← Back to The Art Laboratory
By Julian Sharples

The Life of a Metaphor

On coinage, decay, and the four thinkers for whom trope does not describe reality but produces it — read alongside the illustration Using Metaphor in Science.

A Correction to the Standard Definition

Metaphor is itself a metaphor, a “carrying over” that carries nothing literally. This is as good a place as any to start, because that fact is a case of the pattern this piece traces. Every metaphor has a lifespan. It is coined once, live and effortful, a fresh cross-domain leap the listener has to work to follow. If it catches on, it settles into a second stage: conventional but still generative. Its source domain is dormant but reactivatable, capable of generating new coinages on demand. A “war” is a metaphor that continues to lend “argument” new terms, such as “outflanked” and “indefensible,” centuries after the mapping first took hold. And it can die: not vanish, but calcify into an independent literal sense, the source domain no longer needed to use the word competently. “Understand” no longer asks anyone to picture standing beneath. In contrast, “shark,” for a ruthless businessman, still draws on that predatory logic, which is what keeps it a metaphor rather than a second dictionary entry. Quintilian’s ancient and still-influential definition of metaphor as nothing more than “a shorter simile” is a theory built entirely on this third stage: decoration on top of a literal thought that was supposedly there first. It cannot explain a live coinage, still less the profusion this piece ends with; it describes only metaphor’s corpse.

Reality is shaped by metaphor

At the far end from Quintilian sit four thinkers who argue that trope does not describe reality but produces some order of it. For Proust, in Kristeva’s reading, the analogical is the ontological: metaphor does not point at two objects’ shared feature, it reveals the depth of being beneath both. Here we see alliteration and rhythm, embodied precursors to the fully-formed image. For Hayden White, historical narrative is tropologically prefigured before a historian says anything at all: a chronicle of events becomes an intelligible history only once organized, in advance, by metaphor, metonymy, synecdoche, or irony — trope as the precondition of historical understanding, not an ornament added after the fact. For Lacan, trope turns properly causal where metaphor is condensation, and “the symptom is a metaphor.” It is not a figure describing the symptom but the substitutive mechanism that produces it. Metonymy is displacement, and desire itself is a metonymy, an endless sliding along a signifying chain. And at the furthest extreme, Deleuze’s The Fold replaces trope with a formal mathematical object outright, the cone’s sections. These are the circle, ellipse, parabola, hyperbola, and they become the literal structure by which each monad’s point of view generates a different perspective on one shared world, no monad sitting at any true centre, since an ellipse is organized around two foci, not one. At this end, the question “is this a metaphor, or is it real” stops being answerable — the distinction has folded.

The vocabulary of science

Scientific vocabulary lives permanently astride these two poles, because the same term can be alive and dead simultaneously, depending on who is using it. A physicist for whom “flux” still summons lines flowing through a surface, or a botanist who still feels “stoma” as naming a mouth, is using the term with its source domain reactivated. This is metaphor, still doing epistemic work, still shaping how the phenomenon is intuited. A student who has only ever met the word as an exam definition is using the identical term as pure nomenclature, no more alive than a serial number. The echo of the original coining survives. It is for those with access to the source domain the coiner drew on, while for everyone else the word has already reached the third stage of paragraph one’s life cycle. Science’s technical vocabulary is never uniformly metaphorical or uniformly literal at any given moment; it is split by audience, the same word doing different work depending on who is holding it.

Six concrete coinage mechanisms

And the vocabulary carrying that split was made in remarkably various ways.

  1. A native word pressed directly into service, no Latin or Greek detour at all: field, Old English feld, given a defined physical sense by Faraday.
  2. An already-naturalized Latinate word, centuries old, redeployed because it already carried a technical register from elsewhere: flux, recruited from Newton’s calculus before Faraday ever touched it.
  3. A single Greek root, freshly compounded with an invented ending, minted by one named person on one dated occasion: entropy (Clausius, 1865, built from tropē specifically to echo “energy”); xylem and phloem, both Nägeli, 1858, wood and bark plus a Greek-sounding suffix.
  4. A double compound of two Greek roots, each already a metaphor before the compound existed: cytoplasm (Kölliker, 1863, “hollow vessel” plus “molded thing”), chloroplast (Schimper, 1883, “green” plus “formed”).
  5. A whole pre-existing Greek word borrowed intact, its adoption anonymous and gradual: stoma (“mouth,” applied to leaf pores by 1837, no individual credited), stroma (“bed-covering,” entering general use by the 1830s).
  6. And, now and then, a second hop stacked on a first — a diffuse borrowing narrowed, decades or centuries later, by one named person on one dated occasion: stroma’s general sense, narrowed by Schimper in 1885 to the chloroplast’s fluid matrix; persona, Rome’s dramatic mask, narrowed by Cicero in 44 BC into a theory of social roles.

Native, naturalized, singly coined, doubly coined, borrowed whole, borrowed and narrowed twice — the profusion is the point.