Marry, Merry, Mary

A Bride[Ed. Note: In an earlier version of this article, I suggested I pronounce ‘marry,’ ‘merry’ and ‘Mary’ differently.  The opposite is true.  I pronounce them alike.]

Do you pronounce ‘marry,’ ‘merry,’ and ‘Mary’ the same?  I do, which makes me a fairly typical American.  That being said, I represent a generation shift.  The Linguistic Atlas of New England, published in the 1930’s, showed that the region I grew up in still made a distinction between these three words.  This is yet another New England dialect feature which is receding.

By the way, if you answered ‘yes’ to the opening question, you may not understand how these three words sound different.  In many British (and some American) accents, ‘marry’ (and other ‘-arry’ words) are pronounced with the same vowel in ‘cat;’ ‘merry,’ (and other ‘-erry’ words) are pronounced with the same vowel in ‘pet;’ and ‘mary’ (and other ‘-ary’ words) is pronounced with the same vowel as that found in ‘fair.’

As I mentioned, this three-way split is found in some American accents as well, although it’s a bit mysterious as to why the accents that preserve the distinction do so.  The split can be found in Boston and New York, which might suggest the feature is related to non-rhoticity (i.e. ‘r-lessness’).  But the split is also found isPhiladelphia, a city where the local accent in rhotic (i.e. ‘r-ful’).

It’s not really appropriate to call this distinction a ‘split,’ anyway.  Really, the more important question is why many American accents merged all three of these words into the same vowel (the vowel in ‘fairy’).  My only thought is that it perhaps has to do with the retroflex r common in American English (i.e. an /r/ formed with the tongue pulled back).  I find it slightly difficult to pronounce the vowel in ‘cat’ before this type of /r/, but that’s just a personal observation.

There are also incomplete mergers.  For example, in some American dialects ‘marry’ is kept distinct while ‘merry’ and ‘mary’ have merged.  I also wouldn’t be surprised if some people maintain the distinction only in certain words. For example, I’ve caught myself using the vowel in ‘cat’ in the word ’embarrassment,’ perhaps a rare case of picking up a pronunciation common to New York City (this feature isn’t stigmatized in that dialect, so you can have people maintain the distinction who otherwise have few New York accent features).

Are you an r-mergerer, or do you preserve the distinction?  And why?


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 English Phonetics and tagged . Bookmark the permalink.

108 Responses to Marry, Merry, Mary

  1. Dayna says:

    Marry, merry, and Mary are all the same to me here in the Midwest – NW Indiana to be exact.

  2. dw says:

    Really, the more important question is why many American accents merged all three of these words into the same vowel (the vowel in ‘fairy’)

    Two ideas:

    Loss of distinctive length

    In my native British accent, the main difference between “Mary” and “merry” is length, with the vowel in “Mary” about twice as long as the one in “merry”. In IPA, [ˈmɛɹi] versus [ˈmɛːɹi ~ ˈmɛəri]. However, in American accents the long-short vowel distinction is generally less important that across the pond.

    This explains the “Mary”-“merry” merger (and also the related “nearer”-“mirror” merger), but not why “marry” should also be merged.


    In, say, RP, there is a problem syllabifying a word like “marry”. The word is pronounced /mæri/ and must be syllabified as either




    What’s the problem? In neither of these cases is the first syllable possible in isolation. In other words, neither */mæ/ nor */mær/ is an allowable monosyllable in RP.

    Why is that a problem? Well, in some sense, it isn’t. RP and other speakers of what Wells calls “Type II accents” (e.g. most Southern English, Welsh, Australian, New Zealand) obviously don’t have a major issue with this ambiguity. However, it seems plausible that a language in which every syllable is possible in isolation is, in some sense, simpler than one in which this is not the case. In any case, this would explain how “marry” got added into the mix.

    • trawicks says:

      For many Americans, it seems to be less a matter of there being no long/short distinction as much as there is a ‘tense’ category of vowels that includes FLEECE, GOOSE, THOUGHT (as might be expected in British English), but also TRAP. Which means that, unlike British English, the ‘a’ in MARRY, even if it were distinguished, would be closer to the vowel in MARY in both position and length.

  3. Jan Freeman says:

    In your first paragraph, do you mean to say you DO pronounce the three words the same? Otherwise I don’t understand the rest of the graf …

    My N. Ohio accent, like Dayna’s Indiana one, uses the same vowel for all three. I can hear the difference, but I find it very hard to produce those sounds.

    • Martienne says:

      I was going to make nearly this identical comment. Funny that the two people who noticed this mistake were both from the same region of the same state.

      You also went on to say the accent in Philadelphia “in rhotic” instead of “is rhotic”.

