Kathleen Currie Hall

Notes on: Foulkes, Paul, Gerard Docherty, and Dominic Watt. (2005). Phonological variation in child-directed speech. Language 81(1), 177-206.

Summary:
In this paper, the authors present an analysis of the acoustics of (t) in child-directed speech in Tyneside English.  They combine a sociolinguistic and acoustic phonetic approach, and look at effects of age and gender of the child on the mother's speech.  They are able to directly compare their findings for the CDS with patterns in interadult speech because they have previously conducted numerous studies on the adult speech patterns of this dialect of English.  They find that overall, there is more use of the standard [t] pronunciation of (t) with children than with adults, with girls than with boys, and with younger children than with older children.  They emphasize that these differences may reflect (unconscious) efforts on the part of parents to enforce larger societal gender norms.

Details:
1. CDS is generally assumed to exist for both analytic and social reasons.  Analytic reasons are those that are thought to enhance a child's acquisition of linguistic structure; social reasons are those that are thought to enhance the child's realization that language is a social, interactive phenomenon.

2. However, there has been a lot of controversy / conflicting evidence about the characteristics and functions of CDS.  While some authors claim that CDS tends to be "hyperspeech" and particularly well-suited for analytic processing by a child, others claim that there is in fact more variability in CDS and that its non-canonical nature could make it more difficult for a child to learn language; these authors stress the social functions of CDS.  Foulkes, Docherty, and Watt attempt to address some of the controversy over what characterizes CDS by using CDS in a dialect where the adult speech is particularly well studied as well as by doing a cross-sectional study to compare different age groups.

3. The authors recorded mother-child interactions (mediated by clinicians with particular materials designed to elicit particular segments) for 40 dyads.  They focused on the production of (t) in two positions: word-medial between sonorants, where it is variably realized as [t] or as a "glottal" (creaky, voiced coronal); and word-final prevocalically, where it is variably realized as [t], the glottal, [R] (retroflex approximant), or a voiced coronal (flap or [d]).  In these positions, [t] is the "standard" pronunciation and the others are sociolinguistically marked variants (with the glottal being an overt marker, especially common to this geographic area and particularly among males; the [R] being overt but associated with a larger area; and the voiced variants not being noticed by speakers at all).

4. They found that mothers used more [t] pronunciations in CDS than they do in interadult speech.  Furthermore, they tend to use more [t] pronunciations with girls than with boys, and more [t]s with younger children than with older children, though the age differences were more significant with girls than with boys.

5. While the authors acknowledge that their findings do match well with other studies that have found that CDS tends to have more "standard" forms than interadult speech, they question the typical interpretation that this is to enhance the child's acquisition of language.  (The standard is not inherently easier to produce, and enforcing the standard form means the child is exposed to more variability and may have a harder time correctly learning the varnacular.)

6. Instead, they espouse the idea that the differences seen in CDS (and especially differences within CDS to girls vs. boys or younger vs. older children) are indicative of the mother's expectations for her child's social identity.  The differences between CDS to girls as opposed to boys reflects the tendency for women in Tyneside English to use more standard forms than men; the mother's CDS encourages this differentiation by giving boys more non-standard input than girls and modelling the social differentiation for the children.  As the child grows older, the mother feels less pressure to create the child's social identity for him/her (because the child is more linguistically capable and can express his/her own identity), and so use of the standard forms declines with age.

Questions:
1. Were the studies they cite on p. 180 about the effects of adult and child gender on CDS controlled for language / society?  I would imagine that there are fairly large differences across societies / cultures about the role of men and women in child-rearing -- which groups are these results valid for?

2. The authors note (p. 186) that women in their adult-speech study tended to use the standard [t] variant in the read wordlist almost exclusively, despite using the glottal in unscripted conversation.  How did the amount of [t] usage in their CDS compare to the amount of usage on the wordlist?

3. Clearly there is more than just sociolinguistic marking going on, as mothers use more standard forms with boys than they do with adults . . . is it a prescriptive desire on the part of the mothers (e.g. "I want my child to speak properly")?  Or is there an advantage to acquisition by hearing more variabiltiy?  The adults clearly know the alternations between [t] and the glottals or other forms -- how would the child learn the system if he weren't exposed to such variation?  Why is there the assumption that variability would be bad for a child?