RT list: Procedural-conceptual: arguments from translation, language learning etc.

Procedural-conceptual distinction:

A quick response to the arguments (discussed in the Minh Dang/Christoph
Unger exchange) from the alleged difficulty of raising to consciousness,
foreign-language learning or translation (for the identification of
procedurals vs. conceptuals).

(1) These arguments seem to be rather circular inasmuch as procedural
instructions are by definition harder to bring to consciousness than
concepts, which have declarative content.

(2) I know of no evidence that (primarily) procedural expressions are harder
to learn in a foreign language. They probably seem so when we try to teach
them through declarative definitions rather than through numerous examples
of use (i.e. through acting out procedures, as in learning to ride a
bicycle).

(3) Nor do I know of evidence that procedurals are harder to translate than
conceptuals; they may each pose problems for different reasons; expressions
do not match exactly across languages except possibly for a limited class of
closed referent terms like proper names and technical terms.

Put differently, these arguments seem to me to reflect a confusion between
language as object of description or classification (as for language
teaching or compiling dictionaries), and language in use (as in reading,
writing, translation, and especially, conversation or interpreting).
Definitions and paraphrases of expressions are declarative artifacts
isolated from context for specific purposes like language teaching or
dictionary compilation. In translation (but especially interpreting, as in
conversation), they rise to consciousness only in case of a problem (in
written translation, especially e.g. Bible translation, there is more
reason/leisure to linger on expressions, with the concomitant danger of
lapsing into artificiality).

(4) So a revision does seem necessary. From the viewpoint of a 'user' of RT
(for teaching interpreting in my case), conceding that 'procedurality' or
'conceptuality' must be 'smeared' across discourse, rather than neatly
dividing it into categories of words, obviously complicates things for
explanatory and pedagogical purposes, but seems more convincing. As Deirdre
Wilson said in a conference in Geneva (during a discussion of the
implications of Carston 2002, in particular) better a complicated but
realistic theory than a neat and handy one that doesn't fit the facts.

            Which way should we go? Language processing in normal use seems
to me to be overwhelmingly procedural; computation uses representations, not
the reverse. So I would tend to prefer Dan's approach: all expressions are
procedural, some with more or less conceptual content. This revision would
at least not entail throwing out the computational-representational
distinction, which is of great explanatory and pedagogical value - even if
this now makes it much harder to tie operations in discourse processing to
items in a corpus.

Robin Setton

Conference Interpreter (AIIC)

OECD / GIIT, Shanghai International Studies University

Please note my new email address: <mailto:orobinsetton@gmail.com>
robinsetton@gmail.com
Received on Wed Jan 9 04:10:21 2008