Baby Talk and the Origin of the Word

Peter F. MacNeilage & Barbara L. Davis

University of Texas at Austin

(macneilage@psy.utexas.edu)

Falk (Behavioral and Brain Sciences [BBS], in press) offers a plausible
scenario regarding the communicative context for the first spoken words.
Early hominin mothers, she argues, needed to put down their babies, who had
become unable to cling to them, while foraging. The resultant need for
parental care at a distance created selection pressures for elaborating the
dyadic vocal communication pattern, thus helping to account for the sheer
volubility of modern hominins compared with their nearest relatives. The
repeated occurrence of particular vocal episodes within the mother-infant
communicative dyad could have induced a participant to link a specific
vocalization with a recurring aspect of the context, as a word requires.

We note conducive characteristics of the mother-child dyad: (1) a
non-threatening environment, with intimacy allowing structured
communicative modes, and (2) highly focused attention. The communication
patterns would be, as they are now, stereotyped, simple, and relatively
small in number.

Sound patterns of first words, perhaps formed in this context-baby talk
words-may have shared two properties of speech that infants and languages,
including Creole languages, share today-the universal consonant-vowel (CV)
form and a biomechanical constraint on amount of intrasyllabic tongue
movement (MacNeilage & Davis, BBS, in press).  This constraint takes three
forms: Coronal (or tongue-front) consonants co-occur with front vowels,
dorsal (or tongue-back) consonants co-occur with back vowels, and labial
(or lip) consonants co-occur with central (neutral?) vowels (MacNeilage &
Davis, Science, 2000).

Falk suggests that  nasal demand-vocalizations, (e,g, "m-m-m-") observable
today when the infant is separated from the parent, (Goldman, J. Child
Language, 2000) may have given rise to a first word.  The mother could have
surmised "This sound stands for me." Consistent with this proposal, all the
baby-talk words for "mother" in a corpus of 6 languages presented by
Ferguson (American Anthropologist, l964) have nasal consonants.
Strikingly - and this suggests a systematic cross-linguistic communicative
contrast in baby-talk words-all the baby-talk words for "father" have only
oral consonants. But whereas the nasalized demand-cries of modern infants
occur in the prebabbling stage and are consequently quasi-steady-state
vocalic forms, baby-talk words feature the CV syllable and the
biomechanical constraint against intrasyllabic tongue movement (hence
"mama" and "papa"). Nasality (and orality) might have been incorporated
into protosyllabic CV strings in the first baby-talk words because these
cyclicities may have constituted the core of the already existing
vocal-grooming repertoire postulated by Dunbar (BBS, l993). The implication
of this is that the first imitation in the evolution of words was a partial
imitation of infant vocalizations by the female adult-an imitation of
nasalization incorporated into CV forms, in what was, in effect, a
self-naming operation.

The likelihood that baby-talk forms were a source of parental terms in
languages proper is indicated by Murdock's finding, in a corpus of 474
languages (American Anthropologist, l959), that the first syllable of 78%
of maternal terms had a nasal consonant, whereas the first syllable of 66%
of paternal terms had an oral consonant.  Predictably, from our standpoint,
these first (predominantly CV) syllables of parental terms also have the
intrasyllabic biomechanical constraint on tongue movement.
If parental terms in languages have additional phonetic properties of
baby-talk words, this would provide further evidence for a phylogenetic
link between parental terms in baby talk and in language proper. The
baby-talk forms, like other infant vocal forms, tend to involve syllable
reduplication (e,g, "mama"), whereas languages proper avoid it.  Thus baby
talk has intersyllabic as well as intrasyllabic biomechanical constraints
on tongue movement.  That is, the tongue tends to maintain its position
across as well as within syllables.  Intersyllabic biomechanical
constraints are absent in modern languages, in which successive syllables
tend to be different.  We are currently investigating the hypothesis that
parental terms in languages tend to have more syllable reduplication than
other words of languages, and, unlike other words of languages, have the
intersyllabic biomechanical constraint on tongue movement, as well as the
intrasyllabic constraint.