languages. We first show that the model exhibits predictable behaviors on phone-level and word-level discrimination tasks. We then test the model on a spoken word processing task, showing that phonology may not be necessary to explain some of the word processing effects observed in non-native speakers. We run an additional analysis of the model's lexical representation space, showing that the two training languages are not fully separated in that space, similarly to the languages of a bilingual human speaker. 

Abstract Early changes in infants’ ability to perceive native and nonnative speech sound contrasts are typically attributed to their developing knowledge of phonetic categories. We critically examine this hypothesis and argue that there is little direct evidence of category knowledge in infancy. We then propose an alternative account in which infants’ perception changes because they are learning a perceptual space that is appropriate to represent speech, without yet carving up that space into phonetic categories. If correct, this new account has substantial implications for understanding early language development. 

Before they even speak, infants become attuned to the sounds of the language(s) they hear, processing native phonetic contrasts more easily than nonnative ones. For example, between 6 to 8 mo and 10 to 12 mo, infants learning American English get better at distinguishing English and [l], as in “rock” vs. “lock,” relative to infants learning Japanese. Influential accounts of this early phonetic learning phenomenon initially proposed that infants group sounds into native vowel- and consonant-like phonetic categories—like and [l] in English—through a statistical clustering mechanism dubbed “distributional learning.” The feasibility of this mechanism for learning phonetic categories has been challenged, however. Here, we demonstrate that a distributional learning algorithm operating on naturalistic speech can predict early phonetic learning, as observed in Japanese and American English infants, suggesting that infants might learn through distributional learning after all. We further show, however, that, contrary to the original distributional learning proposal, our model learns units too brief and too fine-grained acoustically to correspond to phonetic categories. This challenges the influential idea that what infants learn are phonetic categories. More broadly, our work introduces a mechanism-driven approach to the study of early phonetic learning, together with a quantitative modeling framework that can handle realistic input. This allows accounts of early phonetic learning to be linked to concrete, systematic predictions regarding infants’ attunement. 

In the first year of life, infants' speech perception becomes attuned to the sounds of their native language. Many accounts of this early phonetic learning exist, but computational models predicting the attunement patterns observed in infants from the speech input they hear have been lacking. A recent study presented the first such model, drawing on algorithms proposed for unsupervised learning from naturalistic speech, and tested it on a single phone contrast. Here we study five such algorithms, selected for their potential cognitive relevance. We simulate phonetic learning with each algorithm and perform tests on three phone contrasts from different languages, comparing the results to infants' discrimination patterns. The five models display varying degrees of agreement with empirical observations, showing that our approach can help decide between candidate mechanisms for early phonetic learning, and providing insight into which aspects of the models are critical for capturing infants' perceptual development. 

In acquiring language, differences in input can greatly affect learning outcomes, but which aspects of language learning are most sensitive to input variations, and which are robust, remains debated. A recent modeling study successfully reproduced a phenomenon empirically observed in early phonetic learning---learning about the sounds of the native language in the first year of life---despite using input that differed in quantity and speaker composition from what a typical infant would hear. In this paper, we carry out a direct test of that model's robustness to input variations. We find that, despite what the original result suggested, the learning outcomes are sensitive to properties of the input and that more plausible input leads to a better fit with empirical observations. This has implications for understanding early phonetic learning in infants and underscores the importance of using realistic input in models of language acquisition. 

Human listeners are better at telling apart speakers of their native language than speakers of other languages, a phenomenon known as the language familiarity effect. The recent observation of such an effect in infants as young as 4.5 months of age (Fecher & Johnson, in press) has led to new difficulties for theories of the effect. On the one hand, retaining classical accounts—which rely on sophisticated knowledge of the native language (Goggin, Thompson, Strube, & Simental, 1991)–requires an explanation of how infants could acquire this knowledge so early. On the other hand, letting go of these accounts requires an explanation of how the effect could arise in the absence of such knowledge. In this paper, we build on algorithms from unsupervised machine learning and zero-resource speech technology to propose, for the first time, a feasible acquisition mechanism for the language familiarity effect in infants. Our results show how, without relying on sophisticated linguistic knowledge, infants could develop a language familiarity effect through statistical modeling at multiple time-scales of the acoustics of the speech signal to which they are exposed. 

The way listeners perceive speech sounds is largely determined by the language(s) they were exposed to as a child. For example, native speakers of Japanese have a hard time discriminating between American English /ɹ/ and /l/, a phonetic contrast that has no equivalent in Japanese. Such effects are typically attributed to knowledge of sounds in the native language, but quantitative models of how these effects arise from linguistic knowledge are lacking. One possible source for such models is Automatic Speech Recognition (ASR) technology. We implement models based on two types of systems from the ASR literature—hidden Markov models (HMMs) and the more recent, and more accurate, neural network systems—and ask whether, in addition to showing better performance, the neural network systems also provide better models of human perception. We find that while both types of systems can account for Japanese natives’ difficulty with American English /ɹ/ and /l/, only the neural network system successfully accounts for Japanese natives’ facility with Japanese vowel length contrasts. Our work provides a new example, in the domain of speech perception, of an often observed correlation between task performance and similarity to human behavior.