David Madore's WebLog: Father, bother, cot, caught, stark and stork

I already wrote something about English vowels versus spelling, now let's concentrate on one small group of vowels versus accents.

Let's start with an exercise for those who (think they) can speak English: here is a list of words with a vowel underlined, you should (without reading this entry any further or consulting a dictionary) try to group the identical-sounding ones, i.e., decide how many different vowels you can hear in this list and which words contain which:

bother, brother, caught, coral, cot, court, dawn, don, farther, father, for, force, forest, four, hoarse, horrid, horse, law, morning, mourning, north, palm, psalm, Shaw, shore, stalk, stark, stock, stork, thaught, thought, war, warp, wash, watt

(Write down your answers and your doubts before reading any more of this, so you won't be tempted to change them. Remember that only the pronunciation matters: e.g., son and sun would be grouped together if they appeared in the list.)

Now, what should be the answer? First, let's cross out the odd word out: the vowel in brother does not sound like any other in the list, it is the same vowel as in son and sun and also mother and other. I included this word as a kind of control: if you think brother rhymes with bother, then either English is not your native language, or you are unaccustomed to noticing the differences between vowels, or your variety of English is unusual and I'd like to know more about it.

Other than that, everyone should agree with at least the following identifications:

• (‘ä’) father and psalm have the same sound, and generally palm also;
• (‘är’) farther and stark have the same sound;
• (‘ŏ’) bother, cot, don, stock and watt have the same sound, and generally wash also;
• (‘ŏr’) coral, forest and horrid have the same sound;
• (‘ô’) caught, dawn, law, Shaw, stalk, thaught and thought have the same sound;
• (‘ôr’) for, horse, morning, north, stork, war, warp have the same sound;
• (‘ōr’) court force, four, hoarse and shore have the same sound, and sometimes mourning also.

(I've used diacritics rather than IPA symbols for these sets, because the actual phonetic realization can vary considerably, as I will describe.)

If you make distinctions among these groups (say, between cot and don), it's probably because your ear is overfussy and cannot ignore the context. On the other hand, I'm definitely not saying that there aren't any more vowel identifications to be made than those described above: for example, if you think father rhymes with bother, that's fine (as I'll be explaining in a minute, most North American speakers should say that). In fact, a sizable number of native English speakers might even consider that all the vowels above (all except brother, that is) have the same sound. And, as we shall see, almost nobody distinguishes ‘ôr’ and ‘ōr’.

Now that we have distinguished seven groups of words, how do people actually pronounce these vowels?

British English Received Pronunciation makes a distinction between ‘ä’, ‘ŏ’ and ‘ô’: the vowel ‘ä’ is pronounced as the long open back unrounded vowel [ɑː], the vowel ‘ŏ’ is short, rounded, and slightly less open, [ɒ], and the vowel ‘ô’ is long, also rounded, and yet less open, [ɔː]. The essential distinction is that of roundness: ‘ä’ is pronounced with unrounded lips whereas ‘ŏ’ and ‘ô’ are rounded. Also, ‘ŏ’ is breve whereas the other two are long. The degree of openness varies (RP ‘ô’ is transcribed [ɔː], but it tends toward [oː]), but this is probably less important. The variants with ‘r’ are pronounced exactly as those without and, since RP is non-rhotic, there is generally no consonant to distinguish. So ‘ä’ and ‘är’ are identical (father and farther are pronounced the same), and ‘ô’ and ‘ôr’ are identical, and so is what we have written ‘ōr’ (caught and court or Shaw and shore are pronounced the same); as for ‘ŏr’, it only occurs with intervocalic ‘r’, so that it is pronounced, but the vowel is otherwise the same as ‘ŏ’. Since I know very little of other British pronunciations, let alone Southern Hemisphere variants of English, I will now concentrate on North America.

North American pronunciations typically merge ‘ŏ’ with ‘ä’ (except in a certain sense before ‘r’, see the end of this paragraph). So American father rhymes with bother, both being pronounced with a long open back unrounded vowel [ɑː] very similar to the ‘ä’ of English RP. The main exception to this is Eastern New England (and most famously, Bawstawn, i.e., Boston) and Pittsburgh: in those areas, ‘ŏ’ merges with ‘ô’ instead, both being rendered as a long open back rounded vowel [ɒː] (furthermore, since Eastern New England speech is partially non-rhotic, con and corn are identical). Elsewhere, the pronunciation of ‘ô’ varies quite a bit, but it is typically more open than in British English: while it is transcribed [ɔː], it could tend to [ɒː] (hence the perception of Bostonian ‘ŏ’ as “aw”). Before ‘r’, it tends to be closer (except where ‘ōr’ has survived, see below), so ‘ô’ and ‘ôr’ may not have identical vowels. Also before (intervocalic) ‘r’, the vowel ‘ŏ’ (hence, ‘ŏr’) has become as in ‘ôr’, except in the North-East where it is unrounded and identical to ‘är’.

In the Western part of the United States and the Northern Midwest (and also Alaska, but excluding the San Francisco Bay area), and pretty much all of Canada, the vowels ‘ô’ and ‘ŏ’ have also merged (this is the caught–cot merger) when not followed by ‘r’: the resulting vowel is transcribed as [ɑː], but it can be slightly rounded; this merger does not take place before ‘r’, so while caught and cot become identical, stark and stork do not (they remain as [ɑːɹ] or even [aːɹ] for ‘är’ versus [ɔːɹ] or even [oːɹ] for ‘ŏr’, ‘ôr’ and ‘ōr’, again with variations).

The distinction between ‘ôr’ and ‘ōr’ is lost in almost all varieties of English. Some isolated areas still have it to some extent (e.g., Louisiana and Mississippi), in which case ‘ōr’ is distinguished by the fact that it is closer and/or partially diphtongized, as is the vowel in toe or goat.

My personal story with regards to all of this is that I learned English in Toronto, Canada, which has the caught–cot merger: so I learned English with ‘ä’=‘ō’=‘ô’ all pronounced as a slightly rounded version of [ɑː], whereas ‘är’≠‘ōr’≅‘ôr’=‘ŏr’ pronounced as [ɑːɹ] and [ɔːɹ]/[oːɹ]. Because of this, I was extremely confused: I could distinguish about three vowels in the whole set, but the distinction I saw did not at all match the one found in dictionaries! I occasionally entertain the idea of revisiting my pronunciation of English and forcing myself to make a maximal distinction in the set (pronounce ‘ä’, ‘ō’ and ‘ô’ all differently, though it is difficult to do so in a way that is compatible with a generally North American accent; and also pronounce ‘är’, ‘ōr’, ‘ôr’ and ‘ŏr’ differently). It is quite possible to change one's pronunciation and to learn to make distinctions: I've done something of the sort in French, and I now distinguish the ‘in’ and ‘un’ nasals ([ɛ̃] and [œ̃]) while initially I did not. It's a good ear training exercise.