Why w is so unstable?

Linguistics question: why the consonant w is so unstable and tends to get changed or lost in many languages?
Once upon a time, all the indo-european languages had a w, inherited from proto-indo-european. Then something happened and it started leaving the scene. Ancient greek initially had w, but lost it, at first between vowels and later at the beginning of a word, sometimes being replaced by h. A few dialects preserved it a little more and changed it into v or b. Latin also had w, but swiftly turned it into v. Anatolian languages had w, although they went extinct quite early, and also they were in a place with mostly w-having semitic languages around them. Sanscrit and all the other indo-iranian languages had w, but turned it into v as well. All the germanic languages had w, but turned it into v or lost it completely at the beginning of the words like in north germanic languages. The curious thing is that english kept it unaltered at the beginning of words, despite the many changes it has undergone, although it seems some modern vernaculars like the african american dialect are in the process of changing it into v. Slavic languages replaced it with v as well, Celtic languages had no w. Modern hebrew has also changed w to v as an influence from yiddish, but most other semitic languages kept it unaltered. Chinese, Japanese, and many southeast asian languages have kept it. Some austronesian languages have changed it to v as well, like hawaiian and tuvaluan. Most african, native american and australian aboriginal languages have kept it as well.
Why w is so unstable? It seems the only way to keep it is either to live in insular semi-isolation, like in the case of english or japanese, or to have a large population base, like in the case of arabic and chinese. Also many languages of indigenous/primitive people throughout the world keep it, perhaps due to isolation as well. Why is it so unstable? And it is not so hard to pronounce.

Approximants are generally unstable; look at all of the variations in liquids, for example. The specific processes you describe above are examples are fricatization (or, more generally, fortition). Since [w] is a semivowel, there aren’t many alternatives; there’s not much of a way for it to weaken (without just becoming or something similar), and becoming a fricative is the ‘easiest’ change.

Japanese has kept it, but it only surfaces before a. For example, most verbs form their nonpast negative form by adding -anai to the stem: mat.u ‘wait’ -> matanai, kik.u ‘listen’ -> kikanai, etc. (I’m simplifying a bit here.) The verb ka.u, however, becomes kawanai, with an extra ‘w’. The same ‘w’ surfaces in some other forms as well. It looks like the stem is actually kaw-, and the ‘w’ simply becomes 0 in derived forms except before ‘a’. (Similar processes occur in other languages as well.)

Thank you, Itself! I’ve read Campbell’s Introduction but it doesn’t go into this much detail in sound change discussion. Is there a good webpage listing likely and less likely changes?

Intuitively ‘w’ strikes me as a consonant somewhat tedious to pronounce, and therefore likely to mutate due to laziness. OTOH, some sound changes strike this layman as less likely, e.g. S>R. How about K>P ? Do the k>p shifts in Western Europe stem from a single mutation?

Untrue, at least in regards to Irish Gaelic (which is the only one I’ve studied). It’s just not spelled with a “w”

Let’s see if I can explain this without mangling it too badly. It’s a phenomena called “eclipsis” where the first consonant of a word changes. In words starting with “b” under certain conditions the “b” is pronounced as a “w”.

For instance, “bord” is table which is pronounced pretty much as an English speaker might expect. “Ar an mhord” is “on the table” and the word “mbord” word is pronounced as close to “word”. “Baile”, or “village”, can become “walleh”. So while the letter “w” doesn’t appear in written Irish it does appear in the spoken form with some frequency.

Welsh does use the letter “w” - the mountaineering term “cwm” is from Welsh - but I don’t know if it’s ever pronounced as “w” is in English.

When I took and Irish class a couple decades ago I was told Irish has something like 60 different sounds and only 20 letters, as opposed to the 40 sounds of English and its 26 letters. So written Irish uses a multiple letters to express a phoneme more often than English does.

I don’t know about Irish, but I’m pretty sure English has a lot more than 40 sounds.

I’m pretty sure that one of those two “m*ord” words is a typo, but I don’t know which one. Is the second letter b or h?

When the OP mentions w changing to v, what kind of v? In English, the v is dentolabial, the upper teeth articulating against the lower lip. I believe that is some languages (German, maybe?) it is bilabial. I do know that the f in Japanese is an unvoiced bilabial, while it is dentolabial in English.

I would be interested in knowing if there is any explanation of the w/g alternation between Germanic and Romance languages, e.g. war and guerre, wardrobe and garderobe, and many others.

Without looking anything up and just going off the top of my head:

First of all, the whole question is immaterial with respect to languages that have an alphabet of their own and do not use the Roman alphabet. Japanese does not have any W, nor any of the other letters if our alphabet, but people using our alphabet have arbitrarily and conventionally assigned our letters to phonemes that sound similar to those of Japanese. The Japanese did not do this, the Europeans did it for their own convenience, and used the W according to their own European phoneics…

Similarly, Arabic. The languages of Sri Lanka and Ethiopia make rich use of a phoneme that is represented by W when respelled in our alphabet, with a sound that corresponds to the English W. Mandarin has many words beginning with the “woo” sound, where W is distinct from U, so W is used in transcribing spoken Mandarin into our alphabet.

