Peenk?
« previous post | next post »
There's an interesting video (tiktok or YouTube) from Alisa @multilingual_nest that starts with the question "Can you guess where I'm from based on what I'm about to tell you about my accent?" Here's Alisa's audio:
As WoodyLing points out on YouTube, the discussion of Alisa's vowel in pink needs to start with the effects of final-consonant co-articulation on vowel color, and specifically the raising of the second formant involved in "velar pinch".
We can see (and hear) this clearly, in the audio examples from the online Merriam-Webster dictionary for "pink" and "pimp":
pink:
pimp:
In the MW audio for pink, F2 is 2616 Hz immediately at the start of voicing, and 2717 Hz in the middle of the vowel. In the MW audio for pimp, F2 is 2268 Hz and 2088 at the comparable points.
This suggests many questions about variation in acoustic-phonetic details across words and speakers and contexts.
- Is this difference phonetic (due to coarticulation with the final consonant), or phonological (representing a symbolic shift of vowel categories in the affected words), or both?
- Is it a near-merger as described in many other cases by sociolinguists?
- What is the distribution of the associated variation across words, space, time, and social structure?
Of course, this is different from things like the merry/marry/Mary (near-)merger, in that there's no lexical difference between /ɪ/ and /i/ before /ŋ/ to start with — that is, there's no analog to peen/pin in /ŋ/-final words. So it's not a question of whether a set of word pairs are distinguished, but rather just a question about the phonetic realization of a set of rhyming monosyllabic words ending in
/ŋ/, and preceded by a vowel whose correspondence with the rest of the phonological vowel space is unclear.
I haven't seen any systematic studies of this phenomenon. Even if there's been some coverage, I suspect it remains a good opportunity for a term project or two.
Laura Morland said,
July 2, 2026 @ 11:18 am
In the meantime, can you guess where Alisa is from based on what she told you about her accent?
Mark Liberman said,
July 2, 2026 @ 11:39 am
@Laura Morland: 'can you guess where Alisa is from based on what she told you about her accent?'
Nope. It's pretty clear where she's NOT from, at least in terms of local accent features. But where she IS from? The comments on tiktok and youtube are all over the place in response to that question.
Cedar said,
July 2, 2026 @ 11:42 am
I guess: Wisconsin / Minnesota, but not based on the "pink/peenk".
ktschwarz said,
July 2, 2026 @ 11:58 am
Phonetics blogs have been rediscovering "peenk" for almost 20 years now; the oldest one that I know of that's still on the web is Neal Whitman's Literal-Minded, from 2008, and he links to a blog from 2007 (now gone) by former Language Logger Eric Bakovic.
WoodyLing thinks it's regional to California or the West Coast, but I'm skeptical of that, since I've heard it from people from different parts of the US. Neal Whitman was surprised to be told in his first linguistics class that the vowel in "sing" was supposed to be the same as in "sin" rather than "seen", and he's from Texas.
DaveK said,
July 2, 2026 @ 12:34 pm
Im from the northeast US and for me “pink” has the same vowel as “pin” and “thin” which is different from the vowel in “ping” and “think”. The speaker sounds vaguely midwestern to my ear.
ktschwarz said,
July 2, 2026 @ 12:48 pm
Sounds to me like it's different for different words in Merriam-Webster's audio: I hear /ɪ/ as in "sin" in their think, drink, rink, vs. /i/ as in "seen" in their mink, ink. (Some male and some female speakers in both cases.) But it's hard to know when I'm just hearing what I expect to hear.
Terry K. said,
July 2, 2026 @ 1:03 pm
DaveK, are you saying that for you "pink" and "ping" have different vowels?
Jonathan Smith said,
July 2, 2026 @ 1:35 pm
Presumably related: "thank" (etc.) raising. And yeah this seems to be everywhere…
This speaker's report says in effect "this is phonological (however far or not it extends beyond me)"… or put differently, underscores again what phonemes even are — ya can't (reliably) hear them, but they're damn sure there.
Urs Gygli said,
July 2, 2026 @ 1:39 pm
Reminds me of New Zealand, where "bitter" and "better" sound – at least for my ears – quite the same. Didn't meet one native who agreed with me; but many foreigners.
Y said,
July 2, 2026 @ 3:17 pm
I hear an open DRESS vowel, as in "best", "guess", etc., which would put her in the northern US.
JPL said,
July 2, 2026 @ 5:27 pm
I interpret that effect as the raising of the blade of the tongue in anticipation of articulation with the velum, which would be in accordance with your approach to phonological description as a dynamic process, rather than a simple series of instances of phonemes.