    • IVV says:

      Agreed. The Marry-Mary-Merry merger is typical for most Americans. The Northeast maintains the distinction, and the Southeast has merged Mary and merry, but not marry. However, the West and Midwest pretty much have a complete merger.

      • Mark Paris says:

        I am from NW Georgia and I pronounce them all the same, as do most of the people I know. I am pretty sure, but not certain, that there was a difference in some earlier generations. My grandmother might have pronounced marry differently, but I can’t remember for sure.

      • Mark Paris says:

        I should also note that I have heard Mary pronounced like fairy but with the syllables strongly separated (may-ree). That is also probably more typical of earlier generations than my own.

      • IVV says:

        I am probably quite incorrect about the situation in the Southeast, thanks.

    • trawicks says:

      Thanks for pointing that out, guys. I do, indeed, merge all three!

  4. Aaron Bauman says:

    In Southeast Ohio (and Southeast Pennsylvania) they’re all pronounced the same. I can hear the difference only if I’m paying close attention, and I do not pronounce them differently unless I’m concentrating really hard.

    This is the same shift between Aaron / Erin, which is particularly applicable to me personally. Folks from the Northeast have a hard time accepting that I pronounce my name with a closed “a”, so much so that I’ve recently begun training myself to pronounce my own name differently. Now I feel like I’m yawning or yelling when I tell people my first name: AAAAAAAHHRon.

    • trawicks says:

      It’s kind of a difficult way of pronouncing -ar words if you didn’t grow up speaking that way. Like you, when I’m imitating someone with the distinction, I stretch out the vowel.

    • Michael says:

      Southeastern PA is one of the areas in which all three are distinct. Did you mean Southwestern PA (i.e., the area of PA near Southeastern Ohio)?

    • Stauffer says:

      You’re wrong about Southeastern PA. I’m for New Hope and everyone I’ve grown up with them differently.

  5. NemaVeze says:

    I’m from Philadelphia, so I have the three-way split (as well as the cot/caught distinction). I have known Philadelphians who merge “merry” and “Murray,” though – also known as the “ferry” / “furry” merger – so they’ll wish you a Murray Christmas.

  6. IVV says:

    They’re pronounced the same in California. I’m not sure how Merry and Mary would be pronounced differently, but comparing the sound to other NY/NJ accents I hear, the big issue is with /ær/. By the time you get west of the Ohio River (and perhaps the Appalachians), I don’t think I’ve ever heard /ær/. There’s /er/ (“air”) and /ɑr/ (“are”) but no /ær/.

    Aaron: It’s crazy, isn’t it? Despite being homophones or near-homophones, Midwesterners and Westerners don’t have problem differentiating between Aaron and Erin, or Don and Dawn, I find. We’ll all say it’s the same sound, and it’s similar enough to confuse the Northeasterners, but somehow, we just know who we’re talking about.

    • m.m. says:

      I’m pretty sure you can ‘chalk’ that up to context :]

    • Ellen K. says:

      Except, I think don/dawn it’s the east coast where they sound alike, backwards of Aaron/Erin.

      • IVV says:

        Not in New Jersey!

        This merger actually caused a problem in one meeting I attended. The meeting was in New Jersey, and most of the people in the room were New Jerseyans. The speaker was a California transplant, talking about how everyone should just get their data in to her, and she’ll pass it all on to Dawn, another analyst. The New Jerseyans all nearly jumped out of their seats, because they thought she said Don–the CFO! A bit later, it became clear what the mix-up was, but she was left protesting “It’s the same sound! Don and Dawn!” She stopped for a moment, and in her best Joizey, she said, “Uh… Dwaaauuun.” The others nodded in approval, with a slight self-deprecating laugh.

        Don is NOT Dawn on the East Coast.

        • Ellen K. says:

          New Jersey is not the entire east coast. I didn’t mean the entire east coast. Just that the area where they sound a like is part of the east coast. In particular, Boston.

          Now looking into it, I see I was wrong on thinking it limited to the East coast. But I was not wrong in placing it there.

          Didn’t actually realize that people with the caught/cot merger pronounce those two words alike. (This midwesterner doesn’t.) Though one can have a don/dawn merger without the cot/caught merger.

        • m.m. says:

          Erm, don-dawn/cot-caught pairs belong to the same merger, which is indeed usually confined to the boston/new england area of the east coast. The father-bother merger would be the reverse, not found in boston, but along the rest of the east coast [though there is some distinction in the NYC region via the low back chain shift]
          It’s interesting you say that you can have the don-dawn pair but not the cot-caught pair, because its been found that the merger is generally more advanced before /n/ than /t/. and recalling your parents being from chicago [generally not cot-caught merged land], growing up in st.louis [generally not cot-caught merged land] and living in kansas city [transitional area] it makes sense you might say that 😀

        • Ellen K. says:

          No, I say that because I can read. Has nothing to do with where I’m from. I know how to, like, read up on these topics. Pointing out information, that, based on what I’ve read, is generally accepted as true by those knowledgeable. No personal investigation involved in that.

