Phonemically weird words

[Travis B.]

In many languages there are words that do not fit into the typical phonemic scheme of things, usually interjections and loanwords. Then there are words which just don't fit into the language's overall phonology for no good reason at all. Have any of you encountered this, as I realized today that I pronounce vanguard as /ˈvæŋˌɡɑrd/, realized as [ˈvɛ̃ːŋˌɡɑːʁt], even though this is not normally phonologically permissible in my dialect; one would expect /ˈveɪŋˌɡɑrd/, realized as [ˈvẽ̞ːŋˌɡɑːʁt] (cf. kangaroo /ˈkeɪŋɡəˌru/, realized as [ˈkʰẽ̞ːŋɡʁ̩ːˌʁu(ː)]).

[Travis B.]

I realize in the case of vanguard this is probably due to being able to be analyzed as two separate words together, van /væn/ and guard /ɡɑrd/.

[aporaporimos]

I merge poor and pore as /pɔɹ/, but at some point I started pronouncing boor as /buɹ/ to distinguish it from bore, even though I find the sequence /ur/ pretty hard to pronounce. (I mostly pronounce pure as /pɪɹ/, homophonous with peer. The other main /ur/ word is lure, which also takes some conscious effort to pronounce correctly.)

[bradrn]

In many languages there are words that do not fit into the typical phonemic scheme of things, usually interjections and loanwords. Then there are words which just don't fit into the language's overall phonology for no good reason at all. Have any of you encountered this, as I realized today that I pronounce vanguard as /ˈvæŋˌɡɑrd/, realized as [ˈvɛ̃ːŋˌɡɑːʁt], even though this is not normally phonologically permissible in my dialect; one would expect /ˈveɪŋˌɡɑrd/, realized as [ˈvẽ̞ːŋˌɡɑːʁt] (cf. kangaroo /ˈkeɪŋɡəˌru/, realized as [ˈkʰẽ̞ːŋɡʁ̩ːˌʁu(ː)]).

What’s so phonemically odd about /ˈvæŋˌɡɑrd/? Both syllables seem fine to me.

[Travis B.]

What’s so phonemically odd about /ˈvæŋˌɡɑrd/? Both syllables seem fine to me.

The regular result of historical /æŋ/ in my dialect and in much of NAE is /eɪŋ/ (even though this may be pronounced with a slightly more open vowel than the usual /eɪ/).

[bradrn]

What’s so phonemically odd about /ˈvæŋˌɡɑrd/? Both syllables seem fine to me.

The regular result of historical /æŋ/ in my dialect and in much of NAE is /eɪŋ/ (even though this may be pronounced with a slightly more open vowel than the usual /eɪ/).

But wouldn’t that still be /æŋ/ at a phonemic level? Sure, this word is weird phonetically, but I still don’t get what’s phonemically weird about it.

[Travis B.]

What’s so phonemically odd about /ˈvæŋˌɡɑrd/? Both syllables seem fine to me.

The regular result of historical /æŋ/ in my dialect and in much of NAE is /eɪŋ/ (even though this may be pronounced with a slightly more open vowel than the usual /eɪ/).

But wouldn’t that still be /æŋ/ at a phonemic level? Sure, this word is weird phonetically, but I still don’t get what’s phonemically weird about it.

Because there normally is not a /æŋ/~/eɪŋ/ contrast, whereas this word is an example of a contrast which, for the most part, does not exist otherwise. (There is a /eɪŋ/~/ɛŋ/ contrast, but the only words which contain /ɛŋ/ in my dialect are penguin and Genghis Khan.)

[Kuchigakatai]

In many languages there are words that do not fit into the typical phonemic scheme of things, usually interjections and loanwords. Then there are words which just don't fit into the language's overall phonology for no good reason at all. Have any of you encountered this

The diminutive/pejorative suffix -uelo (< Latin -(i,e)olus) was applied on Spanish hoyo 'hole' (< Latin fovea 'pit', turned masculine), creating hoyuelo 'dimple', /oˈɟʝwelo/, normally [o.ˈʝwe.lo], which, because my dialect (that of El Salvador) has [j] as the lenited allophone of /ɟʝ/, ends up as [oˈɥelo]. It's pretty much the only instance of /ɟʝw/ (and therefore [ɥ]) in my dialect in common use. (Although note that dimples are referred to a bit more often with the word el camanance, a borrowing from Nahuatl: camac 'mouth' + nantzin 'fruit'.) Technically it's a bit irrelevant for this thread though, because I'm only mentioning it for its weird phonetic realization, not the phonemic structure.

