Publications

Abstract

Both quality and quantity of speech from the primary caregiver have been found to impact language development. A third aspect of the input has been largely ignored: the number of talkers who provide input. Some infants spend most of their waking time with only one person; others hear many different talkers. Even if the very same words are spoken the same number of times, the pronunciations can be more variable when several talkers pronounce them. Is language acquisition affected by the number of people who provide input? To shed light on the possible link between how many people provide input in daily life and infants’ native vowel discrimination, three age groups were tested: 4-month-olds (before attunement to native vowels), 6-month-olds (at the cusp of native vowel attunement) and 12-month-olds (well attuned to the native vowel system). No relationship was found between talker number and native vowel discrimination skills in 4- and 6-month-olds, who are overall able to discriminate the vowel contrast. At 12 months, we observe a small positive relationship, but further analyses reveal that the data are also compatible with the null hypothesis of no relationship. Implications in the context of infant language acquisition and cognitive development are discussed.

Abstract

Previous work suggests key factors for replicability, a necessary feature for theory
building, include statistical power and appropriate research planning. These factors are examined by analyzing a collection of 12 standardized meta-analyses on language development between birth and 5 years. With a median effect size of Cohen's d= 0.45 and typical sample size of 18 participants, most research is underpowered (range: 6%-99%;
median 44%); and calculating power based on seminal publications is not a suitable strategy.
Method choice can be improved, as shown in analyses on exclusion rates and effect size as a
function of method. The article ends with a discussion on how to increase replicability in both language acquisition studies specifically and developmental research more generally.

Abstract

Rapid response latencies in conversation suggest that responders start planning before the ongoing turn is finished. Indeed, an earlier EEG study suggests that listeners start planning their responses to questions as soon as they can (Bögels, S., Magyari, L., & Levinson, S. C. (2015). Neural signatures of response planning occur midway through an incoming question in conversation. Scientific Reports, 5, 12881). The present study aimed to (1) replicate this early planning effect and (2) investigate whether such early response planning incurs a cost on participants’ concurrent comprehension of the ongoing turn. During the experiment participants answered questions from a confederate partner. To address aim (1), the questions were designed such that response planning could start either early or late in the turn. Our results largely replicate Bögels et al. (2015) showing a large positive ERP effect and an oscillatory alpha/beta reduction right after participants could have first started planning their verbal response, again suggesting an early start of response planning. To address aim (2), the confederate's questions also contained either an expected word or an unexpected one to elicit a differential N400 effect, either before or after the start of response planning. We hypothesized an attenuated N400 effect after response planning had started. In contrast, the N400 effects before and after planning did not differ. There was, however, a positive correlation between participants' response time and their N400 effect size after planning had started; quick responders showed a smaller N400 effect, suggesting reduced attention to comprehension and possibly reduced anticipatory processing. We conclude that early response planning can indeed impact comprehension processing.

Abstract

Speaker diarization (answering 'who spoke when') is a widely researched subject within speech technology. Numerous experiments have been run on datasets built from broadcast news, meeting data, and call centers—the task sometimes appears close to being solved. Much less work has begun to tackle the hardest diarization task of all: spontaneous conversations in real-world settings. Such diarization would be particularly useful for studies of language acquisition, where researchers investigate the speech children produce and hear in their daily lives. In this paper, we study audio gathered with a recorder worn by small children as they went about their normal days. As a result, each child was exposed to different acoustic environments with a multitude of background noises and a varying number of adults and peers. The inconsistency of speech and noise within and across samples poses a challenging task for speaker diarization systems, which we tackled via retraining and data augmentation techniques. We further studied sources of structured variation across raw audio files, including the impact of speaker type distribution, proportion of speech from children, and child age on diarization performance. We discuss the extent to which these findings might generalize to other samples of speech in the wild.

Abstract

Children’s songs often contain rhyming words at phrase endings. In this study, we investigated whether infants can already recognize this phonological pattern in songs. Earlier studies using lists of spoken words were equivocal on infants’ spontaneous processing of rhymes (Hayes, Slater, & Brown, 2000; Jusczyk, Goodman, & Baumann, 1999). Songs, however, constitute an ecologically valid rhyming stimulus, which could allow for spontaneous processing of this phonological pattern in infants. Novel children’s songs with rhyming and non-rhyming lyrics using pseudo-words were presented to 35 9-month-old Dutch infants using the Headturn Preference Procedure. Infants on average listened longer to the non-rhyming songs, with around half of the infants however exhibiting a preference for the rhyming songs. These results highlight that infants have the processing abilities to benefit from their natural rhyming input for the development of their phonological abilities.

Abstract

A common assumption in the cognitive sciences is that artificial and natural language learning rely on shared mechanisms. However, attempts to bridge the two have yielded ambiguous results. We suggest that an empirical disconnect between the computations employed during learning and the methods employed at test may explain these mixed results. Further, we propose statistically-based chunking as a potential computational link between artificial and natural language learning. We compare the acquisition of non-adjacent dependencies to that of natural language structure using two types of tasks: reflection-based 2AFC measures, and processing-based recall measures, the latter being more computationally analogous to the processes used during language acquisition. Our results demonstrate that task-type significantly influences the correlations observed between artificial and natural language acquisition, with reflection-based and processing-based measures correlating within – but not across – task-type. These findings have fundamental implications for artificial-to-natural language comparisons, both methodologically and theoretically.