        • m.m. says:

          I was referring to

          >>Though one can have a don/dawn merger without the cot/caught merger.<<

          both word pairs belong the same merger.

          The other stuff was a tangent that explains cases where the /n/ pair is merged but the /t/ pair isn't in transitional mergers.

        • Ellen K. says:

          Dang, how dense are you? We seem to agree. You are saying the same thing I did… one can have the Dawn/Don merger without the cot/caught merger. Just saying it differently. And, again, my background has nothing to do with my knowing about what you call a “transitional” merger. Me, I prefer not to forecast the future.

    • dw says:

      But as Amy points out below, you shouldn’t really think of it as /ær/.

      If you’re a mergerer who wants to know how a non-merged speaker would pronounce “marry”, I would suggest;

      1. Say “mat”
      2. Remove the final /t/ so you have just /mæ/
      3. Convince yourself that this /mæ/ can be a freestanding syllable.
      4. Say /mæ/ and then /ri/.
      5. Repeat until you can get the two syllables close together as a single word.

      • IVV says:

        Or, I can keep saying /meri/.

        • dw says:

          When you said “I’m not sure how Merry and Mary would be pronounced differently”, I interpreted that as an interest in how they might be pronounced differently. (Admittedly I talked about “marry” rather than “merry”, but just change “mat” to “met” in the suggestions above and, mutatis mutandis, everything will still work).

        • IVV says:

          I apologize for being snippy.

          But for me, Mary and merry would still be homophones, even if starting from “met.” From what I hear, it seems to be more about the length of the vowel, instead of a different sound. If that’s it, then that’s where I have the merger–the length of the vowel is the same. Chary is cherry, ferry is fairy.

          Does that sound right?

        • Macias says:

          Yes, if you’re an idiot.

      • IVV says:

        I get it, though, and have been able to say /mæri/ when I think of it. Thanks.

  7. boynamedsue says:

    This is probably not really related, but (in northern England) I used to have a great-aunty whose name was “Mary” pronounced “marry”. She was of my grandparents generation so born 1910-1925, and had obviously always been called that. However, my grandparents knew two Marys who they’d met after the war, Dutch Mary (“she’s really a German you know!”) and Irish Mary (“She’s very nice, but she’s a [whisper] cathlick [/whisper], like”) whose names had the same vowel as “fairy”, which is much longer than that of my aunty Mary’s name.

    I’ve often wondered if that was the last trace of a separating effect, or just an isolated pronunciation difference related to the name “Mary”.

    • Eimear says:

      In Ireland the name pronounced like “marry” is spelled “Marie”, or to look at it another way, women of a certain generation ( say 50+) called “Marie” are likely to pronounce it that way. (see also “Black Maria”)

  8. Amy Stoller says:

    Native of New York City here. I preserve the merry-Mary-marry distinction (as you rightly point out, it’s not a split; where the distinction is not preserved, there is a two-or three-way merger).

    So far as I know, the process creating merger is r-coloration. The r, however, does not need to be retroflexed; it can be, and often is, bunched (also called molar, or braced). This exerts a strong “pull” on the preceding vowel.

    Where distinction is preserved, there is no r-coloration on the vowel before the letter r. Instead, the r is a consonant at the beginning of the “next” syllable.

    In other words:

    Merry is not Merr-y [ˈmɛə˞.i] but Me-rry [ˈmɛ.ɹi]
    Mary is not Mar-y [ˈmɛə˞.i] but Ma-ry [ˈmɛə.ɹi]
    Marry is not Marr-y [ˈmæə˞.i or ˈmɛə˞.i] but Ma-rry [ˈmæ.ɹi]

    I find it useful, when teaching, to think of the letter r as indicating two sounds in English (southwestern BrE and non-regional AmE): The Consonant R [ɹ] and the Vowel R [ɜ˞, ə˞].

    Of course these two sounds don’t exist as black vs white, but are on a continuum from black through grey to white. But when introducing the concept, I start with absolutes. I get around to shades once this starting concept has been absorbed (it seldom takes long, but a few of my clients struggle with it for a while).

    Once we’ve discussed the shades, I can add that this is why r is often called a semi-vowel. You could just as easily call it a semi-consonant. Personally, I like distinguishing the vowel from the consonant better.

