Where a contraction may sit inside a word

There are 33 common words with a two-cell contraction of their own: one, day, time, there, under, question. Used as whole words they are straightforward. Used inside a longer word they are governed by a rule that reads well and cannot be executed: the contraction may be used where the letters keep their meaning.

Someone keeps the one. Component does not, though the letters are identical. Nothing about the spelling of either word says which. Meaning is not a property a translator can see.

What changed on 18 September 2026

Until that date this survey was the rule. Each of the 33 contractions carried the positions where the reference kept it more often than not, and this site's translator used the contraction there and nowhere else.

That was a measurement standing in for a rule, and it went wrong in both directions. It spelled the sign out where the rulebook uses it — stonework, atonement — because the letters of one in the middle of a word are usually not the word. And it used the sign where the rulebook does not — laundering, multimedia, severity — because the letters sat in a position that was usually safe.

The Rules of Unified English Braille do not decide by position. They decide contraction by contraction (10.7.1 to 10.7.9). Most are used wherever their letters occur. Three have a test that spelling can settle: one is not used after an o, ever not after an e or an i, under not after an a or an o. The rest turn on how a word is said or what it means, and there the rulebook gives its answer as a list of words. The translator now follows exactly that, and holds those lists.

From the rulebook's own examples, as the translator writes them now: stonework ⠌⠐⠕⠐⠺; atonement ⠁⠞⠐⠕⠰⠞; laundering ⠇⠁⠥⠝⠙⠻⠬; multimedia ⠍⠥⠇⠞⠊⠍⠫⠊⠁; severity ⠎⠑⠧⠻⠰⠽. Against the reference over the 5,000 commonest English words, the change took the match from 4,993 to 4,997, and all 226 of the rulebook's worked examples for these contractions that this site checks are written cell for cell.

The survey below is kept because it is still true of what it measured: how often liblouis 3.29.0 keeps each contraction, by where its letters sit. Read that way, it does not describe a rule about position. It describes how often a contraction's letters turn up inside an unrelated word — which is what the rulebook's lists already record.

The method

Position carries most of the signal

Three of the four are decided. At the start of a word the contraction is kept almost always; at the end almost always; in the middle, apparently, a coin toss.

That is where the useful part begins.

The fourth position

There are three obvious positions — start, middle, end — and measuring only those gets a real answer wrong.

A word like ones, phones, tones or theirs has the contraction at the end except for a plural s. Counted as “middle”, they drag the middle rate down and, worse, an engine that follows the middle rule spells them out. The reference keeps the contraction in 83.9% of them: a trailing s behaves like the end of the word, not like the middle of it.

So the survey uses four positions and checks the trailing s case before the other three. That change alone recovered several very common words.

The middle is not one case

Splitting the middle by which contraction is involved — every one with at least 20 examples — the aggregate 36.8% turns out to be an average of two populations that barely overlap.

14 of the 18 are above 80 per cent or below 20. Only 4 sit anywhere in the middle.

The words at the top are the ones whose letters are almost always a real part of the longer word: part in department and apartment, work in networking, know in unknown. The words at the bottom are common letter strings that land inside unrelated words: here inside therefore, name inside tournament, one inside component and commissioner and pioneer.

That is why a single rule about the middle of a word cannot work. It is also why position looked like the answer. The contractions at the bottom of the table are the ones whose letters keep landing inside unrelated words, and the rulebook names those words; the ones at the top rarely do, and the rulebook has almost nothing to list for them. Position was a shadow of the word lists.

What this site does with it

Nothing, any more, and that is the finding. The translator follows the rulebook's rule for each contraction, and each contraction's page states that rule with the rulebook's own examples, in the cells the translator writes.

This page once said that the positions on those pages were read from the same table the engine used, so the two could not disagree. When the engine moved to the rulebook on 18 September, all 33 pages went on describing the old positions for nine days, and fourteen of them contradicted what the translator actually wrote, until an audit caught it. So the braille on those pages is now checked rather than trusted: verify/verify-pages.js translates every example the pages show and fails if one differs from what the page prints.

Each contraction's page states its own rule and the words the rulebook sets apart. The finder shows the rule firing on a word you choose, including the cases where a contraction was legal and declined. The other measurements cover the reference tables, contraction positions and page counts.