Linguist Llama

« previous post | next post »

Linguist Llama:

This seems to have come from the meme generator meme, by way of English Major Armadillo. Anyhow there are quite a few others, including Linguist Lioness.

As the Wikipedia article explains, the schwa symbol ə is sometimes used to denote "an unstressed and toneless neutral vowel sound, … transcribed with the symbol <ə>, regardless of … actual phonetic value"; though as far as the IPA is concerned, schwa is just a "mid-central vowel, … stressed or unstressed".

[Tip of the hat to Alex Lorenc]


  1. Eric TF Bat said,

    October 26, 2011 @ 6:12 am

    I describe New Zealand as the Land of the Long Stressed Schwa. I suspect only an Australian linguist (or idle hobbyist) could explain that, but it makes sense to me…

  2. xyzzyva said,

    October 26, 2011 @ 6:21 am

    "I wanna be a schwi, it's usually high and never stressed."

  3. Theodore said,

    October 26, 2011 @ 6:39 am

    I hear stressed schwa in my wife's (Am. English) idiolect all the time, but almost exclusively for emphasis in the word "to," e.g.: "Did you learn that trick from John?" "No, I taught it to him."

  4. Jayarava said,

    October 26, 2011 @ 6:42 am

    The one L lama is a priest.
    The two L llama is a beast

  5. Ray Girvan said,

    October 26, 2011 @ 7:41 am

    Between ourselves, Clare and I always call a llama farm a "llamasery".

  6. Jon Weinberg said,

    October 26, 2011 @ 7:47 am

    @Jayarava: "The one-l lama, / He's a priest." Get it right.

  7. Rod Johnson said,

    October 26, 2011 @ 8:07 am

    And I will bet a silk paja…/there isn't any stressed schwa.

    Theodore: I would distinguish between a low or mid central vowel (which can bear stress) and a schwa (which can't). That's the way I learned it, anyway.

  8. GeorgeW said,

    October 26, 2011 @ 8:15 am

    Are there any English words in which the schwa is regularly stressed?

  9. John Walden said,

    October 26, 2011 @ 8:34 am


  10. Spell Me Jeff said,

    October 26, 2011 @ 8:46 am

    MYL: Your link is broken

  11. Spell Me Jeff said,

    October 26, 2011 @ 8:47 am

    Are there any English words in which the schwa is regularly stressed?

    As the Wikipedia article explains, some languages do this. It is very hard for me to imagine such a thing in English. I think the stress would simply turn the sound into a "true" vowel (not sure what to call that).

  12. Steve said,

    October 26, 2011 @ 8:49 am

    In my dialect (Milton Keynes, UK), the first vowels in 'birdwatcher' and 'furniture' seem to be stressed schwas. To my ear, they don't differ from my unstressed schwas (other than in length and amplitude). The formant frequencies match up on a spectrogram too, though I'm not a phonetician, so might be missing something. In any case, when I teach EFL, I transcribe 'fur' with a long schwa for my students.

  13. LDavidH said,

    October 26, 2011 @ 9:01 am

    I always thought that the schwa was unstressed by definition; if it's stressed, it's not a schwa but a regular vowel (as in purr, fur etc). So talking about a stressed schwa seems contradictory to me. But I could be wrong!

  14. Mark F. said,

    October 26, 2011 @ 9:06 am

    Translating John Walden's comment to American:


  15. Tom Vinson said,

    October 26, 2011 @ 9:26 am

    @Jayarava: A three-L lllama is a conflagration.

  16. noahpoah said,

    October 26, 2011 @ 9:40 am

    Are there any English words in which the schwa is regularly stressed?

    If I'm not mistaken, American English schwa (at least in the dialects I've heard and paid enough attention to to notice whether or not this is the case) is the result of vowel reduction, which occurs in unstressed syllables. If this is the case, then it's probably wrong to talk about a single American English schwa (in the dialects that I've blah blah blah), since it seems unlikely that, say, [i] reduces to the same unstressed vowel that, say, [o] reduces to (feel free to pick your own pair of vowels to better illustrate this point, and hey why not take consonantal context into account while you're at it).

