>>>[...] It really ought to be spelt as "wo'n't" to show
>>>the loss of the double-l as well as the 'o' but double
>>>apostrophes are a bit messy on the page and, as with
>>>"sha'n't", only one of them is conventionally included.

>>Lewis Carroll spelled them so in his "Alice" and "Sylie &
>>Bruno" biologies.  He also used "ca'n't", and explained in
>>one of his prefaces that the apostophe that stands in for a
>>single "o" in so many "-n't" shouldn't be used to stand in
>>for "no" in one other.

>I never thought about it before, but the switch from "i" to
>"o" is as interesting to me as the squelched apostrophe. By
>analogy with
>		     shall not = shan't,

>one would expect to see:
>		      will not = win't.

>I notice that when I pronounce "shalln't" as one syllable, it
>sounds very much like shan't; the "ll" is almost dropped but
>the sound of the "a" is little affected. When I pronounce
>"willn't" as one syllable, it sounds like "won't." The "lln"
>combination hints at a long "o," just as the "l" slides into
>the 'n," causing the "i" to be heard differently as well.
>Also, I think the juxtaposition of "l" and "n" makes the sound
>of "ll" disappear from "willn't" more easily than from
Right. As usual, there are several potentially independent but mutually reinforcing sources for any language change. For instance, the morphological source uses the O-stem woll that shows up in Gm wollen and regularly inflected past and participle forms, whereas English ordinarily uses the I-stem will that shows up in the German present tense. This kind of root-switching is extremely common in all European languages, and often involves switching a front (/i/, /e/) for a back vowel (/u/, /o/). Take a look at English irregular verbs and you'll see tons of such vowel switches (the technical term is "ablaut"). For another instance -- as you note -- two apostrophes is two too many for most English writers to handle, so there's a good sound orthographic reason, too.

And, as you also point out, there's also a phonetic reason, reinforcing and legitimizing the others. The English post-vocalic /l/ is velarized. That means that when we pronounce /l/ after a vowel, the back of the tongue rises toward the soft palate (the "Velum") while the tip moves up to do the major articulation in the front of the mouth. This is the so-called "dark L", not found in most of the Romance languages, but occurring in, for example, Catalan. The secondary articulation in the velar area is often sufficient by itself to signal the presence of /l/, and the primary gesture is then omitted.

That leaves just a velar gesture, which we hear as the closest sound that exists to it in English, which is the postvocalic /w/ that also appears after tense mid back vowels /o/ and /u/. That is, /o/ and /u/ both have a /w/ after them (which Spanish, for instance, doesn't, even though the vowels themselves exist), and we tend to hear a vowel with a /w/ after it as a back rounded vowel. The upshot is that /Il/ --> /Iw/ --> /ow/ without much difficulty.

For extra points, compare and contrast the British RP pronunciation of /o/ as [Ew].

Phonology is destiny.

--- followup on a related topic:

>> For instance: English has only one /l/ phoneme, but it has
>> two variants - the "dark" velarized [l-with-a-bar-through-it]
>> where the back of the tongue is raised toward the soft palate
>> (which English speakers produce automatically after vowels),
>> and the "clear" non-velarized [l], where the back of the tongue
>> remains low in the mouth (which English speakers produce automatically
>> *before* vowels).

>Examples, please?
Well, I'd be glad to give you some, but this isn't really the appropriate medium. How about if I talk you through a phonetics experiment? And forgive me, please, for taking liberties with your tongue.

Let's take two English words (I'm assuming you're a native English speaker), as simple as possible to avoid complicating factors, one with an /l/ phoneme following a vowel and one where it precedes. Say, awl (or all if you pronounce it the same way) and law.

OK, now pronounce the words, several times; repeat them normally over and over; if you're curious, you can look in a mirror while you're doing it, though it's hard to observe velarization without more serious lab equipment. Observing the tongue during the /l/, by feel and by sound, is what's necessary now. /l/ is pronounced with the tip of the tongue touching the gum ridge behind the incisors, while the sides of the tongue don't touch anything, making the air stream from the lungs diverge to the sides before exiting at the front of the mouth. That's why /l/ is called a lateral resonant.

Both allophones (i.e, positional variants) of /l/, the prevocalic non-velarized [l] and the postvocalic velarized [l-with-a-bar], should be recognizable by what the rear of your tongue is doing when you say the words. In law (hold the /l/ as long as you need to get a feel for where the back of your tongue is), it should be low, and the tongue should only start rising at the tip.

In all (or awl), in contrast, the back of your tongue should be raised, forming a distinct hump rising almost, but not quite, to the soft palate (the velum, whence "velarized"). If you alternate between the two words repeatedly, you'll feel the tongue blade rise after the vowel in all and fall before the vowel in law. Do try to pronounce them as separate words, though -- intervocalic /l/ is a bit trickier.

And there are other allophones of /l/ in English, as well. This is just one example. And every other English phoneme has variants, too. Which we rarely notice, but which form our pronunciation habits and mark us as English speakers whenever we open our mouths to speak any other language. If you can produce a "clear L" after a vowel, for instance, you'll improve your Spanish and French accent. And Spanish and French speakers can improve their English accents by learning how to velarize postvocalically.

  - John Lawler       Linguistics Department and Residential College     University of Michigan

    "Language is the most  massive  and  inclusive  art  we know,  a           - Edward Sapir
      mountainous and anonymous work of unconscious generations."       Language (1921)

More English Grammar   More About Language   The Eclectic Company   The Chomskybot