How Phrase Origins Are Established

Ask where a phrase ‘really’ comes from and you will usually be offered a story: a snug little tale involving a Tudor tavern, a hangman, or a sailing ship. Ask a lexicographer the same question and you will get something less satisfying but far more trustworthy — a date, a book, a line of type, and a frank admission of what is still unknown. The gap between those two answers is the subject of this guide. Establishing where a phrase comes from is not a matter of finding the most charming explanation; it is a slow, evidential craft with its own rules, its own tools, and its own habitual disappointments.

Evidence, not intuition

The governing principle of serious etymology is simple to state and hard to live by: a claim about origin must rest on dated, quotable evidence, not on plausibility. It does not matter how neatly a story explains a phrase. What matters is whether someone can point to a printed or manuscript source that actually uses the words, and say when it was written.

This is why the great historical dictionaries are built on quotations rather than assertions. The Oxford English Dictionary was assembled from more than five million slips of paper, each recording a single use of a word or phrase in context, gathered by volunteer readers from literature, newspapers, sermons and letters. The entries that resulted are not opinions about what words mean; they are chronological galleries of the word in use, arranged so that the reader can watch a sense emerge, shift and sometimes die.

Two things follow from this. First, an etymology is only ever as good as its earliest solid citation. Second, the honest answer to many origin questions is ‘we do not yet know’, because absence of evidence is not evidence of absence — it may simply mean nobody has found the crucial quotation.

Attestation: the first solid sighting

An attestation is a recorded instance of a phrase in genuine use. The earliest one a researcher can find is the phrase’s first attestation, and it anchors everything else. If the earliest reliable sighting of an expression is 1856, then any story requiring it to have existed in 1650 has a two-century hole to explain, and the burden of proof lies squarely on the storyteller.

Attestation has to be handled with care. A quotation is only useful if three things are known: what the words actually were, where they appeared, and when. A misremembered line, an undated pamphlet, or a modern edition that silently modernises Tudor spelling can all corrupt the record. This is why reputable sources quote in the original spelling and give chapter and verse. When you see a phrase’s history written as ‘first recorded in’ a named work of a named year, you are looking at an attestation; when you see ‘it is said that sailors used to…’, you are looking at a story.

Attestation also distinguishes between a word existing and a phrase existing. The individual words in ‘the whole nine yards’ are all ancient, but the fixed expression is not attested until the twentieth century — which is precisely why its origin remains genuinely disputed rather than settled.

Corpora: reading at industrial scale

No individual could ever read enough to be confident about a first date. This is where corpora — large, searchable collections of digitised text — have transformed the field. Databases of early printed books, digitised newspaper archives, and searchable runs of periodicals let a researcher scan millions of pages in seconds for a target string.

The effect has been dramatic. Expressions once thought to date from the nineteenth century are repeatedly being pushed back into the eighteenth, and words assumed to be modern coinages turn out to have quiet earlier lives. But corpora bring their own hazards. Optical character recognition misreads old typefaces, so searches miss real hits and produce false ones. Publication dates in databases are sometimes wrong. And the sheer weight of digitised newspapers can bias the record towards whatever happened to be printed and preserved, which is not the same as whatever was actually said.

The discipline, then, is to treat a corpus hit as a lead to be verified against the physical source, not as a conclusion. A responsible researcher confirms the wording, confirms the date, and only then adjusts the phrase’s timeline.

Antedating: pushing the date back

Finding a citation earlier than the one currently on record is called antedating, and it is the daily bread of historical lexicography. Every antedating tightens the story: it narrows the window in which a phrase must have been coined and often kills a folk explanation outright by proving the phrase existed before the event it supposedly commemorates.

Antedating is also a standing reminder of humility. Because the digitisation of old texts is ongoing, first-attestation dates are provisional by nature. Scholars examining the revised third edition of the OED have found that a very large proportion of its entries can be antedated once fresh digital databases are brought to bear, which tells you how cautiously any single ‘earliest known’ date should be read. The date is the earliest we have found, not the earliest that exists.

The best etymologists treat every first date as a challenge rather than a full stop: it stands until someone, somewhere, turns up an earlier page.

Weighing the sense, not just the words

Establishing an origin is not only about finding the earliest appearance of a string of words; it is about finding the earliest appearance of the meaning. Many phrases exist in an older, literal sense long before they acquire their modern figurative one. ‘Break the ice’ meant physically forcing a passage through frozen water centuries before it settled into its social sense of easing awkwardness. A careful history therefore tracks the semantic journey, not merely the first time the letters line up in the right order.

This is why good entries distinguish between the earliest use of a phrase in any sense and the earliest use in the sense we now recognise. Collapsing those two dates is one of the commonest ways a phrase history goes quietly wrong.

When the honest answer is ‘unknown’

The hardest professional habit to acquire is comfort with uncertainty. A well-run etymology will say ‘origin uncertain’ when the evidence does not support a confident claim, and will resist the temptation to fill the vacuum with a good story. The record is full of expressions — ‘the whole nine yards’ among them — whose origins remain genuinely open despite intense searching. Declaring them solved would be more satisfying and less true.

The mark of a trustworthy source is not that it always has an answer, but that it distinguishes clearly between what is attested, what is inferred, and what is merely guessed. A date with a citation is a fact. A plausible-sounding tale with no citation is, at best, a hypothesis awaiting evidence — and, at worst, a folk etymology waiting to be believed.

Etymology and phrase-history are not the same task

It helps to separate two jobs that casual readers often merge. The etymology of a word traces its descent through earlier languages — the Latin, French, Norse or Old English roots from which the modern form grew. The history of a phrase is a different exercise: it asks when and how a particular combination of words acquired a particular figurative meaning. A phrase can be built entirely from words of impeccable ancient pedigree and yet be, as a fixed expression, quite young.

Confusing the two produces a characteristic error. Someone establishes that a component word is centuries old and then quietly transfers that antiquity to the whole phrase, as though the expression must be as old as its oldest ingredient. It need not be. The individual words ‘break’ and ‘ice’ are ancient; the settled social sense of ‘break the ice’ has its own, later timeline. Good phrase-history keeps the two clocks apart and dates the expression on its own evidence.

Manuscript, print and the shape of the record

The evidence itself is uneven, and knowing its texture guards against false confidence. Before the spread of printing, the written record is thinner and harder to date, so first attestations from the medieval period are often hedged with approximation. After printing, the record thickens, and after the rise of newspapers it becomes a flood. This means a phrase can appear to ‘begin’ at the point where the surviving record grows dense enough to catch it — which may be long after it was first spoken.

Speech, of course, almost always precedes print. Slang and idiom in particular circulate orally for years before anyone thinks them worth writing down, so the first printed sighting typically marks not the birth of a phrase but the moment it became respectable, or common, enough to reach the page. A careful researcher treats the earliest citation as a floor beneath the true origin, not as the origin itself, and phrases conclusions accordingly.

How to read an origin claim

Armed with these ideas, a reader can assess almost any origin claim quickly. Does it cite a dated source, or does it rely on ‘they used to say’? Does the earliest evidence actually contain the phrase, or merely a distant cousin of it? Does the proposed explanation predate the first attestation, or does it require the phrase to be older than any evidence allows? And does the account admit uncertainty where the record is thin?

Phrase histories established this way are less colourful than the tavern-and-gallows variety, but they have the one quality the stories lack: they are built to survive the discovery of the next earlier citation. That is what it means to establish an origin — not to tell the best tale, but to leave a trail of dated evidence that anyone else can walk.