Connectionism (Learning by Analogy)

Connectionism models cognition using interacting connected units. Learning by analogy is a distinct process of mapping relevant relations from familiar to unfamiliar cases.

Connectionism is a family of models in which behavior emerges from interactions among connected units, with learning often represented by changes in connection weights. It is not another name for learning by analogy. Analogical learning instead maps relevant relations from a familiar case to a new one; the title combines two distinct ideas.

For everyday study, use the analogy component explicitly: identify corresponding roles, transfer a relation, and mark where the comparison fails. Merely accumulating associations does not guarantee understanding, and a richer knowledge base can also activate misleading comparisons. A useful connection must support a correct inference in the new case.

When to use it

When distinguishing computational accounts of learning from practical analogies, or when testing whether a comparison improves understanding.

How it can help

When using an analogy, map relations explicitly, identify limits, and test a new inference. When discussing connectionism, specify units, connections, and learning rules.