From MEXICA to MEXICA-Impro: The Evolution of a Computer Model for Plot Generation - Computational Creativity Research: Towards Creative Machines - page 264

The Concrete Representation comprises the dictionary of Story-Actions and the
set of Previous Stories in memory. It is used during reflection to break an impasse
(see explanation in the next section).
The Abstract, Tensional and Concrete Representations represent the experience
and beliefs of the agent. They are structures that characterize different aspects of the
Story-Actions and the Previous Stories at different levels of abstraction.
13.2.3 The Engagement-Reflection Cycle
MEXICA generates stories through cycles of engagement and reflection. During
engagement an initial action is performed producing a Story-Context. The system
looks in memory for a Contextual-Structure that matches such a context and retrieves
the set of possible next actions associated to it. One of those actions is selected as
the next event in the story; the action is executed by the system, the Story-Context is
updated, and the cycle starts again. During reflection MEXICA evaluates if the mater-
ial produced during engagement satisfies the requirements of coherence, novelty and
interest. As a result of such an evaluation MEXICA produces a set of constraints
called guidelines; their purpose is to influence the production of material during
engagement.
An impasse is declared when the system cannot match any Contextual-Structure
in memory and therefore cannot continue developing the tale. Then, it switches to
reflection to try to break the impasse. The following lines provide some details of
these processes.
13.2.3.1 The Engaged State
Constraints, rather than goals or predefined story structures, guide the production of
material during the Engaged State. “A writing episode starts not with a single goal, but
with a set of external and internal constraints... As the writing progresses, constraints
provide the tacit knowledge to guide the writing process” ([ 3 ], p. 142-143).
During engagement MEXICA performs the following tasks:
1. The consequences of an initial action chosen by the user produce an initial
Story-Context.
2. This context is used as cue to probe memory; the goal is to match a
Contextual-Structure equal or similar to the current Story-Context and then
retrieve the set of possible next actions associated to it.
3. The system eliminates from such a set all those actions that do not satisfy the
guidelines.
4. The system selects between the remaining actions one as the next event in the
tale. The action is performed and the Story-Context updated.
5. Go to step 2.