JPL said,
July 2, 2026 @ 5:34 pm
(After listening to the video) They should not have used a word with a velar consonant following the vowel; "bit" would have sufficed. I would guess she's from Iowa.
Scott P. said,
July 3, 2026 @ 3:05 am
For me 'pink' has always been 'peenk' — until now, I would not have thought there was an alternate pronunciation. (To me, it rhymes with 'think' and 'sink').
Given that, and some of the subtleties in her accent, I would think she comes from somewhere in the Midwest — Iowa, Missouri, maybe Nebraska (but not the Old Northwest).
Michael Vnuk said,
July 3, 2026 @ 6:42 am
Mention of 'pink' reminds that I was having my mobile phone checked in a shop here in a Australia and the man doing the worked referred to a problem with the 'pink'. Although I asked him several times to repeat the word, it always came out as 'pink', and I was baffled. He appeared to be of East Asian background, and English did not seem to be his first language, but I thought I understood the rest of what he said. Even when he was pointing to the base of the phone where the charging cable goes, I still had no idea what he meant by 'pink'. I got him to write the word and it turned out to be 'pin'.
Peter Cyrus said,
July 4, 2026 @ 2:25 pm
I don't care where she's from. For me, the interesting question is why there are no minimal pairs between FLEECE and KIT before NG. I don't think it's because it's velar, because we DO have *peek* vs *pick* and even *big league*. I think it's because NG never serves as an initial in English, and maybe because of that it also can't serve as the second consonant in a final cluster, even after the semivowel offglides in FLEECE and GOOSE.
Since it never occurs after a long vowel, the short vowels after which it DOES occur are free to lengthen to a position intermediate between tense and lax, thus the ambiguity.
I'm probably grasping at straws. Don't any of you experts have the definitive answer?
Pedro said,
July 5, 2026 @ 8:10 am
@Peter Cyrus: The reason there are no minimal pairs before /ŋ/ (or to put it another way, /ŋ/ is permitted before KIT but not before FLEECE), is that historically /ŋ/ was the consonant cluster /nɡ/ (presumably surfacing as [ŋɡ], as is still heard in Northern England).
A long vowel couldn't have developed before that consonant cluster or any similar cluster such as /nd/ or /mb/, because it was before single consonants that vowels were lengthened (I'm guessing in the Old English period but someone else might know better). However, some combinations arose later due to sound changes (e.g. "find", "pained") or borrowing (e.g. "lounge").
Jarek Weckwerth said,
July 5, 2026 @ 5:40 pm
One frequent context for this is -ing been forms. It never fails to amaze me that people who have this type of pronunciation seem to be so oblivious to it. The other day, I asked a card-carrying Canadian linguist who was giving a talk about including variant pronunciations in a dictionary. S/he did not know she had it
And then there's FLEECE in -ive, this one e.g. from a top-tier sociophonetician who had never noticed it in their own pronunciation…
John Swindle said,
July 6, 2026 @ 2:44 am
According to my map in memory of what I heard in childhood (which was a long time ago, so I even if I had it right I may may forgotten or it may have changed), "peenk" is possible in the USA for the Upper Midwest but not for Kansas or Nebraska. Not sure about Iowa.
John Swindle said,
July 6, 2026 @ 2:51 am
The conversation on YouTube has evolved to say that "peenk" is used in St. Louis, Missouri … and that she's apparently from Utah. Kirtland, Far West, Nauvoo, Salt Lake City … surely "peenk" isn't that old?
Michael Watts said,
July 6, 2026 @ 6:30 am
I was surprised by a different aspect of Alicia's monologue.
I've mentioned before in the comments that I often don't see vowels before /ŋ/ as coinciding with the vowels that IPA transcriptions conventionally assign. That is the case here; to me, the vowel in "pink" cannot be either FLEECE or KIT, and I would reject the idea that "pink" could participate in a partial rhyme with "peek" or "pick".
(Touching on Peter Cyrus's point, I do think the vowel of "pink" coincides with the vowel of "pig"; final /g/ tends to allow the same 'out of standard' preceding vowels that final /ŋ/ does. This is not true of final /k/. "Peek" and "pick" are unambiguous cases of FLEECE and KIT for me. The fact that /g/ shows the same behavior as /ŋ/ leads me to doubt the explanation that the behavior of /ŋ/ is due to its past as a cluster. However, /g/ is unlike /ŋ/ in that it can also be preceded by FLEECE, as in "league", which seems sufficient to demonstrate that the vowel in "pink" cannot be FLEECE.)
What surprised me was that Alicia's pronunciation of "pink" strikes me as completely unexceptional, similar to the way I'd pronounce it myself, but she describes it as sharing the vowel of "green". I find her pronunciation of "green" unexceptional too. But to my ear they are absolutely not using the same vowel.
Eli Driscoll said,
July 6, 2026 @ 8:52 am
Hold on now. I would like to hear her pair 'grin' and 'green'. I think this is just a case of where analysing your own speech is very difficult. It's also hard to understand how someone for whom the idea that there is more than one 'i' phoneme is surprising would come to be designing a pronunciation course for French L1 speakers.
ktschwarz said,
July 6, 2026 @ 4:39 pm
Somebody on Wiktionary has recently started adding [iŋ] pronunciations to words such as pink and sing. They've put in a regional note saying "Upper Midwestern US, Western US, Canada", but without linguistics references. Wikipedia's page on California English (section Urban coastal California English) says "king is pronounced more like /kiŋ/ ('keeng')" and cites a page summarizing vowel shifts by Penelope Eckert, a sociolinguist at Stanford, but that page doesn't give any further references; it mentions the Atlas of North American English, which has a lot on /æ/ raising, but I don't know if it mentions /ɪ/ raising.
The most noticeable "keeng" pronunciations I've ever heard were on the History of English Podcast (recommended!), whose host is from North Carolina.
rpsms said,
July 7, 2026 @ 4:26 pm
I'm thinking she is from the Finger Lakes area in Central NY: Geneva, Ithaca, Cortland, Watkins Glen, Skaneateles etc.
The "something" part though has a whiff of Delaware/Baltimore but that would get more towards "some-theen" (rather than "sometheeng") And the "pink" reminds me very much of the Smokies (Knoxville, Cheeorkee national park, but if that were it, *all* the words ("six" etc. tend towards that).
David Marjanović said,
July 7, 2026 @ 6:30 pm
What little NZ English I've heard* simply uses [e] for DRESS, as in South Africa, RP in the strictest historical sense, and to a lesser extent Australia. The other Englishes use [ɛ], and their speakers are apparently not good at hearing the difference between [ɪ] and short [e] (…which are in fact allophones of a single phoneme in numerous languages, or even parts of a range of free variation).
Long [eː] is a very different matter; it is the FACE vowel of Scottish, Macro-Can[eː]dian and Indian English.
* For example, I've met a Tr[e]vor.
…which, as we learned here, Trump did rhyme: [biglig̚]. This could be part of the broader vowel raising by velar consonants in NYC, where /g/ and /ŋ/ trigger CLOTH.
Vowels were lengthened in open syllables… and before consonant clusters with a single place of articulation: /mb/, /nd/, /ld/ and, yes, /ng/ [ŋg]. Examples are mild, wild, find, found, child, kind, old, climb, grind, ground both noun and verb, wind the verb, wound its past & participle, and a bunch of others. Many others were later* shortened for, presumably, some reason; some were only shortened after their vowel qualities had adjusted to the length, e.g. long. Yet others were analogically replaced; wind the noun was backformed from windy and perhaps windmill, and the lengthened form it replaced still occurred in 18th-century poetry. I don't know what happened to wound the noun, but in Wiktionary there's an "archaic" version pronounced like the verb forms.
However, to the best of my knowledge, the only present or historical long vowels or diphthongs before /ŋ/ that do not come from this particular process are onomatopoetic: oink, boink.
* They were all still long in the Orrmulum.
HS said,
July 7, 2026 @ 11:05 pm
I'm a New Zealander though not a phonologist or phonetician. New Zealand English has undergone a systematic vowel shift in which the short front vowels have effectively moved up a step. (Older speakers may retain more RP-like vowels.) So the DRESS vowel is pronounced [e], as you say. (Foreigners apparently hear the New Zealand pronunciation of my home town of Wellington as Willington, for example.) But the KIT vowel has moved to the centre and become either somewhere between [ɪ] and [ə] or an actual [ə] (or very close). (There may be differences based on the age or education level of the speaker and between urban and rural areas, but I think for most modern speakers it is basically [ə]. In my own speech I can't decide whether it is exactly a schwa or not, but I think it's close enough to make no real difference.) So what foreigners apparently fail to hear is not the difference between [e] and [ɪ] as you say, but the difference between [e] and [ə], which surprises me. I wonder whether it is due to the special nature of [ə] as a reduced vowel (even though our KIT vowel can be stressed, just like other vowels). Perhaps when foreigners hear our KIT vowel pronounced as [ə] they instinctively interpret it as a reduced vowel, but they understand from the context that it is the KIT vowel, so they hear it as their KIT vowel, which is probably [ɪ]. But our DRESS vowel is [e] (or possibly even a bit higher?), which is close to [ɪ], so they can't hear any difference. (I don't know whether this makes sense or not.) Possibly there are other factors involved that I am not aware of (you'd have to ask a real New Zealand phonologist). But what Urs Gygli says is correct – most foreigners apparently can't hear the difference between "bitter" and "better" in New Zealand English (or between "Pin" and "Pen", to take another example which has been discussed on Language Log in the past), but to New Zealanders they are quite distinct. I don't think a New Zealander would be any more likely to confuse "bitter" and "better" than they would be to confuse "batter" and "butter".
