Reasoning with Limited Unification in a Connectionist Rule-Based System

At the intersection between symbolic inference and connectionism, there is interest in producing systems constructed from connectionist components which perform types of inference comparable to symbolic systems. The attraction of such systems is that, under restricted conditions, they may be capable of very fast, \reactive" responses to external stimuli. A major open question in this eld is the extent to which a connectionist architecture can accommodate basic concepts of symbolic inference, such as uniication of terms. One of the current leaders in this area is the rule-based system proposed by Shastri & Ajjanagadde (1993). We demonstrate that the mechanism for variable binding which they advocate is fundamentally limited and show how a reinterpretation of the primitive components of their system can extend the range of inference which can be supported.