limits of IRF

Jay Lemke (JLLBC who-is-at CUNYVM.CUNY.EDU)
Sat, 30 Dec 95 00:40:22 EST

----------------------

I'm not going to specifically debate the pro's and con's of
IRE/IRF further. I think the major points one needs to consider
have been made, and people can judge how specific dialogues are
functioning in particular contexts. I do teach triadic dialogue
as a form to new teachers (in science, in secondary education)
because it is pervasive and they need to know how it works and
some of what it can do, both helpfully and not-so-helpfully. I
try to stress that they should be diverse and not use it
exclusively, and that when they do use it, they should choose
their Initiating questions carefully to open up where possible a
larger range of possible responses (a very middle-class discourse
strategy in this context I think), and to use their Follow-up
move, as Gordon indicates, for a variety of functions, including
encouraging participation, commenting on student answers
evaluatively in multiple respects (conceptual, logical, verbal --
but not all at once!), and developing thematic connections by
elaborating the answers (or actually the Question-Answer pair's
thematics). I also discuss the many variations on the form.

For Eugene and anyone else who'd like to look into this, _Talking
Science_ (1990, Ablex Publishing, 355 Chestnut Street, Norwood,
NJ 07648-2090; 201-767-8455, orders) is still available in its
most recent printing, in softcover (unless you or your library
can afford hardcover prices, also available).

The original work on the IRF by Sinclair and Coulthard (who were
interested in a discourse grammar for a very regular register of
talk, rather than specifically in education) posited many
possible realizations of Initiations and the other moves, not
just Questions. Questions as a semantic class, of course, are
also not always grammatically realized by interrogative clauses
(or by clauses at all). Bud Mehan's classic _Learning Lessons_
explores some of the educational uses of IRF, especially the
interactional-social work done through these exchanges, and my
own early studies tackled particularly the thematic content
development (though we all know these are too interdependent to
cleanly separate, and _Talking Science_ deals with both aspects).

The characterization of the near-peer dialogue (recently cited
again by Eva) about classifying things in different ways as IRF
seems to me to fall somewhere near the outer limits of what can
usefully be accounted functionally or formally similar to the
prototypical teacher-student Question-Answer-Evaluation triad
which motivates the classification. The cited text does not
strike me as significantly outside the register of casual task-
cooperative conversation, though in context it may continue a
pattern that can be usefully seen as specifically IRF. When we
get this far away (in a semantic topology) from the typical
instantiations of the form, I am quite prepared to accept that
many of my objections to IRF forms may not apply. But I would
still like to check and see whether or not in each instance they
still do.

Eva then goes on to cite her 'broken toys' dialogue, which
_could_ also be analyzed as IRF, but I would probably not find it
a particularly useful application of the tool in this case. IRF
in Sinclair and Coulthard can afford to be rather general,
because it is very carefully specified by its possible
realizations by classes of utterances. If we take it so far that
any B-utterance (Response) which forms an interactional and
thematic relation to a preceding A-utterance (retrospectively
then an Initiation), and is relevantly continued by a second A-
utterance (Follow-up), then it is scarcely more specific than
general definitions of cohesive 2-party dialogue. A lot of
analytical experience tells me that even trying to define a
segmentation pattern like ABA-ABA-ABA could be problematic (since
the A-A turn is rarely cleanly decomposable by semantic function
into, say, A(F)-A(I), except when we have the most typical or
canonical realizations). This is probably a more technical point
than would interest most of you, but I make it to show that
analytical value falls when specificity of instances (or narrowly
defined classes) is subordinated to generality of theoretical
tools.

"Who is doing what?" certainly still seems to me a useful
question to address to Eva's dialogue, however. My culture makes
salient for me some work of identity-autonomy-competence in it
(of the 'I can do it myself, Mother!' sort), as well as the
efforts to transmit norms (loyalty/keep, practicality/throw) and
the thematics for adjudicating them (broken: a little; can play
vs. not), and I would probably bring to bear tools (not just
local-semantic, but also intertextual and even social structural)
to explicate what and how in relation to these hypotheses --
hopefully discovering along the way many other things also being
done there.

Perhaps most interesting to me though was Eva's observation that
something like the triadic form prevails even though a book has
been read together many, many times, and core thematic knowledge
is not at stake. Of course we should see a 'phatic' function
here, a maintaining and strengthening of interpersonal bonds in
and through this activity. Perhaps also a continued practice at
the _form_ of the triad, as cultural capital for later schooling,
along with (I suspect) many variant functional uses of the form.
It would be very interesting to see a longitudinal study of re-
readings together of the same book by the same pair: what
changes? what stays the same? what trends? and why? ( -- and
please don't forget the _pictures_!)

JAY.

JAY LEMKE.
City University of New York.
BITNET: JLLBC who-is-at CUNYVM
INTERNET: JLLBC who-is-at CUNYVM.CUNY.EDU