Erotic phonemes

« previous post | next post »

An unusually large number of people have suggested that I should post about the latest SMBC comic. Since I'm on the other side of the world, with slow and erratic internet, I'll just post the link, and note that (implied pornography aside) it would be good phonetics assignment to replace Zach's letter strings with IPA symbols.



  1. Joshua K. said,

    October 23, 2016 @ 9:43 pm

    Well, the one that the couple is primarily arguing about is θ. My IPA skills and/or interpretation of what sounds the comic is referring to aren't up to transcribing the rest.

  2. Maude said,

    October 23, 2016 @ 10:50 pm

    What a scream. My students are overwhelmingly French-speaking bi-and trilingual, and I tell them θ is one of the most beautiful sounds (i.e., one of my favorite sounds) in the English language. At their age, they don't need to know its real name…

  3. R. Fenwick said,

    October 23, 2016 @ 11:01 pm

    My contribution to the erotophonia: t͡ʙ̥!

    As to the interpretation of the phones in the comic: kp [k͡p] (in e.g. Kpelle, Yoruba) and pf [p͡f] ~ [p̪͡f] (in e.g. German, xiNkuna Tsonga) seem straightforward, though gp seems impossible; perhaps the voiceless p is intended to represent a change in laryngeal setting, as in, say, [ɠ͡ɓ] (in e.g. Enu-Onitsha Igbo)?. The lengthened v of bvvvv makes me suspect that here a trill is meant: [ʙ] (in e.g. Mangbetu, Ninde).

  4. GH said,

    October 24, 2016 @ 2:42 am

    Do compounds count when we're considering such sounds or sound combinations? Because in that case, doesn't e.g. a word like "flagpole" or "eggplant" have /gp/? Or is the issue whether they're considered a distinct and "atomic" phoneme in some language? (Though I would think most Germans still consider "pf" to be compounded of two sounds.)

  5. Geoffrey K. Pullum said,

    October 24, 2016 @ 4:31 am

    @ GH: The thing about the sound at the beginning of the word Kpelle is that the velar and bilabial components are produced simultaneously, not one after the other. It's a voiceless labiovelar stop, basically a [k] and a [p] uttered at the same millisecond. Not first one and then the other, as in cockpit.

    No language has or ever could have a simultaneous [g] and [p], because [g] is voiced and [p] is unvoiced. You can't have your vocal folds vibrating and not vibrating at the exact same instant. Not even if it would provoke an orgasm in your partner.

    The three sounds heard only from behind the bathroom door, by the way, are given with IPA symbols. They are velaric ingressive stops (clicks), respectively a postalveolar one, an alveolopalatal one, and a dental one. My recent phonetics teaching involved demonstrating these sounds, and they were included on dictation tests that, I now realize, may have been more arousing than they were intended to be.

    In case there are people out there who have this erotophonetic fetish in real life, I'm a reasonably accomplished practical phonetician, currently without a sexual partner. Don't hesitate to get in touch.

  6. Bob Ladd said,

    October 24, 2016 @ 9:18 am

    Without wanting to cast aspersions on my distinguished colleague Geoff Pullum's practical phonetic skills, I have to point out that instrumental research on labiovelars converges on the conclusion that the velar closure is normally initiated slightly (perhaps 20 milliseconds) earlier than the labial closure and released slightly earlier than the labial closure. The precise figure obviously depends on the overall speed of speech and lots of other factors and I only mean to indicate that the differences, though consistent, are small; moreover, Pullum is correct to emphasise the simultaneous nature of the articulation, in that the velar and labial closures typically overlap during the central phase of the consonant by as much as 100 ms or more. This temporal patterning would actually seem to make labiovelars highly appropriate for any erotophonetic fetish.

    The initiation of the velar closure is of course not audible, and early release of the velar closure inside the mouth while the lips are closed would also be auditorily almost indetectable. Some speakers may have a tiny delay between the velar and bilabial releases (I accept that certain kinds of phonetic instrumentation would be capable of verifying this), but when I pronounce labiovelar stops I release the two closures simultaneously. Which may explain why I have found them (brushing false modesty aside here) so extraordinarily successful in causing excitement in sex partners. —GKP

  7. GH said,

    October 24, 2016 @ 9:50 am

    @Geoffrey K. Pullum:

    Much obliged for the explanation. The "pf" confused me, since it's the only one (apart from "th") that I'm familiar with and can produce, and I don't think of or perceive it (at least in German) as two sounds produced simultaneously, but as a sequence; I now see that the Wikipedia article on affricates discuss this issue.

    This blog is a constant education!

  8. Andreas Johansson said,

    October 24, 2016 @ 10:11 am

    Whether θ is "rare" is actually a good question – does the raw proportion of languages count, or should you weight by the numbers of speakers?

