Kathleen Currie Hall

Notes on: Stoel-Gammon, Carol. 1998. Sounds and words in early language acquisition: The relationship between lexical and phonological development. In Paul Rhea (ed.), Exploring the speech-language connection (Vol. 8, pp. 25-52). Baltimore: Paul H. Brookes.

Summary:
In this chapter, the author outlines some of the research that has been done on the connections between a child's phonological abilities and his lexical development.  She shows that early phonological ability is correlated with larger and faster vocabulary development, and that frequency of occurrence in the lexicon being developed correlates with articulatory ability.


Details:
1. Sounds that are "practiced" a lot (e.g. in babble) tend to be used in first words (though not always).

2. Earlier age of babbling correlates with earlier developmet of speech; larger numbers of vocalizations correlates with larger vocabulary sizes; larger numbers of babbled syllables correlates with earlier age of first words; larger numbers of syllable and sound types early on correlates with much later language tests.

3. There is also evidence that a child may acquire lexical items first that seem to match the child's "personal preference."  

4. There is a "typical" development pattern for order of acquisition of particular sounds (see p. 33-34).  There is also a "standard" list of words that 50% of 30-month-old children have acquired, which Stoel-Gammon treats as a "target lexicon" for acquisition.  She shows that there is a correlation between the sounds acquired early and the frequency of words starting with that sound.

5. Children with larger lexicons tend to have larger inventories of speech sounds and syllable types.

6. Delay in talking doesn't seem to correspond to phonological deviance -- late talkers are simply that; they behave like typically developing children at a younger age.

Questions:
1. While Stoel-Gammon does acknowledge the illegitimacy of this at the end of the article (p. 49), I'm concerned by her tendency to conflate correlation and causation.  For example, p. 28 -- perhaps there is some external factor that makes certain children "better" language learners than others; this means they are more likely to babble earlier and more, and more likely to talk more.  There is no a priori need to posit that "babbling serves as a foundation" or that babbling provides the child with "building blocks" for later speech.  Have there been studies / effective speech treatments that kids are for example taught to babble more and then show actual improvement in their speech abilities?
2. p. 30-31 -- Regarding the notion of a child's "personal preference" -- why assume that the preference is caused by the child's ability to produce the preferred sounds?  Alternatives: perhaps child finds a particular sound perceptually intriguing, and then practices it and uses it in words.  Also mentioned -- frequency in input?  Have there been studies on twins to see if they tend to have the same personal preferences?
3. p. 32 -- The child who moves all fricatives to the end of a word -- seems like a very strange / unattested "sound change" in natural languages, but clearly possible in a human phonology acquirer.  What does this tell us about phonological "naturalness" and possible rules?
4. p. 34 -- "Children must eventually attempt to produce words with phonetic features beyond their capabilities if they are to acquire new words."  Why?  Isn't it hypothetically possible that children just continually extended their babbling to include new phonetic features and didn't try new words until the features had been acquired in babbling?  What does the fact that children DON'T do this tell us about language acquisition?  Just that children are interested in acquiring the sound-meaning pairing?
5. Stoel-Gammon acknowledges that it would be nice to compare acquisition effects with frequency of occurrence in the ambient adult lexicon (p. 50).  However, I kept being confused by her use of "target lexicon" to mean the 598 words that 50% of 30-month-old kids have acquired.  While I see her point that the frequency of sounds in these words mirrors the order and strength of acquisition of individual sounds occurring in those words, I wonder how much overlap there is in these claims; i.e., how tautological are her claims?  If a child is said to have acquired a phoneme when it occurs in words, why should it be surprising that sounds acquired first are those that appear in early words?  (Is it just the _frequency_ of words with a particular sound that is important here?  That is, is there ever a sound acquired early on that is not used in words spoken by the child but is highly frequent in the "target lexicon"?)
6. p. 47 -- What is meant by "accuracy of production was not considered"?  How far was this taken?  E.g. if a child produces [dUk] for [trUk], does he get counted as having a [tr] cluster in his lexicon?
7. p. 47 -- "Precocious talkers at age 20 months had larger phonetic inventories than the typically developing children at age 24 months" -- but at 20 months, the p.t.s had vocabularies almost twice as large as the t.d.s -- why make this comparison??  Why not compare p.t.s at 20 months with t.d.s with vocabularies of equivalent sizes?