Ken, Jonas, Fil, Karen, et al,

I think this discussion hasn't done much to clarify and resolve how
models of designing and of subject domains can be integrated. Ken's
view that " domain models and taxonomic models don't work the same way
that process models do" tends to overlook the dependencies that arise
between them when one attempts to apply either type. More importantly,
it fails to consider models that incorporate both . For example, my
model of design thinking is a process model structured to address
different types of information in any subject domain, even policy,
where complexity, social interaction and emergence over time are
involved. See "A Role Oriented Approach to Problem Solving" at http://independent.academia.edu/charlesburnette
.

Policy or any problematic situation requiring resolution by human
beings are matters for design and need a general model whose
instantiation can be reliably identified in any subject domain or
problematic situation. Most models of design provide no handles by
which this can be achieved. A small number of components is usually a
compromise with reality that seeks to make a model easier to grasp.
Unfortunately they offer few handles to aide application or
integration unless their expansion is specified. (see IDESIGN.com for
a short and long form models) Mappings of similarities between models
with the same number of components is not integration or resolution.
Integration requires conceptual blending of related information, or,
metaphor like interpretations of one model through the lens of
another, or, differentiation between models. During research leading
to the theory of design thinking, I correlated many models to validate
the power and utility of a model based on information types and the
processing each type required. I found many compatible models that
shared the seven part structure. (Scientific disclosure, types of
organization, computational objects, pattern language, etc.) The
correlations often revealed lack of coverage (incompleteness), over
definition of one or more dimensions, or focal inconsistency in other
models. A useful analytic tool that often led to improving the model
being analyzed resulted.

Finally, with regard to Fils' point about the role of the unconscious
in finding resolutions, and recognizing that unconscious acts depend
on the knowledge and experiences of individual thinkers, it seems to
me that the many remarks describing the need for abstract analysis
using formal tools (statistics, ethnography, etc.) fail to state that
the selection of variables and interpretation of results is often
subjectively determined. A social process helps to share and agree on
these interpretations as they arise in the minds of the individuals
involved.

Chuck

Dr. Charles Burnette
[log in to unmask]



On Jul 9, 2011, at 8:33 AM, Ken Friedman wrote:

> Dear Jonas,
>
> It’s not clear to me that all these three-part models map onto each
> other. Not all of these models are process models. They are models of
> different kinds.
>
> The last time we disagreed about three-part models, the model we
> discussed was Frayling’s proposal for three kinds of design research:
> research into design, research by design, and research for design.
> This
> is not a process model, but rather a series of rubrics for three kinds
> of research. In Frayling’s proposal, each of these three rubrics
> describes a different approach to research. This is a taxonomic model.
>
> Your three-part model is a process model. In your model, analysis,
> projection, and synthesis are sequential steps in the same process.
>
> Peirce describes a process model into which he integrates his view of
> the scientific method. He starts with abduction (forming hypotheses),
> moving to induction, and finally to deduction.
>
> But the Nelson-Stolterman model describing the true, the ideal, and
> the
> real is not a process model. It is a domain model.
>
> Kolb’s experiential learning model is a four-part iterative cycle
> involving experiencing, observing, conceptualizing, and experimenting.
> For Kolb, these process steps can follow each other in several ways.
>
> The process models in Boland’s article also work.
>
> Without agreeing that the actual process models you describe do indeed
> map over onto one another, your description of the process models
> seems
> reasonable. My argument with mapping all three-part models onto one
> another is that domain models and taxonomic models don't work the same
> way that process models do.
>
> It’s difficult to see what’s notorious about a pattern of posing
> thoughts, requesting clarification, challenging debatable issues. In
> my
> view, the well-known end to this kind of debate is to clarify and sort
> through ideas, and sometimes to scrub off ideas that don’t really
> work. Isn’t that what research has always been about?
>
> Yours,
>
> Ken
>
>
> Jonas wrote:
>
> —snip—
>
> a notorious communication pattern is showing up again. I would like
> to
> contribute to avoiding the well-known end of this kind of debate.
>
> If I remember right, then this thread started with a discussion about
> 3-phase models of the design process.
>
> Anyway, I come back to this issue and invite you to give this
> debate a
> cybernetic turn. To look at it in a cybernetic spirit. Circularity,
> feedback, self-reference, etc. are the keywords. Also groundlessness
> and
> these infamous concepts.
>
> Many design and many learning process models have a 4- or 3-step
> structure.
>
> Most 4-step models go back to Kolb’s theory of experiential learning.
> In my terminology I call the steps research - analysis - synthesis -
> realization.
>
> And most 3-step models can be mapped to Peirce’s 3-step logic of
> Induction - Abduction - Deduction. In my terminology: Analysis -
> Projection - Synthesis. In Nelson (since Harold’s name appeared) and
> Stolterman’s terminology: the true - the ideal - the real.
>
> 4- and 3-step models can be closely related to each other, an
> advantage
> of the 3-step models is that the abductive step is made explicit.
>
> —snip—