    You can sometimes observe a similar “pull” on vowels followed by a velarized or pharyngealized “dark” l. (In speakers who don’t think of the l in PALM as silent, the vowel is likely to be different from the vowel they use in SPA, whether they actually have an apical- (or laminal-)alveolar component in their dark l or not.)

    While I don’t call the Clear (or Light) L a consonant and the Dark L a vowel, I think it might be useful if we did! L is as much a semi-vowel as r is. So I do point out the analogy to my clients, while explaining that the terminology is different.

    DISCLAIMER: In the interest of total accuracy, I will admit that while I’m perfectly capable of pronouncing Mary with ɛə (and will do so when needed), in my everyday speech, I tend to pronounce Mary exactly the way I pronounce marry; i.e., with æ. I do the same with the name Cary. But I do use ɛə in other words in that set, such as wary. I think this Mary-Cary oddity may be a very NYC thing, but perhaps it’s just my idiolect.

    • trawicks says:

      The problem, as you point out, is that the English ‘r-colored’ consonant is somewhat unstable. I once saw a study that measured people’s tongue position when pronouncing this vowel, and it’s often inconsistent, even within single speakers! It can change from bunched to retroflex depending on the environment.

  9. Chaz says:

    Wait a minute…you’re saying that you’re a typical American because you make a distinction between Mary, merry and marry? I would’ve thought it would’ve been the other way around, considering that most Americans do not make this distinction. Are you sure that’s what you meant to say?

    trawicks: “The split can be found in Boston and New York, which might suggest the feature is related to non-rhoticity (i.e. ‘r-lessness’).”

    But then again the “split” (i.e, the original situation) is also found in the entire island Ireland and in Scotland, which are known for being rhotic areas.

    • dw says:

      Yes. Being rhotic is a necessary, but not sufficient, condition for the merger.

      It is also notable that some (most?) Irish and Scottish speakers have retained the original situation of

      * SQUARE = FACE + /r/
      * START = TRAP + /r/

      and so syllabification of a word like “marry” does not pose the problem for them that it does for RP speakers.

    • Ellen K. says:

      Scotland also has a different kind of r, one that doesn’t color the vowels.

    • trawicks says:

      Sorry, Chaz, that was a mistake on my part: I indeed do NOT make a distinction between the three. Corrected!

  10. Chaz says:

    Also I don’t think the retroflex r is common in American English and it annoys me that I keep reading everywhere that Americans use a retroflex r. I, an American, use the so-called “bunched tongue r”. The retroflex r sounds quite distinct to me. It sounds more like the sound someone from the West Country or Ireland, especially Ulster, would use.

    • trawicks says:

      As I responded to Amy, above, there’s really more a spectrum of r’s in American English.

      • Chaz says:

        Well, I just don’t see how mine could ever be described as retroflex. The tip of my tongue is actually facing downward during articulation.

        • trawicks says:

          There are people who are entirely consistent as well–it apparently depends somewhat on the shape of individual vocal tracts. Even though our tongues move in different positions, the resulting consonants are accoustically rather indistinct. It’s quite fascinating how American English speakers are able to naturally position their tongue in such a way that it accomodates all these factors.

        • Amy Stoller says:

          What you describe, Chaz, is a braced/bunched/molar r. It’s entirely normal, it’s just different from mine, which is tongue-tip up, and often slightly (not fully) retroflexed. Your type of r is extremely common in the US, although not universal. I don’t know which type of r is more common in the US overall, but it wouldn’t especially surprise me to learn that bunched r is in the lead.

          Ben, you’re right, I’m inconsistent, as are many other AmE speakers. I sometimes use a bunched r in /kr/ and /gr/ clusters myself. But I also use apico-post-alveolar in those clusters. I don’t know which I do more often, but I suspect it’s the latter. I know my “pirate r” is fully retroflexed, although many people bunch theirs (I can, too – but I don’t). I think it would be fair to consider apico-post-alveolar my “default” r. It’s a very, very strong tendency.

        • Amy Stoller says:

          Would like to add, Chaz, that anyone who begins a sentence with “Americans use” has forgotten how large a country America is, and how varied its population. I’m not surprised you find it annoying. Of course, to say “many Americans use a retroflex r” would be accurate, as would “many Americans use a molar r.”

  11. Stan says:

    I pronounce them distinctly (west of Ireland), and I don’t think I’ve met an Irish person who didn’t.

    • trawicks says:

      I haven’t met one either. Although there are a lot of other interesting mergers or near-mergers in Hiberno-English regarding these types of words, since Irish dialects (theoretically) don’t have a fully merged NURSE vowel.

  12. Discodoris says:

    ummm, I’m southern English – but I’d note that there’s no consistency to Mary rhyming with fairy in the UK – different British dialects have different pronunciations of it, including the Scot MAHree, or MahREE (sorry, I don’t know the proper phonetic spellings) and the Welsh which I don’t even know how to put on paper. I can distinguish Scouse and Geordie pronunciations from RP as well…

    • boynamedsue says:

      Yeah, but it’s about the vowel system of the dialect, not how the sound would be represented in RP. The word “fairy” almost always has the same vowel as “Mary” in the various British dialects, and that vowel is different from the one in “Carry” and “marry” and the one in “Very” and “merry”, no matter how different these vowels are to RP, the distinction between the three groups of words is the issue.

      • Discodoris says:

        I’ve never heard a Scot pronounce fairy as FAHree at all – that doesn’t sound right to me at all – I have heard the Mary/Marry not being split in some British dialects – not because marry has moved towards fairy, but that Mary is pronounced with the short a vowel, which is entirely unlike the way that some American dialects have moved them together…

        • boynamedsue says:

          It’s hard to say what you mean by MAHree, what would an RP word that rhymes with it?

          It may be that there are older prons of “Mary” in various dialects that are unrelated to the “fairy/scary” group, in fact I posted about this possibility earlier on this thread.

        • dw says:


          It’s hard to say what you mean by MAHree, what would an RP word that rhymes with it?

          “Safari”, perhaps?

        • boynamedsue says:

          If it’s RP “Safaaahri”, then the name is probably Mhari not Mary.

  13. Julie says:

    I pronounce all three the same (as near as I can figure). For years I wondered about the distinction other people made because the alternates seemed impossible to pronounce. There are only so many vowels that can come before r, a. Only recently have I noticed that breaking the syllables differently makes it work.

  14. Julie says:

    Can you please delete that? It glitched and sent itself while I was editing it. Thanks. Delete this one too, if you like.

  15. Katie says:

    FWIW, I pronounce each one slightly differently. I grew up in CT near NYC, but I spent a lot of time with grandparents who had strong eastern CT accents. For example, my grandmother would talk about sitting in a “chaiy-yeh” (chair–two distinct syllables). I now live just south of DC, so still on the east coast.

    • trawicks says:

      Interesting your parents grew up in Eastern CT: that’s where I’m from! Your grandparents are typical of older people of the region. Alas, older New England accent have largely been replaced by General American English in my generation.

  16. Peter S. says:

    I always assumed that the rule for General American is: if a vowel comes before an ‘r’, it becomes one of the r-influenced vowels. The number of them depends on your accent, but they always include the vowels in par, pair, peer, pore, and purr. I think that depending on your dialect, there may be a few more (maybe pure, poor, power, pyre). But note that pyre and power are often pronounced with two syllables: pie-er and pow-er.

    Words with syllables ending in a long vowel, such as highrise and hero, are exceptions.

    • Ellen K. says:

      For me, hero has the vowel of peer, and “high-rise” is two distinct morphemes. I think if it became a single morpheme, it would come to be pronounces (the first vowel) like par. So, not exceptions.

    • Ellen K. says:

      mono-morphemic examples with the vowel in “high” (words mentioned on another thread, on another subject): pirate and tyrant.

      The R goes in the next syllable. Pyre, with no next syllable (and any rhyming words) adds a vowel before the R, the one of purr.

  17. Randy E says:

    There are 28 comments and nobody’s mentioned the Newfoundland traditional song Mari-Mac….

    The song has a lot of fun with the words Mari, Merry, and Marry. They all sound the same to me in the song, but I’m not trained to hear the differences. I’m pretty sure I pronounce them all the same too.

  18. A Mary says:

    Native Arizonan here but I speak with weird mixture of middle Tennessee and East Texas accents courtesy of both sides of my family. I pronounce all 3 the same, as do all other native Arizonans I’ve run across. FWIW, I share the first name of Mary with my mother and grandmother.

  19. Angus-Michel says:

    Most Canadian accents that I’m familiar with have this merger, with the possible exception of old-fashioned Anglo-Montréaler English, which I think merges two of those but distinguishes that from the third.

    Interestingly, because I dated a girl who was originally from Long Island and had a sister named Kerry, I now pronounce Kerry differently than I do ‘carry’. For example, my current girlfriend is named that (only it’s spelled slightly differently), and although we both have accents that have the merger, I instinctively pronounce it [kɛːɹi] or something similar.

  20. Charles Sullivan says:

    This may be a completely different phenomenon (although I think proximity of the letter R plays a similar role):

    I was once talking to my cousin from northern New Jersey (I’m from Pittsburgh) about horror movies. She though I had said ‘whore’ movies.

    For me the first syllable in ‘horror’ and ‘whore’ are pronounced the same (like the word ‘or’). For her, ‘whore’ sounded like the first syllable in ‘or’, but ‘horror’ sounded completely different. Kind of like the vowel sound in ‘car’.

