Early language-specificity of children's event encoding in speech and gesture: evidence from caused motion in Turkish

Abstract

Previous research on language development shows that children are tuned early on to the language-specific semantic and syntactic encoding of events in their native language. Here we ask whether language-specificity is also evident in children's early representations in gesture accompanying speech. In a longitudinal study, we examined the spontaneous speech and cospeech gestures of eight Turkish-speaking children aged one to three and focused on their caused motion event expressions. In Turkish, unlike in English, the main semantic elements of caused motion such as Action and Path can be encoded in the verb (e.g. sok- ‘put in’) and the arguments of a verb can be easily omitted. We found that Turkish-speaking children's speech indeed displayed these language-specific features and focused on verbs to encode caused motion. More interestingly, we found that their early gestures also manifested specificity. Children used iconic cospeech gestures (from 19 months onwards) as often as pointing gestures and represented semantic elements such as Action with Figure and/or Path that reinforced or supplemented speech in language-specific ways until the age of three. In the light of previous reports on the scarcity of iconic gestures in English-speaking children's early productions, we argue that the language children learn shapes gestures and how they get integrated with speech in the first three years of life.

Keywords:

• gesture
• language
• development
• caused motion
• Turkish

Acknowledgements

This work has been supported by a Huygens Scholarship to the first author by the Turkish Academy of Sciences award in the framework of the Young Scientist Award Program given to the second author (AK-TUBA-GEBIP/2001-2-13) and by an NWO VIDI grant and ASPASIA award as well as an ERC Starting Grant given to the last author. Part of the data collection was supported by the EUROCORE programme ‘The Origin of Man, Language and Languages’ (OMLL) and the French CNRS programme ‘Origine de l'Homme, du Langage et des Langues’ (OHLL), main investigator Sophie Kern (Laboratory Dynamique du Langage, UMR5590, Lyon, France). We would like to thank Dilara Koçbaş for data collection, and Beyza Sümer and Nazlı Altınok for their help in establishing reliability.

Notes

1. Note that 20% of all actions that overlapped with caused motion speech were considered for reliability of gesture type. The categorisation by the original coder and the reliability coder had to include a ‘real action’ category (which later was excluded from the analyses reported here) as well the categories of iconics, points and showing gestures. In this way, we obtained reliability for isolating gestures from non-gestures in the same round that we determined the gesture types.

2. Although we started our investigation of caused motion encoding from the age of 12 months, children did not start to talk about these events before they were 14 months old and gesture before 15 months (see Table 4).

3. Since both studies by Özçalışkan and Goldin-Meadow (2009, 2011) have examined the spontaneous speech and gestures children used in their interactions with caregivers and are thus similar to our data, we feel that making comparisons across the results of these studies is justified.