A cross-cultural study examining acoustic features of infant-directed speech. Data were collected on French, Italian, German, Japanese, and British and American English. There is need for cross-linguistic study because of culture-specific features that confound the general trends. (For example, adult speakers of Quiche-Mayan employ a high-pitched register to express deference to other adults.) The purpose of this study was thus to seek universals in CDS prosody.

Procedure

60 parents in major metropolitan areas of their respective countries were recorded while playing games with their infants. The recordings were filtered of noise and analyzed for frequency, amplitude, and duration. Utterances were determined by acoustic, rather than linguistic, features.

100 adult- and infant-directed utterances of each subject were analyzed in regard to two areas:

	voice frequency average extremes range
	utterances length of utterance length of pauses between utterances

Results

Without exception, all parents spoke to infants with higher frequency, greater variability in frequency, shorter utterances, and longer pauses. Mothers increased their frequency relatively more than fathers, regardless of language. Also, American English speakers of both genders used the greatest exaggeration in prosody when speaking to infants (see chart).

Discussion

That mothers are more prosodically expressive supports the Bridge Hypothesis. Research indicates that infants prefer the acoustic exaggeration and variation characteristic of CDS, making mothers' speech more accessible to an infant.

While it is true that the great range of frequency indicates modulation, it should also be noted that CDS also involves an array of short, high-pitched vocalizations such as hmm, yeah, and eh?, which have a strong effect on the data.

It is also hasty to assume that the stark contrast in American English indicates genuine cultural differences. First, because the experimenter was American, those subjects may have felt more relaxed and thus more expressive. The authors also cite Bronfenbrenner's (1979) observation that American mothers speak much more to their infants when they are aware of an observer, and suggest that parents of different cultures might exhibit similar, but not identical, tendencies. Japanese mothers, for example, may actually become less expressive. This does not, however, confound the data. Moreover, it is illustrative of the culture-specific concepts parents hold regarding appropriate behavior.

 

A study of interrogatives and scaffolding in 37 Mexican-American families living along the central coast of California. Parents had a wide range of education, from third grade to graduate degrees. Most, but not all, were bilingual. The researchers also examined the effect of the child's gender on parents' speech. Based on their previous research, they expected mothers to produce more scaffolding responses, and that scaffolding would be more likely in cross-gender dyads.

Procedure

Each dyad was videotaped playing with three pre-determined toys for eight minutes each, a total of 24 minutes per dyad. A later variation introduced gender-specific toys. To code interactions, conversations were broken down into question sequences:

Parent Question > Child Answer > Parent Response

Sequences were further categorized as perceptual (involving description and labeling, such as "What is next to the man?") and conceptual (involving comparison and hypothesis, such as "What would happen if the lion were next to the person?).

Further categories dealt with the child's type of answer and whether parental responses included scaffolding.

Results

Fathers tended to ask slightly more questions, and both parents asked more questions in cross-gender dyads. On the whole, there was little difference in scaffolding between mothers and fathers. However, mothers were more likely to scaffold after:

	correct answers from either gender
	sons' incorrect answers
	daughters' correct answers
	sons' counter-questions

Conclusions

As expected, mothers produced more scaffolding responses than fathers. However, cross-gender dyads did not yield more scaffolding. This indicates that parent -- not child -- gender plays the greater role in communication style. Results were similar to data concerning families of European descent that suggest that fathers are less attuned to their children's cognitive capabilities.

Analysis

While the authors contend that their findings indicate a communicative alienation between fathers and children, this study is better viewed as evidence for the Bridge Hypothesis. Fathers were less likely to scaffold and asked more questions, presenting greater communicative challenges to their children. However, it is difficult to integrate the results with other studies, mostly because it is not clear how much time each parent generally spent with their children. Parental roles have a strong effect on CDS, as documented in Rowe, et al. (below).

 

A study of dyadic interactions with two-year-olds in 33 welfare-eligible families in rural New England. The authors also addressed the bridge hypothesis, which is largely based on the white middle-class of the 1970s. Their intent was to examine the bridge hypothesis in low-income families in which childcare tends to be shared more equally between parents. They also compared speech of resident and non-resident fathers in order to identify changes in the paternal role.

Procedure

In 30-60 minute interviews, each parent was questioned about family composition, parenting practices, beliefs about child-rearing, and their financial and social circumstances.

In videotaped ten-minute sessions, each dyad was given a book and two age-appropriate toys. While the three were always introduced in the same order, parent and child determined the pacing of the interactions.

Interaction was transcribed using CHILDES with respect to

	total number of words
	diversity of vocabulary
	MLU
	frequency and types of questions
	frequency and types of directives

Results

Fathers and mothers showed no significant differences in the first three measures. This indicates an equal level of accommodation for the child's syntactic and semantic skills. However, fathers asked more than twice the amount of questions than mothers, especially Wh-questions and requests for clarification. This is consistent with other observations in middle-class studies. Fathers did not produce more directives than mothers. This conflicts with middle-class research, which showed middle-class fathers using significantly more directives (Hoff-Ginsberg, 1991).

The authors found a negative relationship between mothers' verbosity and complexity of speech. Mothers that talked more tended to use less prohibitives, Wh-questions, and directives per amount of talk. There was no indication of this in fathers' speech.

There were few differences in the speech of resident and non-resident fathers. This reflects data garnered from the interviews which showed no difference in the time either type of father spends with his child. However, non-resident fathers used less indirect prohibitives, indicating reluctance to use more blunt commands.

Conclusions

The greater differences in middle-class mothers and fathers indicates a gender-based contrast in communicative and parental goals. In working-class families, the lack of difference shows a greater similarity. However, there was enough difference to suggest that the Bridge Hypothesis is applicable to low-income families, albeit not quite as prominent as it is in the middle-class. This reflects the more recent role of the father as a co-parent rather than secondary caregiver.