  17. Dw said,

    October 26, 2011 @ 9:54 am

    I've heard some Americans use phonetic stresed schwa for the STRUT vowel. It always sounds odd to me.

    Many non-rhotic accents use a long stressed schwa for the NURSE vowel.

  18. Alan Williams said,

    October 26, 2011 @ 10:16 am

    Stressed schwas are pretty common in dialect from the West Midlands of England. I searched for some audio of Ozzy Osbourne speaking, but couldn't find anything clear enough. But this lady has some nice examples around 30 secs:

    Her pronunciations of the "e" in "her" and "superb" sound like stressed schwas to my ears.

  19. Grettir said,

    October 26, 2011 @ 10:20 am

    I always thought American English did have a stressed schwa. This was very evident to me when I worked in Russia; Russians have a problem pronouncing this, and tend to pronounce it as a short /a/. This is the reason for the typical pronunciations /Rashuh/ (for 'Russia') and /staff/ (for 'stuff'; I've even seen 'stuff' written as 'staff' in memos).

  20. SmR said,

    October 26, 2011 @ 10:23 am

    @noahpoah (or anyone else who's interested in this topic)- there's a paper called "Rosa's Roses" that (as you can guess from the title) discusses the differences between different reduced vowels in American English. (I only know about it because it basically saved my entire phonology seminar paper last year.) They argue that there are at least two distinct reduced vowels in American English- one mid-central ("Rosa's) and the other a bit higher ("roses"). I think they discuss consonantal context there too.

    @MYL: Thanks for the laugh. This is brilliant.

  21. John Ross said,

    October 26, 2011 @ 10:50 am

    Re: Tom Vinson's 3-l lllama meaning 'conflagration', Greg Ross (no relation) at Futility Closet says, "When Ogden Nash published this poem, a reader pointed out that a large fire in Boston is a three-alarmer. Nash responded, “Pooh.”"

  22. Andrew (not the same one) said,

    October 26, 2011 @ 10:57 am

    I take it that in 'He's Winnie ther Pooh. Don't you know what ther means?' the vowel in 'ther' (which is non-rhotic, of course) is meant to be a stressed schwa. (It's not just written 'the', I would guess, to stop people pronouncing it 'thee'.) But this is a use for emphasis, like that of Theodore's wife, not the ordinary pronunciation of the word.

  23. Boris said,

    October 26, 2011 @ 11:08 am

    I'm not sure about the Russia thing. I'm from Russia and I've never heard that pronunciation (unless I'm misunderstanding). If anything, it's pronounced "Rah-shəh" (since there is no differentiation in Russian between a short "o", a short "u" and the "ah" vowel). I could believe spelling stuff as staff, but only if it's produced by a person who already pronounces it "stahff" and confuse the spelling because the Russian a makes an "ah" sound. (obviously I'm talking about the American pronunciations of the short a here).

    The only counterexamples to all this are the borrowing of the word lunch which gets spelled and pronounced in Russion as "lanch" and "Huckleberry Finn" somehow becoming "Geckleberry Finn"

    Anyway, in what way is the "u" in lunch a schwa?

  24. Jon Lennox said,

    October 26, 2011 @ 11:18 am

    Eric TF Bat: That would be the stereotypical New Zealand "Fush and Chups", right?

  25. Grettir said,

    October 26, 2011 @ 11:21 am


    "If anything, it's pronounced 'Rah-shəh'" – this is exactly what I tried to convey with my example (Rah-shuh). It's the first syllable I want emphasize as relevant to this topic.

