Diphthong or L?

Septa Station


The other day, in a Philadelphia train station, I overheard a woman ask a ticket taker if a train would take her to what sounded like “Choatenham.” A moment later, I realized she had asked if the line went to Cheltenham, an inner suburb north of the city. Since many Mid-Atlantic accents feature l-vocalization, the syllables “Choat-” and “Chelt-” can sound quite similar (both close to ʧɜʊt).*

In such accents, remember, /l/ becomes something of a vowel or semi-vowel at the end of syllables. Because this vocalized /l/ is usually produced close to the velum, it tends to come off as o (as in British “paw“), w (as in “wet”), or ʊ (as in British/American “put”). In the example I’ve just mentioned, this results in “el” sounding quite close to the vowel in “go” (especially because this diphthong has a fairly front starting-point in broad Philly English).

Mine is not the first such observation. In his Accents of English, John Wells notes that around London, l-vocalization can result in the words “Paul‘s,” “pulls,” “pause,” and “pools” sounding nearly identical. In certain circumstances, it can cause tremendous confusion to outsiders.

Most intriguing to me is the way l-vocalization wreaks havoc on diphthongs. Beyond the above example, I would be little surprised if there were parts of England where “pal” and “pow” are homophones (pæʊ). A number of sociolinguists note that “dull” and “dole” may be homophones in Cockney**. And some varieties of African American and American Southern English feature a vocalized /l/ that is somewhat close to a schwa, thus rendering “feel” and “fear” strikingly similar in pronunciation.

As mentioned at the beginning of this post, particularly salient is the mid-Atlantic tendency to (almost) neutralize pairs like boat-belt, coat-Celt, and home-helm. You can’t describe this as a merger given that l-vocalization is often inconsistent. But it is a prime example of how a large-scale reordering of vowels ending in /l/ is theoretically possible.

Any other good examples of these interesting conflations?

*This seems especially case because due to “l-coloring,” the ‘e’ in “Chelt-” sounds to be somewhat retracted.

**See Wells, 1982, pp. 315 for a discussion of this.


About Ben

Ben T. Smith launched his dialect fascination while working in theatre. He has worked as an actor, playwright, director, critic and dialect coach. Other passions include linguistics, urban development, philosophy and film.
This entry was posted in American English and tagged , , . Bookmark the permalink.

18 Responses to Diphthong or L?

  1. Charles Sullivan says:

    ‘Pal’ and ‘pow’ are homophones (or pretty damn close) in Pittsburghese.

    • Tom says:

      This is notorious in Philly, too. “How” becomes “hal” and “now” becomes “nal”, particularly when followed by a vowel, and especially before the word “I”. (My partner has a fondness for the old Gilbert O’Sullivan song “Get Down”, and I always wait for him to sing “Nal I’m just like a cat on a hot tin roof”. Strange thing is that he seems to do it more with songs from his youth than newer ones.)

      Also, even “towel” becomes “tal”, no matter what follows it, and I’ve mentioned before the l-vocalization in Philly and dollar.

    • Ivan C. says:


  2. What is the difference between vocalizing the L and velarizing the L (to make a “light L”)?
    Assuming they are the same…

    I speak SAE but teach English in Bulgaria, where people get their TV and movies in SAE, but learn something like British Received Pronunciation at school.

    Bulgarian (like British English) distinguishes between light and dark L, and turns consonant-ending L’s into something that sounds to me like W. The further problem, however, is that the light L changes the vowel before it, so “Fall” and becomes “foe” (homophone of “Foal”) and “Paul” become “Poe” (homophone of “Pole”). The thing is, I’m pretty sure I’ve heard people from London do the same thing.

    I wonder if that’s because of continental influence on British English.

    • dw says:

      Simple (oversimple according to some) answer: a velarized /l/ has contact between the tip or blade of the tongue and the alveolar ridge behind the fromt teeth in the roof of the mouth: a vocalized /l/ does not.

    • Ivan C. says:

      A vocalized L, a velarized (a.k.a. a “dark”) L and a “light” L (a.k.a. a “clear” L) are 3 different things. You can consult the glossary here for the definition of vocalization (and many other terms). This blog posting explains the difference between “clear” and “dark” L pretty well I think.

  3. Ellen K. says:

    I’m puzzled by: “/o/ (as in British “paw“)”

    Surely /o/ = British English [əʊ] as in goat? I can’t imagine hearing the vowel in “paw” as /0/ in any accent, even if it’s pronounced [o]. It’s [əʊ] that maps onto the /o/ phoneme.

  4. I’m originally from Calgary, Alberta, Canada, and the normal way people from there say the name of the city in casual speech is in about 2.5 or even 2.25 syllables (the middle one becoming a lengthened /r/), and the /l/ is almost never fully realized in casual speech – the tip of the tongue doesn’t make it all the way to the alveolar ridge. The back of the tongue is raised, however, but the lips aren’t rounded. The result is that it’s similar but not identical to “cowgry.”

    In more formal or emphatic speech, the /l/ usually manages to make it to the alveolar ridge, and the middle syllable manages to be a whole syllable, if a completely unstressed one. (Anyone who pronounces the name like “Cal, Gary” is obviously not from there.)

    • I pronounce “Calgary” in a similar way. I think it’s because of the tendency to assimilate the /l/ before /g/. Something similar happens when I pronounce “Bulgaria,” which for me is usually bɤgɛɚɹɪə. I’d guess that for a large proportion of English speakers some amount of l-vocalization is typical; we nevertheless tend to associate the feature with those accents where it is particularly frequent or associated with lip-rounding.

    • Ivan C. says:

      There is also a velar L ( n.b. not a velarized L, which is different), just to complicate things further. I’ve read that this can be found as an allophone in some accents. Like a vocalized L, this type of L does not have contact between tip (or blade) of the tongue and the alveolar ridge. But unlike a vocalized L, it is still a lateral consonant.

  5. Ed says:

    In Accents of English, John Wells predicted that L-vocalisation would spread in much the same way that non-rhoticity had spread throughout England. He made the comparison between final /l/ and final /r/. It seems to have spread a long way west from London, but it doesn’t seem to have spread very far north at all. Even in Suffolk, it’s not very common.

    In his review of the book for the Journal of the IPA, Petyt (a Yorkshireman) quoted this with an exclamation mark. I don’t think that he was convinced.

  6. Sooryan FM says:

    Adele sings SKY FOE instead of SKY FALL in her newest song…

  7. Laura says:

    You all sound so smart about all of this!! I guess I’m more of a layperson with a love of regional accents!! 🙂 I just found your blog today and I’m happily hooked!! I may have actually found people who get what an interesting subject this really is!! 🙂

    This post reminded me of when I was little and heard my dad (originally from Pittsburgh) say the word “towel” which to my California ear sounded like “tile” (of the ceramic type)!! We had a great time making him say “tile” and “towel” because they sounded exactly alike!!! 🙂

  8. Nick Davis says:

    It looks like I’m a little bit late to the party. Sorry Trawicks. But I just wanted to say that a person can have pre-L neutralizations without having L-vocalization. For example, I don’t make a distinction between the vowels of dull and bull, but I don’t have L-vocalization. So I don’t think it’s vocalization that causes this to happen (not that anyone said it was).