We began by posing two problems with a frame-based approach to meaning construction. The first problem contrasted the simplicity of frames with the complexity of human behavior, while the second contrasted the static nature of these data structures with the flexibility of the behavior they have been hypothesized to support. In answer to the first, the issue of how simple representations can sustain the complex cognitive operations people seem to perform, part of the explanation lies in the fact the people use language to set up construals rather than to evoke objectively valid models of the world. Although the models used in meaning construction are often schematic, and occasionally even implausible, they are supported by our ability to dynamically integrate models with contextual information, to exploit culturally defined scales, to shift frames, and to blend different models together.
In answer to the second question, of how static frames can sustain the flexible language use evidenced by speakers, we effectively said that they cannot. Though data support the contention that human knowledge representation is hierarchically structured, that slot-filler structure is fundamental, and that people do assume default values when none are provided, it seems that knowledge representation is more dynamic than originally proposed. Rather than just retrieving and instantiating them, speakers are continuously and creatively building and blending frames to yield new concepts, construals, and activities.
In underlining the importance of imagination in information integration, and understanding the dynamic, flexible character of utterance meaning as arising from the demands of situated cognition, the space structuring model affords a promising approach to the fascinating domain of on-line meaning construction. In its present form, the main contribution of the approach is a way of thinking about some old problems. However, it can and should be translated into empirical predictions and computational models, which in turn can be tested with the methods of cognitive anthropology, linguistics, psychology, and neuroscience. Given the tools at our disposal, we should soon be well on our way to the formulation of a precise understanding of the emergence of on-line meanings. Sometimes a journey of a thousand steps begins with a single leap.