How did the number spelled "two" get its pronunciation?

[Space60]

I've read about how "one" got its strange pronunciation, but I've never seen anything written about how "two" came to be pronounced /tu:/.

[Aftovota]

From what I understand, it was originally /twɑː/, then raised to /twoː/ during the Great Vowel Shift, then further raised to /twuː/. w > ∅ / C_V[+round], see "who", "whore", "sword", but contrast "swoon", "swollen".

[Pabappa]

This is just speculation, but the vowel irregularity could be due to restoration from an unstressed schwa form. Perhaps "who" also did this? Or maybe it's a regular change I don't know about ..... I haven't looked up any of this.

[anteallach]

From what I understand, it was originally /twɑː/, then raised to /twoː/ during the Great Vowel Shift, then further raised to /twuː/. w > ∅ / C_V[+round], see "who", "whore", "sword", but contrast "swoon", "swollen".

That's basically the explanation in the OED: the /w/ caused an extra vowel shift (as also in who; it also mentions womb), before disappearing as it sometimes did before /uː/.

It's not quite clear to me whether they think the extra shift happened before or after the GVS, but there's a 1400 citation with the spelling twoo, which might suggest that it was already /twoː/ before the GVS.

There are a few other examples of words which seem to have taken two steps of the GVS, for example lose. In that case the explanation is supposedly influence from loose (now a common spelling mistake of course). The traditional dialect where I live has /lɔɪz/, which is actually the expected development.

[Nortaneous]

There are a few other examples of words which seem to have taken two steps of the GVS, for example lose. In that case the explanation is supposedly influence from loose (now a common spelling mistake of course). The traditional dialect where I live has /lɔɪz/, which is actually the expected development.

With CHOICE?

[anteallach]

There are a few other examples of words which seem to have taken two steps of the GVS, for example lose. In that case the explanation is supposedly influence from loose (now a common spelling mistake of course). The traditional dialect where I live has /lɔɪz/, which is actually the expected development.

With CHOICE?

Yep. Traditional West Riding of Yorkshire (approximately; dialect boundaries don't follow county ones of course) dialects have /ɔɪ/ for standard GOAT in words where the long vowel comes from ME open syllable lengthening, which is what the regular development of lose would have been. A /kɔɪl ɔɪl/ is a cellar used for storing coal ("coal hole"). The royd in the names Boothroyd, Murgatroyd etc. is another example.

Not many people actually speak like this these days of course.

[Space60]

From what I understand, it was originally /twɑː/, then raised to /twoː/ during the Great Vowel Shift, then further raised to /twuː/. w > ∅ / C_V[+round], see "who", "whore", "sword", but contrast "swoon", "swollen".

The words "whore" and "whole" never actually had a /w/ sound in them. The "w" was added to the spelling sometime for some reason. "Who", "whom" and "whose" had a /w/ that was lost.

[mèþru]

The <w> was added for disambiguation.

[anteallach]

From what I understand, it was originally /twɑː/, then raised to /twoː/ during the Great Vowel Shift, then further raised to /twuː/. w > ∅ / C_V[+round], see "who", "whore", "sword", but contrast "swoon", "swollen".

The words "whore" and "whole" never actually had a /w/ sound in them. The "w" was added to the spelling sometime for some reason. "Who", "whom" and "whose" had a /w/ that was lost.

The w in whole is actually connected to the development in one which you mentioned in the first post, except that here it survived in the spelling but not in the standard pronunciation. According to the OED, spellings with wh, reflecting a [w] glide, appeared in the 15th century, and pronunciations of wholemeal with [w] were recorded in the Survey of English Dialects in Westmorland, Derbyshire, Shropshire and Monmouthshire (which are quite well scattered, so we're not talking a very localised feature). There are also pronunciations of home with [w] out there.

[Curlyjimsam]

From what I understand, it was originally /twɑː/, then raised to /twoː/ during the Great Vowel Shift, then further raised to /twuː/. w > ∅ / C_V[+round], see "who", "whore", "sword", but contrast "swoon", "swollen".

swollen presumably sees reinstatement of /w/ via analogy with the root swell (same story for swore and swear). Not sure what's going on with swoon.

[Salmoneus]

From what I understand, it was originally /twɑː/, then raised to /twoː/ during the Great Vowel Shift, then further raised to /twuː/. w > ∅ / C_V[+round], see "who", "whore", "sword", but contrast "swoon", "swollen".

swollen presumably sees reinstatement of /w/ via analogy with the root swell (same story for swore and swear). Not sure what's going on with swoon.

Presumably also analogy with other parts of the verb? Although ironically the verb itself ('sough') has now lost the /w/.

EDIT: it's possible onomatopoeia and sound symbolism may have played a role here? Compare 'swoosh', 'swish', 'swipe',etc.