* For example, I've met a Tr[e]vor.
This line cracked me up, but you would have to be a New Zealander of a certain age to understand why.
FJ said,
July 8, 2026 @ 8:08 am
@Jarek
Do you mean this (quoting Wikipedia):
For some GA speakers from any U.S. region whose accents are derived from, or similar to, those that originate especially in California, other Western states, and even Midwestern areas, including the Upper Midwest, the unstressed I in -ing (/ɪŋ/) is tensed (i.e., raised) and the G is dropped, so that -ing is enunciated to sound like ean (as in mean), een, or ene (as in scene; [in]),[13][14]
Daniel said,
July 8, 2026 @ 10:47 am
What I find interesting is that Alisa pronounces both "green" and "pink" as diphthongs. Both have a palatal off-glide, which makes them seem similar.
Actually, she pronounces "green" twice. The first time is more diphthongized, which I assume is because she was being more causal about pronouncing it. The second time, she begins with a more fronted vowel.
"green" [grɨjn] [grijn]
"pink" [pɪjŋk]
As other commenters have mentioned, the velarization makes the vowel sound take on a palatal off-glide.
I would like to hear how she pronounces "pin". I think we can get a flavor of it of how she tries to pronounce "pink" with the [ɪ] vowel only. The lack of palatal off-glide makes it sound as if it has a weak central off-glide, which perhaps it does.
I think she is paying attention to the off-glide as a contrastive feature, rather than the nucleus of the vowel, which is different between "green" and "pink". The "pink" vowel is more lax (or perhaps really -ATR).
Pronouncing /i/ as a diphthong was something that came up when my wife and I were thinking of names for our daughter. She liked the name "Mia", but she didn't like how I pronounced it. Her friend Mia was from Hawaii, and she pronounced her own name with a very fronted [i] sound, as [mijə], which to me sounds like "MEE-a". I pronounce /i/ more as [ɨj], using the singer's /i/ sound [ɨ], so it came out [mɨjə]. We ended up discarding this name for this reason.
David Marjanović said,
July 8, 2026 @ 1:54 pm
It does! The NZ KIT (of which I was very vaguely aware) is so far outside their experience they compare the NZ DRESS to their own KIT instead of to the NZ one.
For comparison, there are places in Brazil where rr has become [h]. The first time I encountered this, it was from a Brazilian who was speaking French this way. I quickly understood what was going on, but still had serious trouble imagining r instead of h for his [h].
I had no idea, but this is wonderful.
Michael Watts said,
July 9, 2026 @ 11:29 pm
I don't understand what was going on. To my knowledge [r] and [h] are distinct phonemes in Brazilian Portuguese, with [h] spelled "r" at the beginning of words and "rr" otherwise, and [r] spelled "r" when not at the beginning of words. (And with [r] impossible at the beginning of words?)
Do Brazilians see their [h] as a kind of R? I don't see the /ɹ/ in "colonel" as a kind of L.
Michael Watts said,
July 9, 2026 @ 11:37 pm
Or are you saying that this Brazilian was confused by French orthography, thinking for example that "terre" called for a pronunciation along the lines of /tʃehe/ by analogy to Brazilian spelling?
Pedro said,
July 10, 2026 @ 4:20 am
@Michael Watts:
Historically, R represented an alveolar trill/tap (except initially) and RR represented a uvular tell (as did initial R). Later that uvular trill developed into a [h] sound.
The uvular sound is the normal sound of French R (and RR) in any position, so the Brazilian in David M's story was applying the same sound change to his French, because he perceived these sounds to be the same.
Jerry Packard said,
July 13, 2026 @ 12:42 pm
I would guess she’s from what the Labov group would call the northern midland group – somewhere around Omaha, Nebraska/Des Moines Iowa. As to her vowel in ‘pink’/‘think’, to me it seems like @Daniel gets it right in calling it a diphthong that starts at [ɪ] and ends at [j] or [ɨ] due to the coarticulation offgliding influence of the following velar, which we can see in the formant transition on Mark’s spectrogram. The diphthong might start a little higher – closer to [i] – but it can’t be the full, unreduced long [i:] vowel, because then it would sound just like Ricky Riccardo’s (I Love Lucy) or Mel Blanc’s ‘Speedy Gonzales’ “theenk”.