Page 56 -

ITERATION IN SYSTEMS ANALYSIS AND DESIGN 41
both the problem spaces and the solutions spaces, which designers manipulate in order to connect
the solution space with the problem space. “Mental model” here becomes a generic term that is
used to describe (meta)concepts that organize representations of problems and solutions and their
connections. This includes representational metamodels such as frames, schemas, causal models,
situational models, and so on (Brewer, 1987). This notion was popularized by Johnson-Laird (1980)
to refer to cognitive representations that are constructed as required to assist human cognition.
Mental models are not images of problems or solutions, but can lead to such images. Specific,
localized mental models are expected to both draw from and contribute to a global schema of
“generic knowledge structures” within the individual that can later be expected to leverage a new
“episodic” mental model during design (Brewer, 1987).
These ideas underpinning design cognition form the essence of the “rationalistic” tradition of
cognition. Yet, alternatives exist that criticize some of the fundamental assumptions of rational
models (Bruner, 1990; Hutchins, 1995; Suchman, 1987; Weick, 1979; Winograd and Flores,
1986; and others). Any attempt to reconcile these critiques with the rationalistic tradition would
be problematic, as rationalist theories address issues such as “meaning” in a simplistic manner,
whereas many of the other traditions view the meaning of “meaning” as highly nuanced and
situated (Suchman, 1994). In the rationalistic tradition “the machinery of the mind has taken
precedence in theory building, insofar as mental representations and logical operations are taken
as the wellspring for cognition” (Suchman, 1994, p. 188). A family of alternatives that are par-
ticularly salient to research on design cognition can be called the “situated action” perspective,
which calls attention to “the socially constructed nature of knowledge, meaning, and designs . . .
no objective representations of reality are possible; indeed, intelligence is not based exclusively
on manipulating representations” (Clancey, Smoliar, and Stefik, 1994, p. 170).
The situated action view does not focus exclusively on what happens within an individual’s
mind. Rather, it looks at the interactions between social and contextual phenomena within the
ongoing activity of a designer (Suchman, 1987; Winograd and Flores, 1986). An example of an
iterative cognitive activity in this tradition would be the idea of a hermeneutic circle of interpreta-
tion where the individual leverages his “pre-understanding” to understand something within its
context and forms a new “pre-understanding” (Winograd and Flores, 1986). Each hermeneutic
circle can be considered a cognitive iteration.
Although mainstream management and design research generally aligns with the rationalistic
tradition, there is an increasing amount of research that emphasizes interpersonal negotiation and
dialogue as a key to understanding design (Bucciarelli, 1994; Clark and Fujimoto, 1991). In this
stream, the idea of cognitive iteration is not the neat, temporally ordered, and fully formed mental
model of a design in an individual’s mind. Rather, it is a messy, partially formed object and process
of dialogue, laden with meaning and interests and evolving through hermeneutic cycles. In the situ-
ated action view, the notion of a discrete and individual cognitive iteration loses its vividness.
To summarize, we must first become aware of the assumptions of each tradition, as each
tradition offers an alternative view of iteration. The rationalistic tradition assumes fully formed
and well-organized mental models that emerge and are manipulated during design, whereas the
situated action perspective assumes partial, evolving understandings of the design as realized in
dialogue. Either way, both these cognitive iterations share three facets:

1. steps or stages within the design (e.g., generate-test cycles/hermeneutic circles);
2. the design process as a gradual movement of the “mental” object (mental model/under-
standing); and
3. the design object (the representation/the text).