Narratives v/s Grids

Hemant Desai (
Thu, 28 Jul 1994 01:54:53 -0500 (CDT)

To continue the interesting thread on narratives v/s grids --

Jim Mancuso wrote recently:

>> "The move to use a grid-type analysis to look at the sources of
a person's narrative needs to be considered rather thoroughly. We need to
explore the potential of thinking of grids in that way. <...>
If we think of narrative as a construction of a construction process,
we can claim that narrative may be construed as having elements, just as
does any construction process. We can think of those elements as "slots"
in the narrative. As one constructs a narrative to guide his/her self
functioning he/she draws from his/her hierarchically organized system
of two-poled constructs to fill in those slots. A grid approach could be
regarded as having utility in extracting the constructs which a person
retrieves to build a narrative/construction. <...>
Thus, the use of a grid-like assessment device, particularly when
the participant's responses are assessed relative to the response patterns
within the grid, can provide formalisms by which to express the meanings
implicit in the responses in terms of the context provided by the overall
personal construct system.
In short, I believe that as we work out the available approaches,
we will find useful ways to explore the construct systems which persons use
to "fill the narrative's slots" as they construct the narratives which are
used to frame the flow of putative inputs.
And, I believe, we need not shy away from using supplied constructs
in order to explore the ways in which the respondent organizes those
constructs within the system of signs he/she uses to respond to the
constructs which are supplied". Jim Mancuso <<
Rue Cromwell, on the same subject, added:

>> <...> "My own view is that the two groups need each other more than they
realize. Much conceptual development is yet needed in PCP, and I know that
Kelly himself would be the first to applaud and to take pride in building the
platform. Rep grid methodology will not have achieved any substantial
contribution until it is able to subsume the "narrative." The "narrative"
approach in PCP will not have achieved any substantial contribution until it
is also able to subsume the quantitative algorithms of the rep grid and their
recent analytic innovations. <...>
As for rep grid science and narrative, the rep grid is essentially at
the point of dealing with narrative constructions limited to the verb "to
be." Very very little (if any?) is done in the way of rep grid matrices
which deal with "subject-action verb-object" constructions and the
overarching sequential linkings which glue them together into a human or
science story. Call it "antecedent-consequent" or "cause-effect"
constructions for a start. As for narrative type PCP scientists/clinicians,
I likewise see no movement toward consuming the elemental constructive
structures. After all, a lot of people on this side just don't like
numbers!" <<
You said it, Rue. I personally don't have a problem with numbers as long as
they FOLLOW qualitative and idiographic methods rather than preceed them.
Suzanne Huffman joined this thread and made an insightful comment:
>> "Rue, there are a few of us trying to move beyond "to be" - I ask for
stories, use them as elements in grids and do narrative analysis. Pat
Diamond talked about astronomy some, I'll share how I use this metaphor. My
MDS plots are astronomy, locating construct and element points in
experiential space. The stories are involved in the astronomy - naming the
constellations and understanding the personal myths involved in them. This
is confusing to many researchers - if you have numbers, aren't you doing
hypodeductive work? For me, no - I'm using a statistical procedure as a
hermeneutical tool, as Diana Taylor recommends." <<
This dialogue prompts a possibility. I am teaching undergraduates and
I would very much like to obtain written texts from them to understand their
constructs/counstruals in some particular domain. However, there is a
technical difficulty -- that of analysis. Traditional manual coding
(Alphabase, Ethnograph, etc) of qualitative data is not set up to find
semantic differentials in a pcp-type of way. However, (and I'm guessing
there are some computer/AI folks out there who are willing to help), if a
software program could classify words from a text, rank them on the basis of
complexity (e.g., how many thesaurus meanings does the word have, what
context was it used in: verb, preposition, adjective, etc..)

I realise that I am thinking about an intelligent system for construct
analysis which would first use elicited TEXT, (not just bipolarities or
triadic distinctions), and then do the battery of stat measures, that would
be something. I was once told that this is still not possible with the
present level of natural language processing and thus may remain somewhere
between wishful thinking and science-fiction. But I feel it will happen,
hopefully soon. My garden-variety spell-checker, for example, is able to
identify not just characters and words but even grammatical usage and phrases.

So, do you y'all think the quant v/s qual debate could be resolved by
technological means? What do expert programmers think about this?

(Calling Jack, Ken, Mildred, Jeremy, ... anyone online?)

Hemant Desai