The W sound occurs in Native American and African languages, but not in French, so colonials use /ou/ to represent W in French, as in the beginning of Ouagadougou. Similarly, the Spanish, where H is silent, use /hu/ for the W sound, as in Huerta.

As for Roman-alphabet languages, there was no W in Latin, so all languages that use W created it later on, and assigned it the value, usually either U or W. Most of those had no written language, until the Roman alphabet was adopted for that purpose. Romance languages do not use W, but Geremanic and Slavic ones do. Polish uses W to represent the V sound, and L with a diagonal line through it to represent the English W sound, as in Wałęsa = vawensa.

That’s right - for instance, French does have the “w” sound, but spells it differently. For example, in France, dogs go “ouah, ouah!” or “ouaf, ouaf!” instead of their English cousins, who go “woof, woof!”

Here’s an interesting web-page on the topic of dogs woofing:

How to “Woof” in 16 Different Languages

About 40 English phonemes is the number I’ve always heard. There are more sounds if you count allophones separately, like the aspirated P in “pot” vs. the unaspirated P in “spot”.

From Wikipedia:

That gives 38-40 phonemes in General American, and 44-45 in RP and Australian.

–Mark

This sort of study falls under the area of phonology. Wikipedia actually has decent information on linguistics. (The general rule is that the more technical and advanced a subject is, the more accurate Wikipedia is about it.) Offline, I’d recommend Hayes’ “Introductory Phonology” or Odden’s “Introducing Phonology.”

It’s not laziness or difficulty. I find the French /ʁ/ and Japanese /ɴ/ tricky to articulate well, but native speakers have no issues at all. (There are legitimate sound changes that occur to make a worse easier to say, like the epenthetic ‘p’ that sometimes pops up in pronunciations of hamster, or the expected Latin *regre becoming regere.)

Generally, what you’re looking for is the idea of articulation: where sounds are pronounced in the mouth. Nearby sounds in a word generally can assimilate to the same or similar articulations to make them easier to pronounce. For example, Japanese tempura is phonologically [teɴpura], but the ‘n’ is pronounced as ‘m’ (I’m skipping over some technical issues here) to agree with the +labial feature of the following ‘p’. There are many other reasons for sounds changes as well, and some of them (e.g., dissimilation) go in the opposite direction. It’s complicated.

It’s more complicated than that. The one-line answer is that Proto-Indo-European stops changed quite a bit in daughter languages. The words ‘spring’ and ‘frog’, for example, arise from the same PIE root through pretty well-known processes (namely, s-mobile, Grimm’s law, and nasal infix). To take a more familiar example, Greek ‘kardia’ (as in, e.g., ‘cardi-ologist’) and English ‘heart’ come from the same root and are dervied by well-known processes. Similarly, the English word ‘punch’ (as in the drink) comes from the same root (borrowed from Hindi) as the English word ‘five’. (Probably. But the Hindi word for ‘five’ is similar to ‘punch’ regardless.)

You more often see changes occurring under more specific conditions rather than swappings sounds wholesale: German ‘s’ becoming English ‘t’ in word-final positions like ‘nuss’ -> ‘nut’, for example.

It should be “bord” and “mbord” - sorry, my spelling skills are even worse in Irish than in English.

OK, thanks. I figured that was it, but you can never take anything for granted in Irish orthography.
And as an aside, am I the only one who originally thought this thread was going to be about the weak nuclear force?

It is a compound letter.

I didn’t know which w he meant, but I was surprised to see the wall of text about the letter as opposed to one of the other uses. The letter was well down my half-formed list of possibilities.

In hindsight it’s my cluelessness about linguistics that wouldn’t have thought to use “unstable” in reference to the evolution of a letter or a phoneme.

My first thought was that it was talking about tungsten, and I thought “What? It’s not unstable at all! And it should be a capital W…”

No, it was my thought as well.

As a letter, W is relatively new. English hasn’t had the W as long as it has had other letters, and I think that in general it is a relatively new letter. I think the newer letters are, the less stable the sounds they represent tend to be, and the less “universal” the letter tends to be. Not all languages have it, and the ones that do don’t pronounce it the same.

B, for example, is pretty old. All Roman-letter languages pronounce it pretty much the same, and the sound it in nearly every language. The Roman B even looks a lot like the Russian letter with the same sound (and the letter that looks like the B is the devoiced version of the same sound), while the beta looks the same, and even the Hebrew beit is slightly similar. (Actually, it’s more similar to the Russian beh.)

J is another new letter, and is similar to W in having an unstable sound, and not representing the same sound across languages, or appearing in all languages with the Roman alphabet.

Except when most people were illiterate, I don’t see writing influencing speech all that much.

Welsh keeps sneaking into all those other languages and stealing all their Ws.

I was thinking either that or it was a zombie about the Bush administration.