Quebec French has a rhoticized nasal vowel in the number one: un [œ̃ɹ]. I'm not sure if this is normal for word-final /œ̃/ (would brun 'brown' have it? parfum 'perfume'?), but I think it isn't?

Some Taiwanese Mandarin speakers have /jaj˧˥/ [jaɪ˧˥] for the morpheme 崖 'cliff', otherwise normally /ja˧˥/ [ja˧˥]. It's a weird retention of a Middle Chinese triphthong that has been simplified in pretty much all other instances. Standard-like dialects of Mandarin normally only retain triphthongs when the rising element is not the same as the falling element: /waj jaw wəj jəw/ [waɪ jaʊ weɪ joʊ].

[missals]

My American English idiolect seems to have a phonological hapax, or, uh, a phonological dis. For me, historic /ʊl/, /ʌl/, and /oʊl/ all merge to something like [ol] - thus pull and pole; bull and bowl; hull and hole; gull and goal; mull and mole; cull and coal are all homophonous pairs.

Except, the vowel in wolf is not the same as the vowel in these words, even though wolf is historically /ʊl/. For some time I thought it was the only word in my idiolect with whatever vowel this is, but then I realized that wool also has the same vowel. I think these words could be analyzed as having /ʊl/ or perhaps /l̩/ - either way makes sense; the vowels sound somewhat like the vowel in book, but also seems to have significant velar construction throughout, more like [ɫ] than a normal vowel.

[bradrn]

The regular result of historical /æŋ/ in my dialect and in much of NAE is /eɪŋ/ (even though this may be pronounced with a slightly more open vowel than the usual /eɪ/).

But wouldn’t that still be /æŋ/ at a phonemic level? Sure, this word is weird phonetically, but I still don’t get what’s phonemically weird about it.

Because there normally is not a /æŋ/~/eɪŋ/ contrast, whereas this word is an example of a contrast which, for the most part, does not exist otherwise. (There is a /eɪŋ/~/ɛŋ/ contrast, but the only words which contain /ɛŋ/ in my dialect are penguin and Genghis Khan.)

Thanks for explaining — that makes much more sense now!

[hoyuelo]'s pretty much the only instance of /ɟʝw/ (and therefore [ɥ]) in my dialect in common use. … Technically it's a bit irrelevant for this thread though, because I'm only mentioning it for its weird phonetic realization, not the phonemic structure.

If it’s the only instance of /ɟʝw/ in your dialect, then wouldn’t that make it weird phonemically as well as phonetically?

[Pabappa]

i have [uj] in buoy, the only such in the language, as far as Im aware. One syllable, and not homophonous with boy.

[anteallach]

I realize in the case of vanguard this is probably due to being able to be analyzed as two separate words together, van /væn/ and guard /ɡɑrd/.

Possibly related: I have /ng/ in vanguard, not /ŋɡ/.

[sasasha]

So, and I'm rusty on all this, but isn't sound change "phonologically distinct, but lexically diffuse"*? Many of these examples just representing words which for principally sociolinguistic reasons haven't switched pronunciation yet... like outlying dominos in the chain?

Especially missal's 'wool' example.

*So rusty that I remember this idea but not its source.

[Linguoboy]

I have /kətəˈlɑn/ for Catalan. This is obviously modelled on Eastern Catalan /kə.təˈla/, but Catalan is still a naturalised English word and not a direct borrowing. I can't think of any other trisyllabic words in my speech with this stress pattern.

(In general, I'm not one to use endonyms when referring to other languages; Xhosa with [ǁ] [sic] is a rare exception just because clicks are so much fun to do.)

[Travis B.]