  26. Rod Johnson said,

    October 26, 2011 @ 11:25 am

    From a phonological perspective I think it's worth distinguishing between [ʌ] (the vowel in strut) and reduced vowels like [ə]. What people here are calling "schwa" is actually highly variable in quality, sometimes fairly high (the [ɨ] in roses), sometimes slightly rounded (the [ʊ] in await), sometimes fairly low (the prototypical [ə] in about), sometimes r-colored (the [ɚ] in butter in rhotic dialects). Sometimes what people report as "schwa" is actually not a vowel at all, but a syllabic consonant like the [n̩] in button. There's a ton of low-level phonology going on that is very difficult to hear reliably. And if your perception of the vowels differs from mine, that's the point—they're variable, and it's hard to perceive the exact vowel quality. They're so short that the effect of the surrounding segments tends to overwhelm any "inherent" vowel quality they might have. What they have in common is that they're unstressed.

    This is why I say there's no such thing as a stressed schwa. "Schwa" properly should refer to these unstressed, reduced vowels. It's not really a vowel quality. IPA [ə] shouldn't be thought of as a schwa—it's just a mid-central vowel. Or, better, the fully stressed, unreduced vowel with the mid central unrounded quality people are referring to as "schwa" should really be transcribed as [ʌ] to avoid confusion. (I think IPA gets this wrong, by the way—or I should say that its strictly phonetic basis often makes it difficult to use for phonological purposes.)

  27. dw said,

    October 26, 2011 @ 11:57 am

    @Rod Johnson:

    You seem to be arguing that the word "schwa" and the symbol ə ought to be reserved for unstressed vowels. Fair enough: it's an arguable position, although many, including the IPA, disagree.

    This doesn't answer the interesting question of whether a stressed mid central vowel occurs phonetically in English. It does: many sources describe the RP NURSE vowel as an unrounded long mid-central vowel. I also sometimes hear the US STRUT vowel as a phonetic schwa, although it's usually further back.

  28. Rolig said,

    October 26, 2011 @ 12:02 pm

    In Slovene, the stressed short "e", as in the words pès ("dog") and tèmen, is usually described as a schwa, and the standard orthographic reference book conveys the sound of this vowel with the schwa symbol (ə). As an American who lives in Slovenia, I would describe this sound as very close to the "u" in pus or bud, only shorter. In the Ljubljana dialect, by the way, it seems that the tendency is to pronounce all short non-rounded vowels with this sound; at least I can't distinguish any difference between the vowels in pes, zrak ("air"), and nič ("nothing, never mind, forget it").

  29. Michael Sappir said,

    October 26, 2011 @ 12:04 pm

    This post suffers from a serious lack of linkage to lolPhonology:

  30. Mitch P said,

    October 26, 2011 @ 12:54 pm

    Stressed schwas? Oh, Puh-lease!

  31. Grettir said,

    October 26, 2011 @ 1:01 pm

    @Rod Johnson: thanks, I was unaware of these distinctions.

  32. Kapitano said,

    October 26, 2011 @ 2:20 pm

    What do you call the stressed vowel in 'Murder'?

  33. Rod Johnson said,

    October 26, 2011 @ 2:37 pm

    @dw: right—I'm arguing that the word "schwa" as a technical term should be reserved for "reduced" vowels and is therefore always unstressed. I'm agnostic on the symbol, though it is used pretty promiscuously and in ways that don't match its IPA value, which lessens its usefulness.

    As for whether mid central vowels occur phonetically in English: I absolutely agree they do. If you look at a sample of vowels on a formant chart, very few of them approach the IPA cardinal vowels very closely. We have a ton of more or less centralized vowels.

    I should be deferring to myl here—he has a lot more expertise with the down'n'dirty acoustic data.

  34. dw said,

    October 26, 2011 @ 3:08 pm


    I was referring to the vowel of "murder" when mentioning the NURSE vowel above. I was using the convention of the lexical sets introduced by John Wells.

    Many non-rhotic accents spoken in England and elsewhere have a long central mid-vowel (i.e. a "schwa" according to the IPA definition) for this sound.

  35. Max said,

    October 26, 2011 @ 3:44 pm

    This makes me think of when people from my native San Francisco particularly like a lyric in a hip hop song or just express general agreement with a hearty "Uh!"

    Less esoterically, what about the (only Am. English?) polar question response "UH-uh"

    Are these not instances of stressed schwas in English?