    • Dw says:

      Yes: there’s the possibility of a sorry-story or coral-choral merger (anyone got a better name for it?)

      What’s interesting, however, is that I think there are many Americans who do have these contrasts, using LOT in words like “sorry”, even though they may not contrast Mary-merry and nearer-mirror. I’d suggest that this is related to the fact that words like sorry can easily be syllabified as so-rry as a result of the father-bother merger.

      • Chaz says:

        Actually, I don’t think most Americans use LOT in words like “sorry” (although I realize you didn’t say most, you said “many”). I think most use START, so “sorry” is like “sar-ee” with the “r” in the first syllable. The 2 sound quite distinct to me. I think that’s one characteristic of American accents is that the “r” tends to be in the first syllable in words like “Mary”, “marry”, “merry”, “mirror” (although I have to admit I pronounce it just like “mere” and no I’m not Larry the Cable Guy), “borrow”, etc.

        • dw says:

          So, for these speakers, “sorry” rhymes with “starry” rather than with “safari”?

        • IVV says:

          I’m not sure?

          My LOT is different from START. If I were to break down the sound, I’d say that START had a LOT vowel, followed by the R consonant, but I’d wager most linguists would say the vowel is R-colored.

          “Sorry” is /sɑɻi/, which rhymes with both “starry” and “safari.”

        • Chaz says:

          No “sorry” rhymes with both “starry” and “safari”. The “r” is in the same syllable as the vowel that precedes it in all 3 of those words.

        • Peter S. says:

          I think this is another instance of the r-influenced vowel phenomenon of General American. In “safari”, since the broad “a” is followed by an “r”, the vowel becomes the r-influenced vowel in “par”, “sorry” and “start”. It may or may not be pronounced the same phonetically as the vowel in “father” and “lot”, but it’s perceived as different.

    • m.m. says:

      The way I see it: the NYC/philadelphia/carolinas regions are known for having [ɑ] instead of historically [ɒ] before intervocalic R, the rest of the states [like Pennsylvania] have a mix of [ɑ] and [ɔ], ‘horror’ taking [ɔ]. ‘whore’ and ‘or’ typically ranging from [ɔ]~[oʊ] depending on stress etc.

      for us [ɔ] users, ‘horror’ has a ‘weak form’ [hɔɚ] which elides the first R, which can then indeed sound very similar to [hɔɹ] ‘whore’, especially to your [ɑ] ‘horrible’ using cousin who probly isnt used to the shorter form of ‘horror’.

      @Chaz: for many of us, LOT = START, in which case many is right.
      Interesting about the syllabification. im not fully nearer-mirror merged, but my “mirror” is definitely close to “nearer” if its “mir-ror”, but definitely farther when “mi-rror”. And trying to say marry/merry/mary unmerged is definitely easier shifting that r from the first to the second syllable. neat find!

      • Chaz says:

        In order for LOT = START as an American, you would have to be a non-rhotic speaker. And there aren’t many non-rhotic speakers left in America, so I don’t see how it could be many.

        • dw says:

          I think that m.m means that START = LOT + /r/.

          In other words, if LOT is [ɑ] then START is something like [ɑ˞] or [ɑɚ].

        • Chaz says:

          I disagree. I don’t associate LOT with START. I think of them as 2 completely different sounds.

        • IVV says:

          And I think of them as the same sound.
          LOT is /lɑt/ (or /lɑʔ/)
          START is /stɑɻt/ (or /stɑɻʔ/)

        • Chaz says:

          You should be using brackets [ ], not slashes / / if those are truly your surface realizations of start and lot. I highly doubt they are though if you’re American.

        • IVV says:

          I’ll use brackets, then.

          But if it’s not that, what is it?

        • m.m. says:

          Like IVV, they are the same for me [were talking strictly vowel sound of course] The only reason they wouldn’t be the same sound would be if you were not father-bother merged, in which case they would indeed each have a different vowel.

        • Peter S. says:

          In General American, a vowel followed by an “r” is often not perceived as the same vowel when not followed by an “r”, even if they are pronounced the same. I actually pronounce the vowel of “far” nearly the same as, although slightly differently, than that in “lot”. But I perceive them phonemically as different vowels. The same holds for “peer” and “pit”. And “core” and “caught” are pronounced quite differently, even though they are both reasonably close to the IPA vowel /ɔː/.

        • benw says:

          Hum. For me as a more-or-less Gen. American speaker, LOT and STAR have basically the same vowel except for the /r/ in the latter, but START has a rather different vowel, more like in CUT. And contrary to Peter S.’s comment, the vowels in CORE and CAUGHT are not at all similar, even if you ignore the r coloring of the former. For me (with cot-caught distinction), the vowel in CORE is much higher than the CAUGHT vowel; CORE has a vowel close to that of CONE except for the r-coloring of the former and the slight diphthongization of the latter.

