[Wittrs] The mythical "chomskyan refutation of behaviorism"
- From: gprimero <gerardoprim@xxxxxxxxxxxx>
- To: wittrs@xxxxxxxxxxxxx
- Date: Sun, 27 Sep 2009 21:29:30 -0400
(Josh) Skinner has a bunch of core beliefs which exemplify behaviorism. These
are interesting, but after due consideration, were destroyed by Chomsky and
became passe in the 1960-1970 timeframe. Nothing you (or anyone else) have
posted changes that a bit.
(Gerardo) What´s the "great argument" of Chomsky about behaviorism?
I´ve just found the following arguments in Chomsky´s review of Skinner´s Verbal
Behavior:
* Skinner´s proposals have not been tested ,so they should be discarded.
* Skinner´s terms are merely a paraphrasis of more traditional terms.
* Language is too complex, so it requires a mediational theory.
* Language is too complex, so it must be innate (poverty of stimulus argument)
Are they so good?
* Skinner´s proposals have not been tested, so they should be discarded.
It´s a non-sequitur. Skinner´s book is a research program, and it´s not
reasonable to claim that it´s a "failure" just 2 years after the publication.
Nowadays there´s abundant evidence in favour of Skinner´s proposals (Sautter &
LeBlanc, 2006), and other theories have adopted its empiricist research
program, its general learning mechanisms (López Ornat, 2004; Elman et al., 1996
) and its functional approach (Andresen, 1991, López Ornat, 2004; Tomasello,
2000). Chomsky´s strong nativism has shown to be a dead end that generated more
problems than solutions, and even cognitivists have abandoned it (cognitive
linguistics is anti-chomskyan, and has adopted some of the proposals of
Skinner, e.g. "frames" and operant learning).
* Skinner´s terms are merely a paraphrasis of more traditional terms.
No, they´re not, it´s simply that Chomsky has distorted them. For example,
Skinner never considered reinforcement as a voluntary praise of the child´s
utterances, as Chomsky suggests. The same can be said about the concepts of
"stimulus class" and "probability", that Chomsky distorts for attacking a
strawman. When Chomsky finds a discrepancy between Skinner´s terms and
traditional ones, he complains about such discrepancy, instead of doubting his
assumption that they´re merely a paraphrasis.
* Language is too complex so it requires a mediational theory.
Complex phenomena can be explained by the combination of simpler mechanisms
(e.g. natural selection, neural networks). The relevance of neural and genetic
events is not denied by Skinner, and this relevance doesn´t imply the
acceptance of Chomsky´s nativism.
* Language is too complex, so it must be innate (poverty of stimulus argument)
The evidence of universal grammatical features doesn´t imply that they´re
innate: grammatical simmilarities are to be expected if there´re innate general
learning mechanisms.
If every complex behavior must be innate, then almost everything must be
innate, which is implausible.
The solution is a distinction between data an evidence: negative data, positive
data, and non-occurrence can be a positive or negative evidence, depending on
the expected input (see Fiona Cowie´s book What´s within: nativism
reconsidered).
The child receives abundant negative evidence in the parent´s repetitions,
ampliations, and the non-occurrence of expected utterances (Cowie, López Ornat).
The structure dependence is to be expected if (as Skinner and Tomasello
proposed) the child doesn´t begin with the learning of grammatical classes but
with the learning of frames (e.g. "Is X singing?", "X, who is Y"). Therefore,
the structure dependence ("Is the man who is happy singing?") is simpler than
non-dependence ("Is the man who happy is singing?") because it combines two
previously learned frames ("Is X singing?", "X, who is Y").
There´s evidence of general statistical learning mechanisms (Saffran, 1996;
Maratsos, 1982).
Conclusion: the so-called "Chomsky´s refutation of behaviorism" is only a myth.
Chomsky just attacked a strawman with unsound arguments. The appeal to this
myth must be replaced with a rigorous analysis of the soundness of each side´s
arguments. The empiricist behavior-analytic research program of Skinner is
alive and well.
You can see Fiona Cowie´s talk about language learning in:
http://www.youtube.com/watch?v=QSLlMlbhIqo
Regards,
Gerardo.
WEB VIEW: http://tinyurl.com/ku7ga4
TODAY: http://alturl.com/whcf
3 DAYS: http://alturl.com/d9vz
1 WEEK: http://alturl.com/yeza
GOOGLE: http://groups.google.com/group/Wittrs
YAHOO: http://groups.yahoo.com/group/Wittrs/
FREELIST: //www.freelists.org/archive/wittrs/09-2009
Other related posts:
