Friday 21 August 2009

Kerry, Carrie and Cary

Nedecky Jason wrote
As a North American speaker, I have always struggled with the MERRY/MARRY/MARY set of vowels which are preserved outside of North America. Of course, mine are 100% merged. Your blog entry today has finally prompted me to use the shiny, red "Ask Professor Wells" button on the website. The confusion for me comes not in how the vowels sound, but how they are determined. Is there an orthographical rule that covers this effectively?
So I replied
Generally speaking, the MERRY set (with the DRESS vowel) are spelt with the letter e. The MARRY set (with the TRAP vowel) are spelt with a, in positions where you would expect a short vowel. The MARY set (with the SQUARE vowel) are also spelt with a, or ai etc., in positions where you would expect a long vowel.
Merry, very, terror, berry, bury, Jerry, Kerry have DRESS. Marry, carry, arrow, narrow, baron/barren, arid, charity have TRAP. Mary, vary, area, bear(er), fair(y), precarious, Pharaoh have SQUARE.

There are one or two words where speakers very vary, e.g. Charing Cross. Note the alternation in compare (eə) — comparison (æ). Although a baron is ˈbærən, some bearers of the surname Baron pronounce it ˈbeərən.
_ _

There will now be another break in my blogging: back again on 2 September.


  1. There's also tarry (adjective) that has the CAR vowel, but this can be explained through morphology (#tar# + #y#), whereas tarry (verb) cannot be decomposed into two morphemes, am I right?

  2. Correct, anonymous.

    Also, there are plenty of areas in the U.S. where the distinction is well preserved, notably the Northeast, and other regions that make a two-way distinction (MERRY = MARY, but MARRY has [æ], such as the coastal South. This may be fading with generational change, but my Northern New Jersey accent still separates them.

  3. Just to confirm: non-rhotic dialects preserve the distinction while rhotic dialects do not?

  4. A trick which works most of the time is to replace the Rs by Ls. Thus

    MERRY -> MELLY, which is spelled like BELLY, KELLY or JELLY, all of which have DRESS.

    MARRY -> MALLY, which is spelled like DALLY, RALLY, SALLY or TALLY, all of which have TRAP

    MARY -> MALY, which looks like SCALY, which has FACE. One then has to change FACE (and possibly FLEECE) to SQUARE before "R".

    This method doesn't work for everything, of course. Of the list given above, VERY and BURY are irregular. ARID is also irregular, but it works because ALID looks like VALID which has TRAP.

  5. Ryan: no. All non-North American rhotic dialects (e.g. Scotland, Ireland, West of England, some parts of the Caribbean) preserve the distinction.

    I think that the converse is true: all non-rhotic dialects _do_ preserve the distinction.

  6. My wife (born in the US) and I (born in the UK) had a wonderful transatlantic moment last night. She asked whether I wanted to watch a film which to me sounded like "The Errors". Can anyone guess what it was?

  7. I was born merged, but I decided to separate them :-)

  8. Unrelated, but just heard Russel Brand rambling on about 'labial [sic] fricatives', not having a clue what he's talking about!

  9. The word that I hear even native non-rhotic speakers say with what I believe the "wrong" vowel is "parent", which should be SQUARE. However, in Harry Potter and the Half-Blood Prince, Hermione talks about her "pa-rents" with a TRAP vowel. Not that Emma Watson is the gold standard of what I expect a contemporary RP speaker to sound like... in fact, I hear her say things like "it's not FAIR" throughout the series with a SQUARE vowel that is so open it sound like [æ] or even [a]. So maybe that accounts for her pronunciation of "parents"?

  10. /eə/ has an allophonic variant [æə] among speakers of "Refined RP", so Emma Watson may be one of them.
    In General RP today, a long monophthongal [ɛ:] is a well-known allophonic variant of /eə/.

  11. @Ryan "Just to confirm: non-rhotic dialects preserve the distinction while rhotic dialects do not?"

    The situation is a little more complex in the US, because most US accents are variably rhotic, except perhaps in their broadest versions. For example: natives of a historically non-rhotic region such as New York City will have a three-way distinction in Mary-merry-marry even if their speech is fully rhotic. On the other hand, there are historically rhotic regions of the US that have a two-way split, but which differ from US areas that have a different two-way split.

    Speaking very generally, however, it is certainly easier to preserve a three-way-split if your underlying accent in English is non-rhotic, or at least does not use r-coloration where the letter R occurs between to vowels. In other words, the syllable division must occur before the R, not after it. I have taught any number of fully rhotic speakers how to do this. It's very simple, though of course not easy without a great deal of practice.

    @vp: "MARY -> MALY, which looks like SCALY, which has FACE."

    I take it that in your idiolect, the first syllable of "scaly" has the diphthong [eɪ]. In mine, however, the diphthong is [ɛə], which does not require any "conversion" to the SQUARE diphthong, since in my speech, that IS the SQUARE diphthong. I know plenty of people, however, for whom the SQUARE diphthong beɡins with [e].

    "One then has to change FACE (and possibly FLEECE) to SQUARE before "R"."

    I don't understand what [iː] has to do with the matter. I suppose you must be thinkinɡ of an accent in which Mary-merry-marry all take [iə˞]. I don't think I've heard any with a more closed sound than [ɪə˞], but it's a big world.

  12. “Labial” refers to the lips. “Labial fricatives” include [f] and [v].

  13. I was surprised to see "vary" in the SQUARE list -- I pronounce it with the TRAP vowel. Is that unusual?

    @John Cowan: I'm from Northern NJ and 30 years old, and I definitely preserve the three-way distinction.

    @vp: ooh, ooh, was the movie "The Heiress"?

  14. @Andy: LPD lists /væri/ as an alternative, though less common, pronunciation of "vary" in General American, /veri/ being a more usual one.

  15. what happens to this blog when prof wells dies? i guess he should nominate somebody to take over this blog or bury it when he is gone.

  16. You think there'll still be blogs by then?

  17. As a speaker of Appalachian, Mary-merry-marry all have the same vowel for me.

    As a linguist, it's interesting to note that some people distinguish these; as a human being, I celebrate my homophones.


  18. It's a real pity I only found out yesterday that you were in Argentina last weekend! Had I known, I would've taken a trip to attend your seminar in Buenos Aires.

    Oh well, I guess the two months university "break" caused by the so-called swine flu disrupted all means of communication between professors and students, like myself, tho at least one of them, especially those in the phonetics department, should have mailed us or something.

    Anyway, hope you visit our country again in the future!

  19. Amy Stoller: Do you really find that the rhotic North American accents are variably rhotic? I agree that the non-rhotic ones are variable, of course, with rhoticity increasing (as Labov found) with the class of the speaker, the seriousness of the subject, and other factors.

  20. In my normal northern California speech, all three sound alike. I've even seen "Merry" as an alternate spelling for "Mary."

    I wonder if the non-American rhotic accents which maintain the Mary/merry distinction are those which do not use the retroflex 'r'?

    Just wondering, since otherwise, as Amy Stoller suggested, in order to imitate the vowel distinction, I have to change the syllable break so the r-coloring doesn't change the vowel. "May-ree," instead of "Mar-ee."

  21. @John Cowan: Forgive me for late return to this thread. I just re-read my post, and of course you are right and what I wrote about "most US accents" is just plain wrong. I think that's due to an error in editing on my part, since I certainly know better. I would say that most American accents are rhotic, and with the passage of time, fewer and fewer are non-rhotic. In historically non-rhotic parts of the country (for example, New York City and environs, much of New England), rhoticity, or at least variable rhoticity, is apparently becoming the norm. Labov wrote about this trend in his study of New York accents more some 40 years ago. One would have thought non-rhoticity might be extinct in NYC by now, but I've lived here my whole life and I'm on solid ground when I say such is not the case. I've had occasion to study some Boston accents in the course of my work, and what amazes me is how many speakers will pronounce the same word twice within a sentence or two, once with r-coloration, once without. You can't even pin it down to whether or not the word is being stressed. I think this variability is one of the things that makes learning to speak in a Boston accent so challenging to anyone not native there and to the manner born.

    I'll stand by the rest of that post.

  22. Carrie. Even youtube is giving out the non trap pronunciation of this(my) name. No bueno! Met a guy named Jared who told me even his sister calls him Jered.


Note: only a member of this blog may post a comment.