Date Published: March 16, 2016
Publisher: Public Library of Science
Author(s): Lama K. Farran, Chia-Cheng Lee, Hyunjoo Yoo, D. Kimbrough Oller, Kim A. Bard.
Infant-directed speech (IDS) provides an environment that appears to play a significant role in the origins of language in the human infant. Differences have been reported in the use of IDS across cultures, suggesting different styles of infant language-learning. Importantly, both cross-cultural and intra-cultural research suggest there may be a positive relationship between the use of IDS and rates of language development, underscoring the need to investigate cultural differences more deeply. The majority of studies, however, have conceptualized IDS monolithically, granting little attention to a potentially key distinction in how IDS manifests across cultures during the first two years. This study examines and quantifies for the first time differences within IDS in the use of baby register (IDS/BR), an acoustically identifiable type of IDS that includes features such as high pitch, long duration, and smooth intonation (the register that is usually assumed to occur in IDS), and adult register (IDS/AR), the type of IDS that does not include such features and thus sounds as if it could have been addressed to an adult. We studied IDS across 19 American and 19 Lebanese mother-infant dyads, with particular focus on the differential use of registers within IDS as mothers interacted with their infants ages 0–24 months. Our results showed considerable usage of IDS/AR (>30% of utterances) and a tendency for Lebanese mothers to use more IDS than American mothers. Implications for future research on IDS and its role in elucidating how language evolves across cultures are explored.
Descriptive statistics are presented in Table 2. The results for Mean Rate in Utterances per Minute show, perhaps surprisingly, a substantial amount of IDS/AR in the sample, accounting for 33% of IDS utterances (35% for Lebanese mothers and 32% for American mothers). Not a single mother failed to produce at least some IDS/AR utterances in her sample, and even for infants < 7 months of age, IDS/AR accounted for >10% of maternal talk. Similarly, the alternative analysis in terms of IDS Seconds per Minute also showed that mothers used IDS/AR quite frequently: for Lebanese mothers, an average of 31% of the time in recordings being occupied by IDS was IDS/AR, and for American mothers, 26% of the time occupied by IDS was IDS/AR.
While others have reported differences in patterns of IDS for different developmental levels  and across cultural contexts , this exploratory study is the first to report quantitatively on the role of language differences in rate of IDS and on use of different registers within IDS (IDS/BR vs IDS/AR) for different languages. The results revealed the expected higher frequency of IDS/BR compared to IDS/AR for both languages, but the high rate of IDS/AR was unexpected, with mothers producing 33% of their utterances in IDS/AR. Especially surprising was the fact that mothers at all infant ages and in both languages produced at least some IDS/AR in these 10-minute samples. In contrast, the higher rate of IDS/AR at older infant ages for both language groups was not surprising, presumably because infants become increasingly able to understand adult speech as they get older  and thus may not need the facilitative support in comprehension brought about by features inherent in IDS/BR [37, 38].
This study was opportunistic, resulting in samples that were less than perfectly matched. It also relied on cross-sectional data, relatively brief durations of mother-infant interactions, and slightly different instrumentation for recording mother-infant interactions across the Lebanese and American cultures. The location of the recordings (home versus laboratory) also differed across the two groups. The impact of this difference is less likely to change the pattern of results we obtained, however, as research suggests similar results in mother-child interactions across the home and laboratory settings [57, 58], especially when mothers from different groups are instructed to interact with their infants . In this study, both Lebanese and American mothers were instructed to interact with their infants. Of similar importance is the limited number of infants studied. A fully longitudinal effort would also be advisable. Improvement on such factors would undoubtedly improve generalizability of the findings.