May 26, 2007

The doors of infant perception

A lovely study from Janet Werker's lab at UBC was published in Science yesterday. The paper is Whitney Weikum, Athena Vouloumanos, Jordi Navarra, Salvador Soto-Faraco, Nuría Sebastiaán-Gallés and Janet Werker, "Visual Language Discrimination in Infancy", Science, 316(5828) 1159 May 25, 2007. The abstract:

This study shows that 4- and 6-month-old infants can discriminate languages (English from French) just from viewing silently presented articulations. By the age of 8 months, only bilingual (French-English) infants succeed at this task. These findings reveal a surprisingly early preparedness for visual language discrimination and highlight infants' selectivity for retaining only necessary perceptual sensitivities.

You can see pictures, a demo movie, and examples of the stimuli here.

How do you test a four-month-old's perceptual abilities?

Discrimination was tested by using silent video clips of three bilingual French-English speakers reciting sentences in each language. Every trial contained a video clip of a different sentence by one speaker in one language ... The infants (n = 36) were presented with video clips from one of the languages until their looking time declined to a 60% habituation criterion. Test trials using the same speakers but different sentences from the other language were shown to examine whether the infants' looking time had increased, indicating that they had noticed the language change. The test trials where the language was switched were compared with a control condition (n = 36) for which the test trials were always different sentences but in the same language as the habituation trials.

The results:

Fig. 1. Mean looking time in seconds to silent talking faces. The y axis represents infant looking time; the x axis represents the trials that the infant was shown (final habituation trials and test trials). Error bars represent the standard error of the mean. (A) Experimental (language switch) and control (language same) conditions for monolingual infants at 4, 6, and 8 months. (B) Experimental conditions for monolingual [replotted from (A)] and bilingual infants at 6 and 8 months.

The authors observe that

The finding that infants can visually discriminate their native language from an unfamiliar language at 4 and 6 months but not at 8 months parallels declines in performance seen in other perceptual domains. Indeed, across the first year of life, infants' performance declines on the discrimination of nonnative consonant and vowel contrasts (6, 7), nonnative musical rhythms (8), cross-species individual faces (9), and cross-species face and voice matching (10). Thus, it appears that specific experience is necessary for maintaining sensitivity to some initial perceptual discriminations.

The references from that last paragraph:

(6) J. F. Werker, R. C. Tees, "Cross-Language Speech Perception -- Evidence for Perceptual Reorganization During the 1st year of Life", Infant Behav. Dev. 7, 49 (1984)
(7) P. K. Kuhl et al., "Infants show a facilitation effect for native language phonetic perception between 6 and 12 months", Dev. Sci. 9, F13 (2006)
(8) E. E. Hannon, S. E. Trehub, "Tuning in to musical rhythms: Infants learn more readily than adults", Proc. Natl. Acad. Sci. U.S.A. 102, 12639 (2005)
(9) O. Pascalis et al., "Plasticity of face processing in infancy", Proc. Natl. Acad. Sci. U.S.A. 102, 5297 (2005)
(10) D. J. Lewkowicz, A. A. Ghazanfar, "The decline of cross-species intersensory perception in human infants", Proc. Natl. Acad. Sci. U.S.A. 103, 6771 (2006)

I expect that a preprint of yesterday's Science paper will appear soon in Janet's well-maintained publications list. Meanwhile, there's a good journalistic account in Tu Thanh Ha, "French and English look different to babies", Globe and Mail, 5/25/2007.

To understand this work fully, you need to see it in the context of what we might call the Blakean tendency in modern developmental studies, according to which infant learning is associated with a selective loss of perceptual abilities. As Blake famously wrote in The Marriage of Heaven and Hell:

If the doors of perception were cleansed, every thing would appear to man as it is: infinite.
For man has closed himself up, till he sees all things thro' the narrow chinks of his cavern.

There is a debate (mostly implicit) about whether this "closing up" is a functionally-essential part of learning (because you can only focus on what matters by ignoring what doesn't), or an unavoidable but secondary effect (because making some distinctions more salient tends to put others into the background), or an inessential and avoidable effect of the infant's perceptual environment (because perceptual abilities that are not exercised will atrophy). This context may help you to understand the discussion at the end of Tu Thanh Ha's article:

It is normal, researchers say, for the young child to prune out unused linguistic skills, a pruning process that helps limit information overload.

"As babies grow up, it wouldn't really be advantageous to continue to hear sound differences in languages they don't use," said the study's research supervisor, UBC psychology professor Janet Werker, a long-time researcher on how babies perceive speech.

At the same time, Prof. Vouloumanos said the recent findings also made her believe that there is no harm in getting children to learn new languages at the earliest age.

There is no reason that a healthy child should not be exposed to multiple languages that are spoken in their natural environment, she said, emphasizing that she was stating a personal opinion.

"Much research has shown that young infants can do many things with language, and, in many cases, better than adults can."

She and Prof. Werker noted that, outside of North America, the majority of young children grow up in multilingual environments.

"The brain is definitely set up to acquire more than one language," Prof. Werker said.

For some additional perspective on the topic of visual language discrimination, see Salvador Soto-Faraco, Jordi Navarra, Whiney M. Weikum, Athena Vouloumanos, Nuría Sebastiaán-Gallés and Janet F. Werker, "Disciminating Languages by Speech-reading", Perception and Psychophysics, 2006. The abstract:

The goal of this study was to explore the ability to discriminate languages using the visual correlates of speech (i.e., speech-reading). Participants were presented with silent video-clips of an actor pronouncing two sentences (in Catalan and/or Spanish), and asked to judge whether the sentences were in the same or in different languages. Our results established that Spanish-Catalan bilingual speakers could discriminate running speech from their two languages (Spanish and Catalan) based on visual cues alone (Experiment 1). However, we found that this ability was critically restricted by linguistic experience, as Italian and English speakers, who were unfamiliar with the test languages, could not successfully discriminate the stimuli (Experiment 2). A test of Spanish monolingual speakers revealed that knowledge of only one of the two test languages was sufficient to achieve the discrimination, although at a lower level of accuracy than that seen in bilingual speakers (Experiment 3). Finally, we evaluated the ability to identify the language by speech-reading particularly distinctive words (Experiment 4). The results obtained are in accord with recent proposals arguing that the visual speech signal is rich in informational content, above and beyond what traditional accounts based solely on visemic confusion matrices would predict.

[Randy Alexander wrote:

What you took as a lovely study in your post, I took as a sign of impending doom.

I design methods for children to learn English as a foreign language.

As I read this news yesterday, I saw in my mind's eye crowds of mothers pushing their way into my office begging me to teach English to their four month old babies.

I'm praying that this doesn't make it to the Chinese media.

It's gonna be like the Mozart effect all over again. :-(

What, is business so good that the prospect of opening up a whole new market feels like impending doom? ]

Posted by Mark Liberman at May 26, 2007 07:56 AM