Abstract

The mutual exclusivity (ME) assumption is proposed to facilitate early word learning by guiding infants to map novel words to novel referents. This study assessed the emergence and use of ME to both disambiguate and retain the meanings of novel words across development in 18‐month‐old monolingual and bilingual children (Experiment 1; N = 58), and in a sub‐group of these children again at 24 months of age (Experiment 2: N = 32). Both monolinguals and bilinguals employed ME to select the referent of a novel label to a similar extent at 18 and 24 months. At 18 months, there were also no differences in novel word retention between the two language‐background groups. However, at 24 months, only monolinguals showed the ability to retain these label–object mappings. These findings indicate that the development of the ME assumption as a reliable word‐learning strategy is shaped by children's individual language exposure and experience with language use.

Abstract

We report a large‐scale electrophysiological study of infant speech segmentation, in which over 100 English‐acquiring 9‐month‐olds were exposed to unfamiliar bisyllabic words embedded in sentences (e.g., He saw a wild eagle up there), after which their brain responses to either the just‐familiarized word (eagle) or a control word (coral) were recorded. When initial exposure occurs in continuous speech, as here, past studies have reported that even somewhat older infants do not reliably recognize target words, but that successful segmentation varies across children. Here, we both confirm and further uncover the nature of this variation. The segmentation response systematically varied across individuals and was related to their vocabulary development. About one‐third of the group showed a left‐frontally located relative negativity in response to familiar versus control targets, which has previously been described as a mature response. Another third showed a similarly located positive‐going reaction (a previously described immature response), and the remaining third formed an intermediate grouping that was primarily characterized by an initial response delay. A fine‐grained group‐level analysis suggested that a developmental shift to a lexical mode of processing occurs toward the end of the first year, with variation across individual infants in the exact timing of this shift.

Abstract

Humans differ in innumerable ways, with considerable variation observable at every level of description, from the molecular to the social. Traditionally, linguistic and psycholinguistic theory has downplayed the possibility of meaningful differences in language across individuals. However, it is becoming increasingly evident that there is
significant variation among speakers at any age as well as across the lifespan. In this paper, we review recent research in psycholinguistics, and argue that a focus on individual differences provides a crucial source of evidence that bears strongly upon core issues in theories of the acquisition and processing of language; specifically, the role of experience in language acquisition, processing, and attainment, and the architecture of the language faculty.

Abstract

A developmental relationship between symbolic play and language has been long proposed, going as far back as the writings of Piaget and Vygotsky. In the current paper we build on recent qualitative reviews of the literature by reporting the first quantitative analysis of the relationship. We conducted a three-level meta-analysis of past studies that have investigated the relationship between symbolic play and language acquisition. Thirty-five studies (N = 6848) met the criteria for inclusion. Overall, we observed a significant small-to-medium association between the two domains (r = .35). Several moderating variables were included in the analyses, including: (i) study design (longitudinal, concurrent), (ii) the manner in which language was measured (comprehension, production), and (iii) the age at which this relationship is measured. The effect was weakly moderated by these three variables, but overall the association was robust, suggesting that symbolic play and language are closely related in development.

Abstract

Word count estimation (WCE) from audio recordings has a number of applications, including quantifying the amount of speech that language-learning infants hear in their natural environments, as captured by daylong recordings made with devices worn by infants. To be applicable in a wide range of scenarios and also low-resource domains, WCE tools should be extremely robust against varying signal conditions and require minimal access to labeled training data in the target domain. For this purpose, earlier work has used automatic syllabification of speech, followed by a least-squares-mapping of syllables to word counts. This paper compares a number of previously proposed syllabifiers in the WCE task, including a supervised bi-directional long short-term memory (BLSTM) network that is trained on a language for which high quality syllable annotations are available (a “high resource language”), and reports how the alternative methods compare on different languages and signal conditions. We also explore additive noise and varying-channel data augmentation strategies for BLSTM training, and show how they improve performance in both matching and mismatching languages. Intriguingly, we also find that even though the BLSTM works on languages beyond its training data, the unsupervised algorithms can still outperform it in challenging signal conditions on novel languages.

Abstract

Before infants become mature speakers of their native language, they must acquire a robust word-recognition system which allows them to strike the balance between allowing some variation (mood, voice, accent) and recognizing variability that potentially changes meaning (e.g. cat vs hat). The current meta-analysis quantifies how the latter, termed mispronunciation sensitivity, changes over infants’ first three years, testing competing predictions of mainstream language acquisition theories. Our results show that infants were sensitive to mispronunciations, but accepted them as labels for target objects. Interestingly, and in contrast to predictions of mainstream theories, mispronunciation sensitivity was not modulated by infant age, suggesting that a sufficiently flexible understanding of native language phonology is in place at a young age.