    • Rodger C says:

      Once in California I was talking to another transplant, she being from the Northeast, and mentioned “the horrors of war.”

      She: “The whores of war?!”

      Me: “No, no, the HAHRUZ of war!”

      She: “Oh.”

  21. CaitieCat says:

    This is one of the ways my Canadian accent is seen through as not-native, as I have this distinction fully. I was born in the UK (Watford area), and moved here when I was 10, but for a range of reasons, never fully adopted the Canadian accent. I can ape it, if I try, but it’s not natural for me, and I’m occasionally caught out by vocab.

    And when I sing Mari-Mac, by the way, there’s a clear distinction in each word:

    Mari Mac’s mother’s making Mari Mac marry me
    My mother’s making me marry Mari Mac
    Well I’m going to marry Mari for when Mari’s taking care of me
    We’ll all be feeling merry when I marry Mari Mac

    (and by the time the song ends – it quickens as it goes – it’s up around 170 bpm, so that last verse is a serious test of articulation)

    • ella says:

      I’m the same – moved to Canada when I was a kid and usually sound Canadian when speaking with other Canadians, but my vowels (& some vocab) have tended to stay the same. So merry-marry-Mary are pretty much always different for me, tho the difference may be a bit subtle depending on how deep in the accent I am. I do have a pretty fully chamaeleon accent tho – most people can’t tell.

      • CaitieCat says:

        Oh, totally – if I have a friend or relative over from the UK, I’m back in my native accent in seconds. For a while, I made voice posts on my blog, and people would hardly think it was the same person when I’d a visitor. 🙂

  22. Pingback: This Week’s Language Blog Roundup | Wordnik ~ all the words

  23. dw says:

    An interesting consequence of the Mary-Merry merger is that the word “err”, which traditionally had the NURSE vowel, now often has the SQUARE vowel in North America.

    This must be by back-formation from “error”, which by the Mary-Merry merger comes to rhyme with “rarer”. I’d guess that the history must have been something like this:


    Middle English ɛr ɛrər reːr reːrər

    Fir/fern/fur ər ɛrər reːr reːrər

    Mary/Merry ər ɛrər rɛr rɛrər

    Back-formation ɛr ɛrər rɛr rɛrər

  24. AL says:

    I pronounce Mary, marry, and merry all the same. If I had to split those words into syllables, I would tend to make it mair-ee (so I would group the R with the first syllable) instead of mae/meh-ree.

    Mary, marry, merry, hairy, fairy, dairy, parry, Perry, etc. all rhyme for me.

    I never realized I made this merger until I moved to Boston and heard people pronounce America with a distinctive sound for the “e” (distinctive to me). I say Uh-mair-ih-kuh while some locals say Uh-meh-rih-kuh.

  25. says:

    I’m from Cleveland and I pronounce marry, mary and merry the same, the vowel is about 500,1800 if you are familiar with formant values.

    I can easily say them differently if I say them like mae-ry, meh-ry, and may-ry. But the sound I use before the are is with the r in the first syllable. It sounds somewhere between the /I/ in ‘it’, the /e/ in ‘say’ and the /ae/ in cat.

    When it comes to my LOT vs. START, my LOT is /lat/ at about 750,1250 and my START is an r colored at about 650,1050

    Also, I do realize that people from Cleveland have the NCVS, but I don’t know if I do. My parents are both educated, and my mom is from the South, and my Dad had foreign parents, and I live in a neighborhood with many transplants (upper income), with very conservative accents, and I spend alot of time with AAVE speakers. This could all affect my accent.

  26. benw says:

    I have an odd distribution here. I wonder if anyone can comment (and is still reading this!). I was a child in the 1970’s and early 1980’s and grew up from age 4 to 17 in Tucson, AZ and for the most part my speech is pretty typical “General American”. This means that I ought to pronounce marry-merry-Mary the same.

    But I don’t. Rather:

    1. Mary = marry = something like /me̞ri/
    2. Merry = something like /mɛri/

    That is, the two have vowels very similar to each other, but slightly different, in that the vowel of Mary and marry is slightly higher than the vowel of merry. Same length.

    It’s possible in fast speech for me, there may be a tendency for the two to merge. But the distinction feels psychologically real to me, and I’ve had certain incidents that suggest that it must indeed be real. E.g. I had a professor from Michigan who during one lecture kept talking about his friend named “Barrel”, which struck me as a very odd name. It wasn’t till several minutes later that I understood he really meant “Beryl” (!) — but it still sounded like “Barrel” (not “Beryl”) to me when he said it.

