Slide 7 of 25
Notes:
Atomistic symbol tradition runs from Descartes to Early Wittgenstein to Minsky and Papert (Dreyfus & Dreyfus, 1988) to Fodor, and Newell and Simon.
The assumed correspondence between the surface phenomena and the underlying representations (semantic transparency and systematicity) presumed to underlie SPSS implies too narrow a range of possible representational systems. Each element of the model stands for a conceptual element. We can interpret the meaning of each theoretical element in terms corresponding to conceptual units. Although it would be convenient, for example, to identify activity of a certain brain area with having a certain idea, there is no intrinsic reason to believe that the brain was designed for the convenience of neurobiologists or that the programs executed by the brain are labeled for convenient debugging.
In many situations there is a tradeoff between representations that are easy to describe or communicate and those that are more effective. For example, Payton, argued that effective spatial planning algorithms that utilize a gradient field have largely been overlooked because "we are so accustomed to representing plans in ways designed for human interpretation" (Payton, 1990, p. 94).
Attributing to symbols properties corresponding to words does not help us to understand what makes those symbols carry any significant weight because words themselves are not solidly enough connected to their meanings or to referents. All, or at least most, words have ambiguous meanings. Polysemy is the rule, rather than the exception.
