data and analysis

[Arnold M. Zwicky]

On Jan 28, 2006, at 4:15 PM, RonButters at aol.com wrote:

>
> In a message dated 1/28/06 6:33:52 PM, zwicky at CSLI.STANFORD.EDU
> writes:
>
>> ... i'm not promising an analysis; i don't have a lot of insight into
>> my mental processes.  i'm  giving you an account of how i understood
>> the utterance in the context provided.  period.
>>
>> it's like a grammaticality judgment.  it just is.  i can try to
>> figure out the system of grammatical knowledge that gave rise to the
>> judgment, but that's not the same thing as analyzing my mental
>> processes, which are not available to me directly.
>>
>
> Surely it is not only "grammatical" knowledge that affects one's
> judgment about interpretation--even grammatical judgments ("The
> horse raced by the barn door fell" and all that). If all we looked
> at were bald judgments, we wouldn't get very far in the analysis of
> conversation, for example.

i didn't say it *was* a grammaticality judgment; i said it was *like*
one -- a bald report on how i understood something.  here i'm
behaving just like any speaker, who has no access to the processes
behind the mechanisms of understanding.  of course, linguists can
(and should) examine systems of grammatical knowledge and systems of
language use; there are ways for inferring what these are like from
the various sorts of evidence available to us.  but introspection is
not a good source of such evidence.

i happen to be a speaker of english who finds things like "She's
going to San Francisco and talk on firewalls" entirely acceptable.  i
can tell you about that judgment without having a clue as to *why* i
accept it (many others do not).  discovering what the system is like
requires looking at a pile of other examples (and my bald judgments
on them), seeing what turns up in corpora, investigating other
speakers, and so on.

>> i will, eventually, i hope, post about the larger issues.  and point
>> out (surprise surprise) that unmanliness, femininity, and (male)
>> homosexuality are connected in complex ways.
>
> Arnold and I agree about a lot of things, and the complexity
> hypothesis suggested here is certainly one of them. But can one get
> to it without atempting to discern "a lot of insight into
> ... mental processes"?

these are really very different sorts of questions.  yes,
investigating the nature of the social order (and differences between
groups and individuals as to the details of this social order)
requires that we get some insight into the mental worlds of people.
but these worlds are not accessible via simple introspection, and in
fact introspection usually produces epiphenomenal accounts that
reflect explicit/overt beliefs and tend to be seriously misleading.

arnold

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org