Mechanism of Semantic Drift

A theory of linguistic change sits inside a single claim: that language changes through its use by those who took no care to understand what the words they use meant to others who had previously used those words. It identifies a mechanism—imperfect transmission between generations of speakers—and at once assigns something close to blame. “Took no care” frames the result as a lapse, a failure of diligence, as though a word arrives with a fixed prior meaning that an attentive speaker could have preserved and a careless one let slip.

The insight is genuine and worth keeping. Meaning is not handed down intact. Every speaker learns a word from a finite, accidental sample of its uses and infers what it means from that sample. No one receives the word; everyone reconstructs it. To that extent the claim points at the engine of semantic drift: the gap between a word’s history and any one person’s exposure to it.

But the framing mistakes a structural necessity for a contingent omission. The “care to understand what the words meant to others who had previously used those words” is not merely neglected—it is unavailable. No speaker has access to the full record of a word’s prior use; the prior users are mostly dead or unrecorded, and those on record often disagreed. The picture of a stable original meaning eroded by inattentive newcomers presupposes a uniformity that never existed. There was no single thing a word “meant to others,” only a spread of overlapping uses. Carelessness cannot erode what was never solid.

The framing also fails its own predictive test. If change flowed from a lack of care, then care should arrest it. It does not. The domains of greatest lexical diligence—law, science, theology, lexicography itself—are not exempt from drift. Precise technical coinages broaden into popular use; statutes accrete interpretive layers; “energy” and “momentum” carry senses their seventeenth-century users would dispute. The closest attention to prior meaning sometimes slows change locally but never stops it, because the next speaker still learns from a sample and still infers.

And the mechanism named is only one of several. Sound change is largely unconscious and systematic, indifferent to meaning altogether. Children acquiring a first language are not careless; reconstruction from limited input is the only thing acquisition can be, and it is a principal source of grammatical change. Speakers routinely extend words through metaphor, metonymy, and analogy—understanding the current sense perfectly well and stretching it deliberately. Much change is the work of competence, not negligence.

A further problem concerns agency. The claim locates change in individuals’ failure to do something, but no individual changes a language; change is an aggregate outcome that no participant intends or perceives. It resembles a path worn across a lawn: true that no walker “took care” to preserve the grass, but the path is a collective product, and treating it as the sum of individual omissions misdescribes how it formed.

The claim is also, unavoidably, self-exemplifying. It is made in words—“language,” “use,” “care,” “meant”—whose senses have shifted under its own author exactly as the thesis describes. If the proposition is true, the author cannot have fully grasped what those words meant to their predecessors either. This is not a refutation so much as a demonstration: the statement enacts the condition it laments.

Strip away the note of grievance, and what survives is a sharper and more interesting claim. Language changes not because speakers are careless but because inheriting a word’s full history is impossible, and a word reconstructed from a partial sample will always differ, however slightly, from the word that produced the sample. The same logic governs more than language. Traditions, institutions, rituals, and concepts all pass through hands that received only a fragment of what they were and reassembled the rest. Change of this kind is not decay and not error; it is the unavoidable signature of anything transmitted by learning rather than by copying. The statement is most useful read against its own grain—as a description of how continuity works, mistaken by its author for a description of how it fails.