In RP English, the 'w' in "sword" is silent. Wiktionary suggests /sɔːd/ and /soʊrd/.

Why? Are there other words like this? The 'w' is pronounced in words like "swollen", "swoop", "sworn" and "swore".

  • 3
    The W was pronounced in Old English, the pronunciation changed over time, but not the spelling.
    – GEdgar
    Commented Jul 5, 2011 at 16:32
  • 4
    I'm wondering why this happened with "sword" but not with "sworn", and whether "sword" is the only word in which the 'w' drifted away.
    – user4727
    Commented Jul 5, 2011 at 16:39
  • 6
    @Tim: 'answer'? 'coxswain'? 'who'?
    – Mitch
    Commented Jul 5, 2011 at 17:05
  • 8
    Churchill supposedly said the Royal Navy's traditions consisted of "rum, sodomy, and the lash". The word "coxswain" (pronounced coxun) -- like "boatswain" (often spelled "bo's'un" and always pronounced bosun), "forecastle" ("fo'c'stle", foxul), and "gunwhale" (gunnel) -- is a victim of a fourth proud tradition: reducing the King's English down to a mumble. Commented Jul 5, 2011 at 17:12
  • 5
    Not only "sword" and "answer" but also "two"...
    – user20543
    Commented Apr 26, 2012 at 9:37

5 Answers 5


I have actually heard someone pronounce the /w/, and not facetiously. It must be rather rare, though, and I might say it's a hypercorrection. In any case, my guess is that [wɔ] and [wo] have a tendency to become [ɔː] and [oʊ] in the neighbourhood of consonants such as [s] that don't change much in the presence of labialisation. The matter is complicated a bit by the spread of non-rhotic accents.

The original word was /sweord/, which according to the usual Old English reconstructed pronunciation would be [sweort] or [sweʊrt]. I imagine [eʊ] dropped to [əʊ] or [əu], which is an allophone of [o]. By this time the [w] had become vestigial and was ultimately dropped, but English spelling is far more conservative than its pronunciation, so the /w/ was retained.

  • Can you come up with any examples to support the theory of [swɔ] and [swo] drifting to [sɔː] and [soʊ]?
    – user4727
    Commented Jul 5, 2011 at 17:18
  • 2
    @Tim: I was going through the eo diphthong, which made a wide variety of transitions: eoweryour, (a)beornanburn, seofonseven, seoþanseethe, and seocsick.
    – Jon Purdy
    Commented Jul 5, 2011 at 17:33
  • Quite right, but your answer doesn't explain why 'swore' is pronounced with a W. See my answer below Commented Jan 3, 2021 at 19:29

Appropriately it's silent in answer. I couldn't find any other root -sw- words with a silent w.

As for sword, I found this from H.L. Mencken's The American Language from 1921:

As for the consonants, the colonists seem to have resisted valiantly that tendency to slide over them which arose in England after the Restoration. Franklin, in 1768, still retained the sound of l in such words as would and should, a usage not met with in England after the year 1700. In the same way, according to Menner, the w in sword was sounded in America “for some time after Englishmen had abandoned it.”

(There's a whole host of silent w s at the start of words but not after s, but they can be considered another "family": wring wrap wrong wrist write wraith wreath wraparound wreck wrath wreak wreck wrench wreckage wrecker wrestle wren wriggle wretched wrest wrinkle wristlet wristwatch writ write writhe wrought wry.)

  • 2
    I'd say that the "wh*" belong to another family. Didn't think of "answer". Yes, the line is fuzzy and subjective.
    – user4727
    Commented Jul 5, 2011 at 17:08
  • I'm with @Tim: those are "silent" for a different reason.
    – Jon Purdy
    Commented Jul 5, 2011 at 17:11
  • Yeah, there's a whole host of those (wring wrap wrong wrist write wraith wreath wraparound wreck wrath wreak wreck wrench wreckage wrecker wrestle wren wriggle wretched wrest wrinkle wristlet wristwatch writ write writhe wrought wry). I'll edit my answer with some more info.
    – Hugo
    Commented Jul 5, 2011 at 17:31
  • 2
    Those /wr/ ones aren't really silent as such—it's just that labialised and non-labialised /r/ have merged in most dialects. In AmE and BrE, they're normally both labialised (so ‘ring’ sort of really starts with an unwritten w); in IrE, they tend to be unrounded (so the written w is ‘silent’); and in ScE and some other dialects, they're still distinct, so /r/ is [r] and /wr/ is [wr]. Commented Apr 29, 2014 at 20:31

<w> is there because it was pronounced after the advent of printing press. So, the spelling stayed, but pronunciation changed. In three cases, andsƿarian > answer, sƿeord > sword, tƿa > two, <w> is there, but not pronounced.

On the other hand, <w> (or its OE <ƿ>) is not there in some words because they were not pronounced by the time printing took over: OE sƿilch > such, sƿa > so


This answer on ELL has explained the reason why the W is silent in sword but pronounced in swore. That answer is rather long, but I'll simplify it here:

Between 900–1400 AD, there was a sound change through which a W was lost when it was preceded by a [s] or a [t] and followed by a back vowel [ɒ ɔ o ɑ]. [Wikipedia]

The W in two and sword is silent because of that reason (i.e. [t, s] [w] [back vowel]).

Swore, sworn and swollen also lost their W's at one point due to the same reason, but were later on restored due to analogy1 with swear and swell. Swear and swell hadn't lost their W's because they had front vowels after the W. [Trask's Historical Linguistics]

1. Analogical change is a type of language change in which some forms are deliberately changed merely to make them look more like other forms.

  • I recently came across another example of analogy mentioned in this answer, it says that the plural of 'book' was 'beek', but due to analogy with other regular forms, its plural has changed to 'books'. Commented Jan 3, 2021 at 17:40

Just a theory: As sword is a thing of knights and noblemen. The silent w may be due to Norman-French pronunciation habits and this pronunciation was generally accepted.

The w in spelling shows the origin of the word and its connection with German Schwert. With the w-spelling it is clear at once that the word does not belong to the word family "sort".