Just as we cannot identify the Old School with 'quantitative methods', but
only with a particular kind of an way of using what are mostly
quasi-quantitative qualitatively-based methods, so we cannot identify it
with expository argument discourse genres, or the New School with narrative
ones.
For starters, narrative is not a genre, but a large class of loosely
related text-types, and cross-culturally (including subculturally, and so
among nearly all of us as subcultural hybrids) there is a lot of variation
in even core notions of narrative text organization, and along many
dimensions, formal and semantic. Narrative is more like a meta-texttype, a
sort of grand category of discourse form that stands in contrast to not
only analytical-conceptual-expository-argument form, but also to
conversational discourse form, and to lyrical-poetic discourse form. Some
people would also include a dramatistic discourse form at this level of
categorization, though I tend to think of it as a hybrid of narrative and
conversational forms. These are 'ideal-types' in the Weberian sense (i.e.
idealized abstractions); any real narrative of sufficient length is likely
to incorporate features, if not actual passages, or each of the other
forms, and so for all the others.
These very general discourse types are not really very useful as analytical
categories because they do not co-vary with much else of interest. A
semantic analysis of register (the frequency or probability of various
relevant lexical items or grammatical constructions and their typical
combinations), or of genre (how the semantic register varies systematically
as we proceed linearly through the text), is highly multi-dimensional and
produces little in the way of convincing categories for real texts much
above the level of much more specialized genres and text-types, which do
co-vary with other aspects of a network of social practices (situational
context, institutional context, individual and communal activity, etc.). I
do not really believe that one can say what 'narrative' as such is
functionally specialized for, though it is possible to contrast HOW each of
the grand discourse types performs various common discourse functions.
Narratives, poems, conversations, and explanations all make
generalizations, but they do so in characteristically different ways (even
though they can and do borrow each other's typical strategies). They can
and so all get very concrete and specific about viewpoint, values, actions,
and participants, but again in typically different ways.
There does not seem to me to be any motivated reason to limit research
discourse to any of these types; we have seen all of them on xmca. Nor is
there any reason why 'theoretical' discourse (a function) should be
confined to any of these _forms_. What makes a discourse 'theoretical' is
that it contributes to either a deeper understanding of a particular case
or to a wider understanding that relates some case to one or more others.
It does not need to be 'abstract' or 'impersonal' or 'universally
generalizing' to be theoretical in function. Those views belong to the Old
School.
It is very hard to innovate a discourse form that does not appear to be
merely a mixture or recombinant of the discourse strategies characteristic
of these four grand types; among them they pretty well span the discourse
repertoire, if only because they do the basic things people do with
language. Discourse innovation happens at more specialized levels of
differences among text-types, and in how these text-types are integrated
with other aspects of social and ecological activity.
There is no reason at all why any or all of these forms cannot be combined
in research discourse, either empirically descriptive discourse, or
theoretical-analytical-explanatory-interpretive discourse (not mutually
exclusive). But the combinations need to be motivated by discourse
functions, by what we are trying to do. Research discourse seems to me to
lack integrity if it merely juxtaposes Old School and New School genres to
please both sides, but without any functional integration or sense that the
whole becomes more than the simple sum of these parts. Integrity here does
not mean 'coherence' in the Old School sense of conformity with some one
traditional genre; I am quite happy with postmodern 'fragments' forms, or
hypertexts, so long as more meaning can be made this way.
I certainly do not believe that narrative forms are inherently more 'emic'
or more 'ethnographically valid' or even more 'phenomenologically
experiential' than the other three. One 'virtue' of catgories at that level
of abstraction is that they are (trivially) valid across cultures
(interesting cultural differences are at more specific levels of analysis):
everybody makes generalizations, does 'abduction', tells recounts or
stories, uses metaphors, exchanges information, negotiates social
relationships, in talk. Everybody uses some sort of 'theoretical discourse'
as well as narrative discourse in daily life, including in the construction
of identities and values. You do not get closer to members' perspectives by
narrating them; you may distort them by not also projecting them as
argument, generalization, metaphor, description, explanation, exchange, and
negotiation. By the same token, you certainly do distort them if you
confine yourself to a specific 'scientific' discourse form of your own
culture in speaking 'of' them.
This one is getting rather long, too; so, I'll write elsewhere about the
original issue: making sense of multiple scales of process in the analysis
of individuals and communities. JAY.
---------------------------
JAY L. LEMKE
CITY UNIVERSITY OF NEW YORK
JLLBC who-is-at CUNYVM.CUNY.EDU
---------------------------