My American English idiolect seems to have a phonological hapax, or, uh, a phonological dis. For me, historic /ʊl/, /ʌl/, and /oʊl/ all merge to something like [ol] - thus pull and pole; bull and bowl; hull and hole; gull and goal; mull and mole; cull and coal are all homophonous pairs.

Except, the vowel in wolf is not the same as the vowel in these words, even though wolf is historically /ʊl/. For some time I thought it was the only word in my idiolect with whatever vowel this is, but then I realized that wool also has the same vowel. I think these words could be analyzed as having /ʊl/ or perhaps /l̩/ - either way makes sense; the vowels sound somewhat like the vowel in book, but also seems to have significant velar construction throughout, more like [ɫ] than a normal vowel.

My thought here is that the vowel in wolf and wool is being colored by the preceding /w/ (note that Standard English diachronically has similar w-coloring in the first place).

[Pabappa]

when i was very young and hadnt heard of the IPA, but did have paper dictionaries to look through, i came to think of /ʊ/ as the null vowel .... its hard to explain how because 99% of kids didnt go around babbling nonsense words like I said ... but I played games and would add a stressed /ʊ/ to the end of words and people's names "because it's nothing". I never did the same with the /ʌ/ vowel and I have never accepted the widely held belief that /ʌ/ and /ə/ are allophones of a single phoneme. If any strsssed vowel in English is an allophone of /ə/, for me it can only be /ʊ/, the vowel in all of full pull bull wool wolf.

Since all of these begin with labial consonants and end with /l/, I would say that they are an isolated set, and arguably distinct from the /ʊ/ vowel seen in words where an obstruent consonant follows, such as put woof brook crook roof rook root room (i dont use /ʊ/ in "root" but Ive heard it from people outside the Midwest and so I dont consider it wrong). But I dont really get into the fine details of phonetic analysis ... whether pull for me is best analyzed as /pl̩/, /pʊl/, or /pəl/ i leave up to scholars.

[Travis B.]

I realize in the case of vanguard this is probably due to being able to be analyzed as two separate words together, van /væn/ and guard /ɡɑrd/.

Possibly related: I have /ng/ in vanguard, not /ŋɡ/.

Whereas for me I have regular /nɡ/ > [ŋɡ] across the board except in the word nightingale, which has been discussed here before.

(Note that inside individual words I perceive this as a phonemic change,/nɡ/ > /ŋɡ/, and it is only apparent from affixation and word boundaries that it is a synchronic process.)

[Kuchigakatai]

I have /kətəˈlɑn/ for Catalan. This is obviously modelled on Eastern Catalan /kə.təˈla/, but Catalan is still a naturalised English word and not a direct borrowing. I can't think of any other trisyllabic words in my speech with this stress pattern.

(In general, I'm not one to use endonyms when referring to other languages; Xhosa with [ǁ] [sic] is a rare exception just because clicks are so much fun to do.)

Last year I met a black man from South Africa, who had arrived to Canada literally only four days before that day. I asked him to tell me about the situation of languages there in order to elicit a pronunciation of "Xhosa", and then I anxiously waited for him to say it. To my disappointment, he said he was a speaker of [kʰ]osa.

[aporaporimos]

Another one I remembered: in normal speech I break up the sequence /ɚl/ into two syllables, so pearl is /ˈpɚ.l/ and rhymes with squirrel /ˈskwɚ.l/. The one exception is girl, which is always a single syllable /ˈgɚl/ and as a result does not rhyme with pearl, or any other word. (Of course, when it comes to poetry and song lyrics I pronounce pearl as one syllable when appropriate, and likewise for many other words that are disyllabic in normal speech like fire, owl, boil, etc.)

Edit: fix IPA mistakes

[Pabappa]

you remind me of my childhood self again .... i used to play a game where all animal names were 1 syllable, so "džræf" etc, and abbreviated to "croc" etc if there was no convenient way to drop the unstressed vowels. squirrel was easy though, so for me it rhymed with pearl, girl, etc. and still does today even though i usually speak very carefully.

i imagine some people must rhyme quarrel with whorl ... but to me they are distinct.