    [(myl) "Uh-uh" is usually transcribed with the same vowel as "dud" or "but", which is often given (quasi-phonemically) as [ʌ].

    As I explained in the OP, there's an ambiguity of interpretation here — schwa (the symbol or the term) can mean either (1) "a neutralized unstressed vowel phoneme (or syllable nucleus) that takes on a wide variety of sound values, depending on context and the phase of the moon"; or (2) "a particular place in the dead center of the IPA vowel space".

    So the answer to the question "isn't X a stressed schwa?" is "what do you mean by 'schwa'?"

    If your answer to the "what is schwa?" question is (1), then the response to your "is X a stressed schwa?" question is "By definition, no." If your answer is (2), then the response is "give me a recording, along with recordings of a sample of other vowels from the same speaker, and we'll see."

    If what you mean to ask is "do some possible pronunciations of this stressed vowel fall in the same region of the vowel space as some possible pronunciations of schwa-in-sense-(1)", then the answer is "probably" — because the range of possible pronunciations of schwa-in-sense-(1) is quite broad, basically covering the central 30-40% of the whole vowel space…]

  36. Ray Girvan said,

    October 26, 2011 @ 4:17 pm

    @Andrew (not the same one): "'He's Winnie ther Pooh … 'ther' (which is non-rhotic, of course)"

    I'm not sure that's an "of course". As a rhotic speaker, it took me decades to realise this (I probably read it on LL). Nor did I realise that the hesitation sound "erm" in print represented "əm".

  37. Jon Weinberg said,

    October 26, 2011 @ 6:48 pm

    I'm with Ray. I learned the non-rhotic pronunciation of Winnie-ther-Pooh in my late 40s, from . Not only had I not understood it when I read the book as a boy, but the non-rhotic pronunciation hadn't occurred to me thirty years later when I read the book to my child.

  38. Rod Johnson said,

    October 26, 2011 @ 7:29 pm

    Me too. In fact, the British use of -r to denote long vowels had always been sort of a puzzle. I remember being baffled reading Lord of the Flies at one character yelling "Sucks to your ass-mar!", which was almost completely opaque to teenage me.

    That convention got picked up by Burmese, who use it when transliterating in the Roman alphabet to indicate "heavy" tone, which is part of the reason the word which is pronounced [mjə.ma:] or [bə.ma:] (i.e., "Burma" in colonial-era Brit-spelling) got respelled as "Myanmar," which people now pronounce as []. It's a weird combination of straight transliteration and British spelling conventions.

  39. Joshua said,

    October 27, 2011 @ 12:19 am

    And similarly, when the singer Sade was launching her U.S. career after first breaking through in the U.K., her record company explained the pronunciation of her name as "SHAR-DAY". Unfortunately, they forgot to take into account the fact that most Americans would use a rhotic pronunciation, and some (including myself) referred to her as [ˈʃaɹdei] for years. "SHAH-DAY" would have worked much better for an American audience.

  40. Xmun said,

    October 27, 2011 @ 1:40 am

    Alas, here in NZ people (even linguists) sometimes give the written words "fush and chups" a spelling pronunciation, making the words rhyme with "hush" and "pups". But that's a misreading. The spelling is meant to represent something like the NURSE vowel, but short.

  41. LDavidH said,

    October 27, 2011 @ 3:02 am

    I find it most fascinating that Brits would be unaware of the rhotic American accent (as in the Sade example), and Americans equally unaware that BrE is generally non-rhotoc (which is why Winnie-ther-Pooh is "of course" non-rhotic; he was English, after all!). As an ESL speaker, having worked with Americans and married a Brit, I take it in my stride that pour and paw rhyme in the UK but not in the STates!

  42. LDavidH said,

    October 27, 2011 @ 3:21 am

    Sorry – I meant to say that pour and paw sound more or less the same in the UK. They rhyme with lore, law, saw, sore etc.

    @myl: thanks for explaining that schwa has two different meanings. As usual, English turns out to be more complicated that anyone thought… although I still favour the first definition, it seems more "useful" somehow.