    The general rule seems to be that the “marry” sound is the default one: the only one occurring when a vowel doesn’t follow the “r” (“air”, “bear”, “where”, “werewolf”), and also whenever the vowel spelling the sound is written “a” or “ai”; also “ae” in “aero-“, maybe also “ea”. The “merry” sound is usually written “e” (“merry”, “Kerry”, “very”, “era”, etc.) but also occurs in “bury”.

    Can anyone shed light on this? Is this some unique pattern I made up, perhaps trying to reconcile the speech around me with the speech of my mother (from NYC, with a rhotic accent but who clearly distinguishes Mary vs. marry vs. merry)?

    • Phil says:

      I thought my family were the only ones! Merge mary/marry to that vowel but keep “merry” distinct.

  27. Brother Lawrence says:

    Put me on your mailing list. Tx.

  28. Rose Eneri says:

    This is one of my pet peeves. I am from Philadelphia and I pronounce these 3 words differently. Why? Because they are 3 different words with 3 different spellings and 3 different meanings. What a concept! Generally in English, a vowel followed by a single consonant is long (Mary), a vowel followed by a doubled consonant is short (marry) and an E is actually different from an A (merry).

    I now live outside of Baltimore where there is pretty much only one vowel – the one in Mary. This vowel is so ubiquitous as to make the accent largely unintelligible. I was listening to a weather forecast, and I thought a new Portmanteau had been invented (like smog = smoke + fog). The announcer said there was a chance of “flairies.” Hum, what could this new word mean? It took me a while to realize there was a chance of flurries.

    • Paul says:

      “Fern”, “fir” and “fur” are also 3 different words with 3 different spellings and 3 different meanings, but I’m willing to bet that you use the same sound in all 3 of them. I bet you pronounce “bear” and “bare” the same too.

  29. Will says:

    A lot of time has been spent explaining how they can sound different–as a Brit I want to know how they can ever be pronounced to sound the same. To me this is a classic example of what I (obviously in a very prejudiced and prescriptivist way) regard as a ‘laziness’ inherent in a number of American accents (see also the death of a distinct ‘t’ in ‘wahder’ and ‘budder’ for ‘water’ or ‘butter’).

    • Nico says:

      Is the American/Canadian /wɑɾɚ/ really any lazier than saying the English /wɔʔə/?

    • Ralph says:

      Or what about the growing number of English people who say “fink” for “think” and “ravuh” for “rather”? Or what about the English people who say “appen” instead of “happen”? You don’t think those things are lazy? Very few (if any) Americans do either of those things.

      • Will says:

        I do rationalise those things as ‘lazy’ too.

        • Nico says:

          But the thing is they really aren’t. They’re simply sound shifts in progress. It doesn’t take any more or less effort to pronounce certain phonemes that deviate from ‘standard’ pronunciation.

          Vowels and diphthongs have always been unstable in English and most of the Germanic languages in general. So many shifts have occurred throughout the language family’s history. But English is quite dynamic in this regard. While certain vowels and dipthongs are merging in one variety, they’re splitting in another. The cycle keeps going.

      • Ralph says:

        Actually I take back what I said about “ravuh”. “Ravuh” is a possibility in black American speech I think, although the “a” would be pronounced differently from how the English pronounce it. But you still wouldn’t hear “fink” or “appen” from any American, no matter what their social class, ethnicity or education level was. Those simply aren’t American pronunciations.

    • dw says:


      A lot of time has been spent explaining how they can sound different–as a Brit I want to know how they can ever be pronounced to sound the same.

      If you have a typical English accent (you identify yourself only as a Brit), a typical North American merged pronunciation would probably be like your “Mary”, but possibly with a shorter first vowel.

      To me this is a classic example of what I (obviously in a very prejudiced and prescriptivist way) regard as a ‘laziness’ inherent in a number of American accents (see also the death of a distinct ‘t’ in ‘wahder’ and ‘budder’ for ‘water’ or ‘butter’).

      Assuming you have typical features of an English/Welsh accent (again, you identify yourself only as “a Brit” so sorry if this doesn’t apply to you), then you probably pronounce at least some of the following pairs as homophones:


      All of these are distinct words; all were pronounced differently in earlier versions of English, and all are pronounced distinctly in at least some part of the world today. Indeed, all except that last are pronounced differently in nearly all of North America. Do you regard yourself as “lazy” now? 🙂

  30. Random anonymous user says:

    Mary married on December 25th, merry christmas!!!

  31. Heya i am for the firstt tme here. I found this board and
    I find It really useful & it helped me out a lot.
    I hope to give something back and help others like you aided me.