LINGUIST List 5.6

Sat 01 Jan 1994

Disc: Internal/external evidence

Editor for this issue: <>

Directory

1. , Internal/external evidence

I think that there ARE different kinds of evidence and that
in particular Don Churma is right in saying that different
kinds of evidence are MORE or LESS telling with respect to
the question of what the native speaker knows. However,
I do NOT think that the conventional dichotomy of external
vs. internal corresponds to the desired distinction between
data which indicate something crucial about the organization
of the speaker's knowledge and those which do not.

Thus, data from so-called secret languages do NOT in general
constitute crucial evidence about how speakers organize their
internal knowledge, since in particular it can be shown that
many secret languages have long histories. Thus, their rules
may accumulate all kinds of historical detritus along the way.
Moreover, it is not always obvious which secret-language phenomena
reflect the supposed rules of the speaker's linguistic system
as opposed to being idiosyncratic rules of the secret language
which are learned as one learns the secret language. Before
we can conclude anything froms e
secret language data, we would need to know how the secret
languages are transmitted (see my forthcoming paper "L'arbitraire
de Chine", currently under review by a journal I guess I better
not name).

On the other hand, so-called internal (i.e., distributional)
data CAN, I think, be quite indicative of how speakers structure
linguistic information IF they bear on a wide enough variety
of languages (with the right kinds of test and control cases, etc.).
For example, Paul Kiparsky's well-known (I believe) observation
that we do not find, as a rule, languages in which phonological
rules are sensitive to whether a given feature is phonemic (distinctive)
or not provides, if true, the best kind of support for Halle's critique
of the classical theory of morphophonemes and phonemes.

Thus, while there is a distinction to be made between crucial and
merely suggestive data, I think it has little to do with the
distinction between external and internal. What do you all think?