  43. Theophylact said,

    October 27, 2011 @ 8:03 am

    In a footnote, Ogden Nash said his attention had been called to a type of conflagration known as a three-alarmer. His response was "Pooh".

  44. Rodger C said,

    October 27, 2011 @ 8:06 am

    @Boris: That odd mistake with the vowel in "Huckleberry" occurs in the Soviet Yiddish version too. Did the translator, I wonder, confuse the meaningless "huckle-" with "heckle," a word any good Soviet learner of English would presumably have known?

  45. Brett said,

    October 27, 2011 @ 8:19 am

    @LDavidH: I can't speak for all Americans, but I certainly was aware that British prestige dialects (among others) were non-rhotic. However, I (like several other commenters above) never realized that a Brit would spell "uh" as "er" (until I saw it explained on Language Log). The possibility simply never occurred to me. I suspect this mistake is pretty common, and I'm sure it illustrates something interesting about the relationship between dialect recognition and written language, but I have no idea what.

  46. Ellen K. said,

    October 27, 2011 @ 9:16 am

    Two things I don't follow.

    Rod Johnson mentioned "Sucks to your ass-mar!" as something that was opaque because of the R not representing an R sound. It's still opaque to me, even knowing that.

    Xmun wrote: Alas, here in NZ people (even linguists) sometimes give the written words "fush and chups" a spelling pronunciation, making the words rhyme with "hush" and "pups". But that's a misreading. The spelling is meant to represent something like the NURSE vowel, but short.

    I don't understand what the difference is between the non-rhotic NURSE vowel and the vowel in "hush" and "chups".

    Also, along the lines of what Brett said just above, many of us, I think, are aware of non-rhotic dialects, but not aware that they use silent R's when sound-spelling.

  47. Rod Johnson said,

    October 27, 2011 @ 9:26 am

    Ellen, "ass-mar" = "asthma." One of the boys has asthma, and another—the villain—is basically saying "who cares, ya big baby?" It was also the "sucks to" idiom that was an obstacle to me, even after I realized what "ass-mar" was.

  48. Andrew (not the same one) said,

    October 27, 2011 @ 10:39 am

    Also, along the lines of what Brett said just above, many of us, I think, are aware of non-rhotic dialects, but not aware that they use silent R's when sound-spelling.

    Well, why wouldn't we? If there is a sound which in our dialect is regularly represented by 'ar', naturally we would use 'ar' to signify it when sound-spelling. It's true that in this case we could equally well write 'ah'. In the case of 'ur', though, I can't think of any way of representing it without the R – I'll leave it to others to explain exactly what the difference is, but 'nurse' certainly doesn't sound like 'nuss'.

  49. Xmun said,

    October 27, 2011 @ 12:26 pm

    @Ellen K: I don't understand what the difference is between the non-rhotic NURSE vowel and the vowel in "hush" and "chups".

    I meant that the vowel in "hush" and "pups" is the STRUT vowel (see Dw's comment above, October 26, 2011 @ 9:54 am), while the non-rhotic NURSE vowel is the vowel of (e.g.) "fur", "her", "purr", "cur", "sir", "whirr". I've even heard "duh" sounding pretty similar.

    (Though I live in NZ, my accent is that of Home Counties BrE.)

  50. Ellen K. said,

    October 27, 2011 @ 12:49 pm


    Yes, I understood that. I don't understand the difference.

    I conceive of all those words, when pronounced non-rhotically, as having the STRUT vowel. "Duh" too. Perhaps IPA would be helpful, if someone can describe what you're saying with IPA symbols.

  51. dw said,

    October 27, 2011 @ 1:08 pm

    @Ellen K.

    In many dialects (including my own which is like RP in this respect) there is a contrast between schwa and the unstressed STRUT vowel. A minimal pair would be "an ending" vs. "unending". For me, both initial syllables are unstressed, but they differ in vowel quality, with that of "AN ending" being more central and that of "UNending" being more back and open. I'm going to assume that this is also the case for Australian and New Zealand English.

