Dynamic Symbol Grounding, State Construction and the Problem of Teleology

Symbol grounding has originated within the connectionist-symbolic debate so as to gap the bridge between the two approaches. This paper provides an overview about recent results concerning symbol grounding , which is critically reviewed here. A thorough analysis reveals that symbol grounding parallels transcen-dental logic and is best viewed as automated model construction. If this diagnosis is true, the necessary next question must be which sort of models are generated in symbol grounding systems. The answer to this question very much depends on the kind of network architecture employed for grounding. An illustrative neural network architecture is used to explain how a dynamic symbol grounding system generates a formal model. It is argued that, depending on the architecture, very diierent kinds of signs ranging from input to goal descriptions can be grounded. Based on a fundamental criticism of symbolic models and Searles \Chinese room" argument Searle 80], Stevan Harnad introduced \The Symbol Grounding Problem" in his 1990 paper. Symbol grounding (SG) tries to answer the question as to how it is possible for a computer program to use symbols which are not arbitrarily interpretable. Whereas the meaning of signs in conventional programs is just \parasitic on the meaning in our heads", grounded symbols should possess at least some \intrinsic meaning" Harnad 90]. The problem can also be formulated as how it is that formal symbol systems can acquire a semantics which is not based on other symbol structures but on the system's own sensory experience. It should be noted here that Searle|as opposed to Harnad|is more concerned with intentionality than with meaning. However, as of today, it is quite obvious that symbol grounding cannot solve the problem of original intentionality. The reason for this lies in the observation that intentionality is connected to consciousness, cf. Frixione & Spinelli 92]. But Harnads proposal is obviously directed towards the generation of correlational semantics. (For the more philosophical discussion see Christiansen & Chater 92, Christiansen & Chater 93]. Harnads answers can be found in Harnad 94].) Consequently, symbol grounding has mainly been performed in the context of linguistic research so as to clarify the question of semantics or the origin of reference in words (\word meaning"). The general idea is to equip a symbol system with some kind of measurement device. The goal of the system is to nd invariances in the sensory data whenever a speciic symbol is shown to the system. …