El reconocimiento temprano de la lengua materna: un estudio basado en la voz masculina

Resumen

El trabajo analiza la capacidad temprana del bebé para reconocer la lengua materna y diferenciarla de otra lengua no familiar. Los bebés estudiados, de 4 meses de edad, proceden de entornos familiares monolingües, en los que se habla el castellano o el catalán. La comparación se establece, pues, entre dos lenguas pertenecientes a una misma clase fonológica (ambas son de ritmo silábico) por lo que la diferenciación puede verse comprometida. El estudio incorpora además la novedad de utilizar la voz masculina y múltiples locutores en la preparación de los materialespara la prueba de discriminación. Los resultados confirman la capacidad del bebé para diferenciar este par de lenguas con independencia del tipo de voz y de la variabilidadde los locutores.

Abstract

This study analyses infants' early capacity to identify their maternal language and to differentiate it from a non-familiar language. A group of four-month-old infants, from Spanish-Castilian or Catalan monolingual environments, participated in the study. The languages being compared belong to the same rhythmic class, i.e., they can be considered syllable-timed languages. This makes differentiation more problematic for young infants. What is new in this investigation is the use of male-voice and different speakers in the recordings of utterances used in the discrimination test. The results give evidence of young infants' capacity to tell apart these two languages regardless of voice and talker variability.

Extended Summary

This study analyses infants' early capacity to identify their mother tongue and to differentiate it from a non-familiar language. Previous research in this domain indicates that maternal language recognition abilities emerge very early on, but they are limited by the prosodic characteristics of the languages being compared (Mehler et al., 1996). For languages that belong to the same phonological (rhythmic) category, discrimination may not be achieved during the first months of life. Newborns (Nazzi et al., 1998) and two-month-old infants (Christophe and Morton, 1998) were unable to tell apart utterances form two stress-timed languages such as Dutch and English. For the first time in a series of recent studies (Bosch and Sebastián, 1997a and b; Bosch, 1998; Bosch and Sebastián, 1999) evidence of discrimination between a pair of syllable-timed languages, Spanish-Castilian and Catalan, was obtained for infants aged 4.5 months.

In the present paper, two groups of four-month-old infants (n=20), from Spanish-Castilian and Catalan monolingual environments, have been tested to further analyse their discrimination capacity. The novelty of the present study lies in the use of male-voices and four different speakers in the materials testing discrimination. A total of 28 different utterances, from four different male speakers, were selected from spontaneous narrations of a story in child-directed speech. The utterances were equated in both number of syllables and total duration. The lower fundamental frequency values of male voices compared to female voices, and the amount of variability in the material, could challenge infants' ability to differentiate this pair of syllable-timed languages.

Infants were tested using a procedure that had already been successfully employed with two-month-old infants, in a French vs. English comparison (Dehaene-Lambertz and Houston, 1998). The procedure is a visual orientation technique that measures orientation latencies (RTs) towards utterances in familiar vs. unfamiliar languages. It is based on previous results from visual attention studies in which young infants, carrying out visual recognition tasks, responded faster towards familiar stimuli.

The present results show faster orientation latencies towards utterances spoken in the infant's the maternal language; that is, Spanish-Castilian infants oriented faster towards Spanish-Castilian utterances, while Catalan infants showed a similar pattern for Catalan utterances. The interaction between the two factors (environment language and test utterance language) was highly significant [F (1,18) = 32.179, p < 0.0001].

These results not only confirm previous data from our laboratory in that they support the hypothesis that languages from the same rhythmic class can be differentiated at least by 4 months of age, but they also indicate that this early discrimination capacity is probably based on very robust acoustic cues in the speech signal that are not affected by type of voice and speaker variability. It is hypothesized that the sequence of vowel sounds in the language can offer reliable cues that might aid infants in this recognition/discrimination task, either because of the rhythmic information they carry (as a sequence of elements with different duration and loudness), or because of their distributional properties (type of vowel sounds more frequent in the language).

Palabras clave:

• Discriminación de lenguas
• bebés
• voz masculine

Keywords:

• Language discrimination
• young infants
• male-voices