Semantics guide infants’ vowel learning: Computational and experimental evidence

Abstract

Semantics guide infants' vowel learning: computational and experimental evidenceter Schure, S.M.M.; Junge, C.M.M.; Boersma, P.P.G. General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons). Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible. a b s t r a c tIn their first year, infants' perceptual abilities zoom in on only those speech sound contrasts that are relevant for their language. Infants' lexicons do not yet contain sufficient minimal pairs to explain this phonetic categorization process. Therefore, researchers suggested a bottom-up learning mechanism: infants create categories aligned with the frequency distributions of sounds in their input. Recent evidence shows that this bottom-up mechanism may be complemented by the semantic context in which speech sounds occur, such as simultaneously present objects. To test this hypothesis, we investigated whether discrimination of a non-native vowel contrast improves when sounds from the contrast were paired consistently or randomly with two distinct visually presented objects, while the distribution of speech tokens suggested a single broad category. This was assessed in two ways: computationally, namely in a neural network simulation, and experimentally, namely in a group of 8-month-old infants. The neural network, trained with a large set of sound-meaning pairs, revealed that two categories emerge only if sounds are consistently paired with objects. A group of 49 real 8-month-old infants did not immediately show sensitivity to the pairing condition; a later test at 18 months with some of the same infants, however, showed that this sensitivity at 8 months interacted with their vocabulary size at 18 months. This interaction can be explained by the idea that infants with larger future vocabularies are more positively influenced by consistent training (and/or more negatively influenced by inconsistent training) than infants with smaller future vocabularies. This suggests that consistent pairing with distinct visual objects can help infants to discriminate speech sounds even when the auditory information does not signal a distinction. Together our results give computational as well as experimental support for the idea that semantic context plays a role in disambiguating phonetic auditory input.

Distributional Information Shapes Infants’ Categorization of Objects

“…Since this seminal study (Maye et al., ), there have been several studies evidencing that the shape of frequency distributions guides categorization, yet all focused on (auditory) phoneme acquisition (e.g., Maye, Weiss, & Aslin, ; ter Schure, Junge, & Boersma, ; Yoshida, Pons, Maye, & Werker, ). It thus remains an open question whether infants employ this learning mechanism also outside the spoken language domain, that is, whether it serves as a domain‐general cue.…”

Can infants learn phonology in the lab? A meta-analytic answer

“…After applying the inclusion criteria, there were 13 papers that could be considered for a quantitative meta-analysis, from a variety of sources: 6 journal articles (Liu & Kager, 2014;Maye et al, 2002Maye et al, , 2008ter Schure, Junge, & Boersma, 2016;Wanrooij et al, 2014;Yoshida et al, 2010), 1 article in proceedings (Liu & Kager, 2011), 1 chapter in a collection (Capel, De Bree, De Klerk, Kerkhoff, & Wijnen, 2011), 1 unpublished manuscript (Cristia, 2011), and 4 sets of data that had not been included in public written reports but had been presented at conferences as posters or talks (Fennell, Hudon, & Spring, 2012;Pons, Mugitani, Amano, & Werker, 2006;Pons, Sabourin, Cady, & Werker, 2008).…”

“…As mentioned in the Introduction, it has been proposed that infants could (additionally) exploit prelexical, lexical and/or multimodal learning strategies when learning about the sound categories in their native language, and some of these proposals are even supported their own proof-of-principle experiments (ter Schure et al, 2016;Feldman et al, 2013). Supposing that further research along these lines confirms the robustness and reliability of such early results, future work could assess whether multiple such approaches may be fruitfully combined, or whether at certain ages or in certain situations, infants solely exploit a subset of such sound category learning strategies.…”

“…There was an exception to this general procedure, namely three papers for which difference scores and SD of the difference were available (Pons et al, 2006(Pons et al, , 2008ter Schure et al, 2016). To maximize comparability with the other data, the effect size was calculated as the ratio of the difference scores to the standard deviation, when the latter is corrected by a factor appropriate to the correlation in individuals' repeated measures.…”

Proceedings of the 21st Conference on Computational Natural Language Learning (CoNLL 2017)

“…The forms around peaks in this distribution are then perceived as being a distinct category. Recent computational models showed that infant-directed speech contains sufficiently clear peaks for such a distributional learning mechanism to succeed and also that top-down processes like semantic knowledge and visual information play a role in phonetic category learning (ter Schure et al, 2016).…”