    In New Zealand English, in fact, schwa is merged not with STRUT but with KIT. Hence the stressed vowel of "fish" or "chips" is qualititively similar or identical to the unstressed initial vowel of "about". Australians are aware of this fact, and make fun of it (I would guess that it's one of the most salient distinctions between Australian and New Zealand accents). However, they make fun of this by saying that New Zealanders say "fush and chups", because this sounds close (though probably not identical) to what in an Australian accent would be "fush and chups".

    However, to a New Zealander, "fish and chips" would be quite distinct from "fush and chups". When a New Zealander sees "fush and chups", (s)he is seeing an imperfect Australian orthographic impression of a New Zealand accent. It's somewhat similar to non-rhotic Brit saying that "Sade" is pronounced "Sarday", and then Americans pronouncing it rhotically.

    Phew. Hope that made sense.

  52. Rod Johnson said,

    October 27, 2011 @ 1:21 pm

    In some (non-rhotic) UK English dialects, the vowel of nurse is a lowish central vowel ([nɐs] or [nʌs])–this sounds very posh to me. In others it's higher ([nəs] or [nɛs] or even [nɨs], with significant velarization). I don't think you can point to one pronunciation–I think it's all over the vowel space.

    Is there an archive of speech samples for this sort of thing online anywhere?

  53. Xmun said,

    October 27, 2011 @ 1:44 pm

    @Ellen K: dw's latest comment explains exactly what I meant. I think the relevant symbol is /ə/ but maybe the sound I’m thinking of is closer to one of its near neighbours in the mid-central area of the IPA vowel chart I've been looking at (see

  54. kktkkr said,

    October 28, 2011 @ 1:35 am

    Now I get it! Linguist Llama is being mentioned on Language Log because llama starts with LL, and Linguist Llama and Language Log are both LL! The title of the page in my browser has all the 4 words starting with L!


  55. J Lee said,

    October 28, 2011 @ 3:27 am

    i have heard a consistent stressed schwa in words like 'tourist' in hawaiian creole english, which lacks vowel reduction in words like 'tomorrow.'

  56. Ellen K. said,

    October 28, 2011 @ 8:58 am

    @XMUN. But DW did not at all mention the NURSE vowel, which is what I was commenting on. Does NURSE vowel (in your description) = schwa (in the "mid-central vowel, … stressed or unstressed" sense)? If so, I get the distinction. Not that I can hear the sounds in my head, but I can picture them on a vowel chart, at least.

    Simpler to understand is that Australians sound=spelling how New Zealanders say something looks inaccurate to New Zealanders.

  57. dw said,

    October 28, 2011 @ 11:17 am

    @Ellen K:

    In RP and similar accents, the NURSE vowel is usually a long mid central vowel [əː]

  58. Occupy Google Reader: My God, Google, Why Have You Forsaken Us? | Business News – Tech News – Entertainment – Mobile – Social Media said,

    October 28, 2011 @ 2:47 pm

    […] – I’ll see a post about some crazy architectural feat in the Middle East, followed by a ridiculous gif or meme, followed by a physics abstract, followed by an old interview between John Ashbery and Kenneth […]

  59. Dylan said,

    November 3, 2011 @ 9:49 am

    Catalan has dialects with stressed IPA [ə] (in the Balearic Islands). I've examined both stressed and unstressed realizations of [ə] in Praat (for my dissertation), and they were statistically indistinguishable from one another. So stressed IPA [ə] most certainly exists.

    I recently looked at recordings for Comanche made in the 70s, 90s, and mid 00s. Six speakers all produced an accented vowel that occupied something close to dead-center of the IPA chart – maybe just a bit higher than that (the ever-elusive turned ‘e’ / ɘ / perhaps?).

    If you call IPA [ə] 'schwa', then stressed schwa exists. If you reserve the term 'schwa' for unstressed / reduced vowels, then, by definition, schwa is never stressed.

RSS feed for comments on this post