ry-schachter1974reaction

Schachter, J. (1974). An error in error analysis ry - See also [|Young, R. . Error Analysis: Avoidance] which is based on Schachter's paper.
 * ry - How does Schachter's analysis support the claim that Chomskyan linguistic competence is needed to explain SLA?**

Wow. Either I’m understanding SLA better now or Schachter’s article was very well written and her argument for abandoning CA aposteriori is most convincing. Although I am not as familiar with all the details of her predictions for RFC areas of syntax difficulties between the chosen language pairs, I thought her logic was brilliant.

The question is “How does Schachter's analysis support the claim that Chomskyan linguistic competence is needed to explain SLA?”

If we accept the initial and even later principles and parameters refinement of the Chomskyan notion that humans have a language acquisition device which becomes less available once we reach a certain age range near puberty, one way that we can overcome the logic problem of known successful L2 language acquisition in spite of weak L2 exposure and little or no L2 instruction, is by taking advantage of our particular implementation of L1 knowledge and the underlying knowledge of language that we used to implement it. This, of course, is what Chomsky calls our “linguistic competence”.

Schachter is basically saying that contrastive analysis (CA) is insufficient to provide good guidance for L2 learners in its weaker or more accurately termed "CA aposteriori" (this name implies no judgement calls) version. In effect, Schachter uses her own linguistic competence (and logic!) with syntax and phonetics to show that CA aposteriori analysis will not only miss some of the difficulties certain L1 learners of a given L2 = English language will have, but CA aposteriori can also invite the wrong conclusions.

She uses principles of linguistic competence to predict which L1 to English L2 learners will have difficulty with subordinate postnominal relative clauses and then makes it clear how aposteriori CA only evaluates production while apriori CA includes both production and comprehension in its predictions. Combining this with the observed 50% or so less relative clauses produced by Chinese and Japanese speakers, she makes the believable deduction that these L1 speakers do not create these clauses because of their fear of making an error and because they can paraphrase their way out of them and still communicate. And why should they avoid them more than the Persian and Arab speakers? Because they don’t have postnominal relative clauses in their languages and therefore would not be able to use language transfer with their English. The Persian and Arab speakers, on the other hand, have this in their languages, so they are more willing to use these phrases and attempt to apply syntax rules from their respective languages to English. Hence, more phrases are attempted *and* more errors are committed as they overgeneralize on the rules. This resonated with me because I am aware that I produce similarly when trying to converse in French. If the speed of the conversation is at stake, I’ll tend to stay in the present and not stray much into passé composé, or other more involved grammatical constructions.

Schachter impresses further in noticing that most CA analysis involves phonetic research where phoneme avoidance à la "alleged" syntax avoidance by the Chinese and Japanese speakers would be extremely difficult if not impossible. Just think, she says (I’m paraphrasing), of the difficulty speakers who cannot pronounce “th” would have if they try to, on-the-fly, come up with synonyms which mean the same thing but have no “th” in them. The processing burden is such that the “th” words could not be avoided and the speakers would make the necessary observable mistakes to allow good error analysis. Based on aposteriori CA, because there would be less phrases to evaluate, one might think that the Chinese and Japanese speakers had mastered these phrases, which would be the exact opposite of what the situation most likely was.

So Schachter does appeal to linguistic competence to explain SLA. There’s more to SLA than just production. Comprehension is involved as well. We need to take into account the understanding of language that we use to produce utterances in a given language. Aposteriori CA only recognizes production. Apriori CA does not distinguish between comprehension and production and therefore uses both.