What Is Etymology? Where Words Come From and How to Trace Them

Most English words did not begin in English. Roughly a quarter of the vocabulary is Germanic, inherited from the Anglo-Saxon settlers of the fifth century and shared with Dutch and German: house, bread, king, the numbers, the pronouns and nearly all the commonest words. Another large share came in with the Normans after 1066 from French, and a further wave was borrowed directly from Latin and Greek by scholars from the Renaissance onward. That is why English so often has three words for one thing at rising levels of formality, such as kingly, royal and regal, one from each source.

Behind those languages stand older ones. Latin, Greek, the Germanic and Celtic languages, Sanskrit, Persian and Russian all descend from a single language spoken somewhere north of the Black Sea about six thousand years ago, called Proto-Indo-European, which was never written down and is known entirely by reconstruction. The English word father, the Latin pater, the Sanskrit pitar and the Greek pater are the same word, changed by six thousand years of separate drift.

The method rests on the discovery, made in the early nineteenth century, that sound changes are regular. When a language shifts a sound, it shifts it in every word where the sound occurs in the same position, not at random. Jacob Grimm, of the fairy tales, described the pattern by which the Indo-European p, t and k became f, th and h in the Germanic languages, so that Latin pes matches English foot, tres matches three and cornu matches horn. A proposed etymology that requires an irregular change is suspect; one that follows the known rules is strong evidence even without documents.

The other tool is the record. The Oxford English Dictionary traces each word through dated quotations from its first appearance in writing, so that a claimed origin can be checked against when the word actually turns up and what it meant then. An etymologist's working questions are:

• •What is the earliest recorded form of the word, and in what language?
• •Do the sound correspondences with the proposed source follow the regular rules?
• •Does the meaning of the source make sense as an ancestor, and can the change of meaning be paralleled elsewhere?
• •Is there a plausible route of contact by which the word could have travelled?
• •Is the popular story about it older than the word, or invented afterwards to explain it?

Sounds change by rule; meanings change by use, in a few recurring ways. Words narrow: meat once meant any food, and deer any animal. They widen: dog was one breed, and bird a young bird. They slide up or down the social scale: knight once meant servant and villain once meant farm labourer. They shift by metaphor, which is how a computer acquired a mouse and a window, and by euphemism, which is why toilet, originally a cloth, has replaced a series of earlier words that each became too direct. And they reverse: awful meant awe-inspiring, and terrific meant terrifying, and both were worn down by exaggeration into general intensifiers.

Because people want words to make sense, they invent origins for them, and the invented origins are often better stories than the true ones. Posh does not stand for port out, starboard home; the phrase appears decades after the word, and the word probably comes from a slang term for money. Crap has nothing to do with the sanitary engineer Thomas Crapper, being Middle English for chaff. The rule of thumb is not from a law permitting wife-beating with a stick of that width, which no court ever recorded. Any origin story that is an acronym and dates from before about 1900 is almost certainly false, because acronyms were rare before the twentieth century.

Folk etymology can also change the word itself. Asparagus became sparrow-grass in English kitchens, a crayfish is not a fish but a French ecrevisse reshaped, and a bridegroom's groom is an Old English guma, meaning man, replaced by a familiar word after the original was forgotten. The habit that produces the false stories is the same one that shapes the real history.

An etymology is not a definition. Decimate once meant to kill one in ten, and insisting on that meaning now is a way of misunderstanding what people say. What an etymology gives is something else: the record of how a word was used before it reached us, which is a record of what the people who used it were doing. The vocabulary of English holds the Roman occupation, the Viking settlement, the Norman conquest, the Renaissance, the Empire and the internet, and the words for a thing tell who brought it and when. Learning where a word comes from does not settle what it means, but it makes it harder to use thoughtlessly.

Take the word window as a worked example. Its earliest English forms, in texts from the thirteenth century, look like windoge and windowe, and they match an Old Norse word, vindauga, made of vindr, wind, and auga, eye. The route of contact is plain: Norse speakers settled across northern and eastern England from the ninth century, and their language lent English many everyday words, including sky, egg, knife and they. The meaning makes sense too, since early windows were unglazed openings that let in air and let you look out.

The same evidence explains what the new word replaced. Old English had its own terms, including eagthyrel, eye-hole, whose first part is a cousin of the Norse one: Old English eage and Norse auga both descend from the same Germanic word for eye. That regular correspondence is the kind of check an etymologist relies on. Neither form tells you what a window is today, when most are sealed sheets of glass. But the history preserves a picture of a draughty opening in a wall, and of a time when two related languages lived side by side in England.

Most words drift into existence unnoticed, but some were deliberately coined, and their origins can be pinned down closely. The word gas was invented in the seventeenth century by the Flemish chemist Jan Baptist van Helmont, who based it on the Greek chaos, pronounced in his Dutch with a throaty g. Robot entered English through the Czech writer Karel Čapek's 1920 play R.U.R., where it named artificial workers; Čapek credited his brother Josef with the word, which comes from a Czech term for forced labour. The physicist Murray Gell-Mann took quark from a line in James Joyce's Finnegans Wake.

Other words come from people. Boycott preserves Captain Charles Boycott, a land agent in County Mayo whom his neighbours refused to work for, serve or speak to during a campaign for tenants' rights in 1880, and sandwich is usually traced to an eighteenth-century Earl of Sandwich, said to have eaten meat between slices of bread so that he need not leave the card table. Blends fuse two words, as smog joined smoke and fog early in the twentieth century and brunch joined breakfast and lunch. Even here, the colourful story attached to a word, like the Earl's card game, can be harder to verify than the word itself.

Etymology traces words back through their recorded forms and through the regular sound changes that connect related languages, most of Europe's descending from a single unrecorded language of six thousand years ago. Meanings shift by narrowing, widening, metaphor and euphemism rather than by rule, and the popular stories that explain a word's origin are usually later inventions, which is why a good etymology rests on dated evidence rather than on a satisfying tale.