    (And if you do weight, how do you handle the fact that some speakers of English and most of Spanish don't use the sound?)

  9. Daniel Barkalow said,

    October 24, 2016 @ 2:40 pm

    Don't mind me, I'll just be out here saying "ʭ" (and not saying "ʬ").

  10. Christian Weisgerber said,

    October 24, 2016 @ 3:06 pm


    Though I would think most Germans still consider "pf" to be compounded of two sounds.

    I suspect most naive speakers' opinion on this is colored by the spelling. Also, since German doesn't contrast affricates with stop+fricatives, it's something of a moot point. I like this part of Wikipedia's article on German phonology:

    The phonemic status of affricates is controversial. The majority view accepts /p͡f/ and /t͡s/, but not /t͡ʃ/ or the non-native /d͡ʒ/; some accept none, some accept all but /d͡ʒ/, and some accept all.

  11. Harold said,

    October 24, 2016 @ 3:24 pm

    I think it makes an interesting point about the assumed equivalence between rarity and exoticity.

  12. Counterbander said,

    October 24, 2016 @ 4:16 pm

    The cartoonist missed one: there is a genuinely rare phoneme that is routinely used in making love – the IPA [ʘ] – or what we know as a kiss. Phonetically a bilabial click, often affricated, often lip-rounded, either voiced (as in 'mwah') or voiceless. The phoneme occurs (voiced) in the Tuu and Kx'a language families of Southern Africa.

  13. AntC said,

    October 24, 2016 @ 6:58 pm

    @Counterbander I laughed out loud. thank you

  14. David Marjanović said,

    October 25, 2016 @ 1:31 am

    most Germans still consider "pf" to be compounded of two sounds

    Well. Most Germans live in places where Upper German dialects have never been spoken. Consequently, they pronounce pf at the beginnings of words simply as f, so that Pferd "horse" and fährt "moves on wheels, by ship, to hell or up into heaven" end up as homophones*. Elsewhere, they could get away with interpreting it as a consonant cluster, so that's probably what they do.

    For varieties with an Upper German sound system, including Austrian Standard German, I do prefer analyzing pf as a single phoneme. In particular, it occurs as a part of the word-initial consonant clusters pfl- and pfr-.

    As pointed out on Wikipedia, this /p͡f/ is unusual for an affricate in having two successive places of articulation instead of just one: it is a bilabial stop released into a labiodental fricative – the mouth starts out closed, and when it opens, the lower lip doesn't lose contact with the upper teeth. In contrast, the Tsonga version and its voiced counterpart are labiodental throughout.

    * Unconditional lengthening of -er- is a(nother) Low German substrate feature.

  15. John Walden said,

    October 25, 2016 @ 2:29 am

    I hope GKP clicks with somebody.

    But I wonder if they'll be allowed to comment.

  16. Christopher said,

    October 25, 2016 @ 6:24 am

    This joke is about mathematics as much as linguistics; a humorous misunderstanding about what measure is implied (language weighted vs speaker weighted).

  17. Johan P said,

    October 25, 2016 @ 6:51 am

    @Christian Weisgerber

    Ha! I never thought about the spelling thing, but I think it's right – I pronounce the "ts" in English "bets" and the "c"s in Hungarian "cucc" exactly the same, but think of them as different sounds. Nice. :)

  18. January First-of-May said,

    October 26, 2016 @ 2:31 pm

    This comic reminded me – for years I've been telling everyone that I'm pronouncing /t͡s/ (as in Russian) instead of proper English /θ/, but on a whim a few days ago I actually decided to pay attention to what my tongue is doing when I say English words with /θ/ – and realized that it does, in fact, go between the teeth, in a very different way from where it goes when I say Russian words with /t͡s/ (or, for that matter, English words with "ts", such as "bets"… and, for orthographic reasons, a bunch of words with soft "c", such as "cent", as I've mentioned in one of my previous comments).
    I just couldn't properly hear the difference between one and the other (or, at least, between my attempts to say one and the other).
    Oh, and the same for the voiced versions of both, obviously.

    I can't actually think of a minimal pair for the unvoiced sounds (at least, if we stick to English words) – it would presumably have to involve a soft "c" versus an unvoiced "th"; the voiced versions have a nice minimal pair of "then" and "zen", but I can't think of any English word other than "zen" that I even use that sound for.

  19. Jerry Friedman said,

    October 26, 2016 @ 4:15 pm

    January First-of-May: I think the first thing you need to do is start pronouncing English "soft c" as /s/ and "Zen" with a /z/.

    Is the minimal pair you're looking for something like "path" and "pats", or "myth" and "mitts"?

    John Walden: :-)

RSS feed for comments on this post