Image schemas and conceptual metaphor (writing rule)
Part of Writing principles.
What the rule is
Image schemas are recurring sensorimotor patterns (CONTAINER, SOURCE-PATH-GOAL, VERTICALITY, FORCE/BLOCKAGE, LINK/PART-WHOLE). Conceptual metaphor theory says abstract talk systematically borrows those patterns. The writing rule: make the schema visible so spatial cognition does binding work — and name where the metaphor stops so it does not lie.
What good looks like
-
“Memory lives inside a thread’s private room” (CONTAINER) with an explicit boundary (what cannot cross).
-
Causal stories told as paths with a start, middle, and end state.
-
Vertical metaphors used sparingly and consistently (e.g. call stack up/down).
Failure modes for LLM writing
-
Mixed metaphors that fight (“the lock rises while flowing through a container”).
-
Pretty metaphors that contradict the mechanism.
-
Schema name-dropping without actually structuring the explanation.
Micro-example
Weak: “The allocator reclaims unreachable objects.”
Stronger (schema-forward): “Imagine the heap as a room of boxes. A box is still in play if you can walk a path of pointers from a root to that box. If no path reaches a box, the collector treats it as outside the living set and reclaims the space.”