The question exposes an ontological ambiguity in the word parameter. If a transient activation pattern implements the same input-output transformation as altered weights, functionalism treats the two as equivalent. A physicalist account does not: one state survives deployment and the other vanishes when the context window is cleared. The distinction resembles the difference between remembering a rule in a notebook and reconstructing it afresh from instructions. Both guide action, but only one changes the enduring system.