clark comments on grammar

The discussion of this, based on the WS forms, was a bit vague, abstract even, and really didn’t give anyone who didn’t already know about two-word combinations and the longer sequences that follow the two-word ‘stage’, what kinds of utterances you were talking about. Maybe some illustrations, from selected languages, of the kinds of utterances in play here, examples from the CDI forms themselves? (I know you have some forms in the tables, but some examples in the text might be worthwhile.)

Also, you need to add some caveats about how consistent children actually are, when learning a word-order language, in observing conventional word orders: Bowerman (1973a, b) noted a number of violations (OV, SOV, OVS, etc.) in children acquiring English – and I have similar data in my diary notes (also from our 2-year-old granddaughter, who goes from OV to OVS, to SVVO, and so on, in successive utterances; also true for early word orders in French, over and above the flexibility found in colloquial spoken French. This tends to suggest that even children learning a language that marks grammatical relations via word order at first make use of word order for pragmatic purposes: to emphasize new information, for example.

What seems critical here is the semantic relations being conveyed at a point when children don’t yet know how word-order conveys grammatical information. Of course they will get word order ‘right’ on some occasions in their very short utterances, but it is certainly not as generally observed (pace Susan Goldin-Meadow) as has been claimed. For languages that are highly inflected for case on nouns (anywhere from 2 cases to over 20, depending on the language, as well as gender and number), and for verbs – tense, mood, and aspect as well as gender and number, these are the elements that typically convey grammatical relations, and word order is used to carry pragmatic information—what is given vs. new, for example.

There are also some questions about the earliest inflections added to words on the CDI forms: How many early uses of added morphemes and of irregular past-tense forms in English, for instance, are actually formulaic (e.g., words that are always produced in plural form), and so effectively unanalyzed in these early productions? Take some of the irregular past tense forms, where longitudinal observations have shown that children often treat forms like fell or went as stems; they then use these stems with other inflections, as in they felled down, Daddy boughted that, or she’s wenting, etc., for some time before they arrive at the actual conventional analysis of such forms as marking ‘past’ (or perhaps ‘non-present’). This makes a difference to assumptions about what is learnt when, and casts some doubt on the rather general view that frequent irregular past tense forms, for example, are learnt first as past tense forms, with regular, and less frequent by type, past tense verb forms being learnt only later, along with subsequent (over)regularization of irregular verbs, and all this before children finally acquire the conventional paradigms of English (or some other language). Just a note of caution to add here.

In 13.2, Tables 13.1 and 13.2, there seem to be different (and different numbers) of languages being sampled from. In one listing you have some 10 languages (plus Mandarin with zero entries, for an eleventh), but later you have only 6 languages represented (with 2 variants of one of them). So I found I was not always clear on just the comparisons were here. And in Figure 13.7 (age effects of form, morphology and syntax), I am not sure why there is (a) such a gap between the Australian and US English data; (b) no information on complexity for French (nor for Cantonese, Mandarin, Korean, and Hebrew); (c) no syntactic information for Slovak; and (d) no morphological information for Mexican Spanish. Are there other references you can give that would help make up for these gaps in the data? (Rojas-Nieto, also Maldonado, have good data on Mexican Spanish; and there is a lot of data around on French [in France], and a lot on Hebrew––especially in the work of Berman, Ravid, etc. – I can supply further references here if you want).

langcog / wordbank-book

clark comments on grammar #64