Metaphor and when it helps

A metaphor is frequently treated as ornament, which underestimates what it does: it hands the listener a structure and they reason inside it.

What metaphors do

Beyond illustration.

They map an unfamiliar thing onto a familiar one, which allows inference — the listener can work out consequences you never stated.

Experimental work has found that the metaphor used to describe a problem changes what solutions people propose, even when the underlying facts are identical.

Which means the choice of metaphor is a substantive choice rather than a stylistic one.

And it means an inappropriate metaphor produces incorrect inferences that are difficult to dislodge.

Choosing one

What makes a metaphor useful.

The source is genuinely familiar to this audience, which is where most fail.

The structural relationship is the same, not merely the surface appearance.

The inferences it invites are ones you would endorse.

It is concrete and physical, which is what makes it easy to hold.

And it is short, since a metaphor that requires explanation is not doing its job.

Stating the limit

The step that is always skipped.

Every metaphor breaks down, and listeners extend them past that point.

Which means saying where it stops: "this is like a filter, except that nothing is removed permanently."

Which costs one sentence and prevents a class of misunderstanding.

And where the limit is important, it is worth stating before the metaphor rather than after.

Dead metaphors

Which are everywhere.

A great deal of ordinary language is metaphor that has stopped being noticed: grasping an idea, a strong argument, falling behind.

These are generally fine and are how language works.

They become a problem when the underlying mapping is doing unnoticed work — describing a discussion in terms of combat, for instance, shapes how it is conducted.

Which is worth noticing occasionally rather than constantly.

Mixed metaphors

The visible failure.

Combining two mappings produces incoherent inferences and, generally, unintended comedy.

The fix is to notice the metaphor you have started and stay inside it for the length of the passage.

Or to abandon it entirely rather than switching mid-sentence.

And to read aloud, which is where mixed metaphors become audible.

Overuse

The subtler failure.

A piece dense with metaphor is harder to follow rather than easier, since each one requires processing.

One extended metaphor, developed across a passage, is more effective than five brief ones.

And plain statement is frequently better than metaphor: not everything needs to be like something else.

Metaphor in argument

Where the stakes are higher.

A metaphor can carry an argument without making it, which is powerful and is why it is used in persuasion.

Which means being alert to metaphors in what you read: the framing frequently does more work than the evidence.

Asking what the metaphor implies, and whether the implication is defensible, is a useful analytical habit.

And offering a competing metaphor is frequently more effective than arguing against the first one directly.

Technical explanation

Where metaphor is most valuable.

Abstract systems are almost impossible to explain without one, since there is nothing to visualise.

Which means the choice matters most exactly where it is most necessary.

Good practice: introduce the metaphor, use it consistently, state its limit, and then move to the literal account once the structure is established.

And retire it once the listener has the model, rather than continuing to lean on it.

Building your own

Practically.

Ask what the thing is structurally: does it accumulate, flow, filter, branch, compete, decay?

Then find a familiar thing with the same structure.

Test it by working out what it implies, and check each implication.

Try it on someone and ask what they concluded, which reveals the inferences you did not intend.

And keep the ones that work, since a good metaphor for a difficult concept is worth more than any amount of careful phrasing.

Metaphors that have become policy

Where this stops being a matter of style.

Some metaphors are so embedded in how a field talks that they shape what gets proposed: describing an organisation as a machine invites different interventions from describing it as an ecosystem.

Neither is correct; both are partial.

Which means the useful habit for anyone in a specialist field is to notice the dominant metaphor and ask what it hides.

The things a metaphor makes invisible are generally more interesting than the things it illuminates, precisely because nobody is looking at them.