I think the phenomena you describe do indeed make problematic many present
assumptions about 'proficiency'. These assumptions, not surprisingly, are
ones that arise generally in the psychometric paradigm: that behavior is
governed by internal factors ('abilities' 'competences'), that these are
relatively situation-independent, and that they are stable and measurable
'traits' of individuals. It seems likely that all these assumptions are
seriously misleading. We are more or less sure today that this is the case
for 'intelligence', and for various sorts of 'problem-solving', and it
seems likely to be the case also for 'language proficiency'.
How can we reconceptualize these notions so that better approaches to
academic counseling and guiding students' language development can be
developed?
Perhaps a start would be to look at the kind of data your work points to as
most directly relevant: language-using habits and strategies in the
relevant settings. How much do students use the L2? for what communicative
(and other) functions? with how much success by the various relevant
criteria? How do these habits and strategies tend to accelerate or inhibit
processes leading to change in language-use patterns?
I certainly know from my own experience in using languages other than my
first (English), that the initial stage of learning is the acquisition of
at least one functional strategy for each communicative task I encounter on
a regular basis. Some of that can be learned outside contexts of use, but
then must be adapted to them. The contexts, moreover, set the priorities
and perhaps the sequence for acquisition. Once I have a minimal functional
repertoire, the main factor in my communicative effectiveness seems to be
how willing and able I am to mobilize these strategies adaptively for
various purposes. At this point inhibiting factors (fear of errors, of
being misunderstood or misjudged, of embarassment, etc.) play a significant
role. If I've been drinking a little, for example, I become much more
fluent, I make errors, but my communication seems to function pretty
successfully. I also quickly find myself in situations where others can
correct me, or where I notice my own errors and either self-correct or make
mental notes to look up words, grammatical conventions, etc. When I allow
inhibition to dominate, my rate of progress in the language slows to a
standstill, so far as language-in-use goes, though I may continue to learn
vocabulary and grammar from textbooks or through reading and using a
dictionary. Language-in-use has its own logic and trajectory of
development, quite apart from that of knowledge about the language. The
latter can enable the former, but need not do so. The former can easily get
rather advanced with minimal increase in the latter (because there are so
many periphrastic ways to say anything).
So from the point of view of assessment and prediction and advice regarding
what students are ready to handle academically, 'decontextualized' (i.e. in
artificial contexts) assessment methods focussed on knowledge of the
language, or even on reading comprehension, miss the perhaps more important
factor of where the student is on the trajectory of improving
language-in-use. One would have to assess each major context of use
separately (reading, writing, conversation, intensive listening and
notetaking, etc.) as well as knowledge about the language (vocabulary,
grammar, genre patterns, etc.).
A particularly important factor here is the way in which 'live' settings
trigger behavioral responses in a way that simulated ones may not. A
pencil-and-paper test that says: Imagine you are being interviewed for a
job and the interviewer asks you "..." What would you reply?" will not
elicit the same behavioral responses as the real event. Even a video or
computer simulation might be only slightly better than the paper test. And
clearly very important for the differences are matters of interpersonal
'face', social inhibition, power and respect relationships, interactional
synchronies, etc.
Is there a research literature investigating such matters? it would appear
that such work is critical for any attempt to get beyond 'proficiency'
measures that are all but useless in practice, and are probably
particularly unfair as well as inaccurate across cultural differences.
Absent context-relevant validity, most such tests merely privilege good
test-takers (i.e. those most socially and culturally like the test designers).
JAY.
---------------------------
JAY L. LEMKE
CITY UNIVERSITY OF NEW YORK
JLLBC who-is-at CUNYVM.CUNY.EDU
---------------------------