English questions

[Travis B.]

Does anyone else have a long vowel in karate, as if it had /d/?. For instance I have karate [kʰəːˈʁˤɑːɾi(ː)]~[kʰʁ̩ˤːˈʁˤɑːɾi(ː)] but spotty [ˈspaɾi(ː)] (when I don't elide).

[bradrn]

It is well known that in American and British English, the short form of "mathematics" is spelled differently - "math" vs "maths". But what I wonder is, is it also pronounced differently? Because I'm not sure how to pronounce "maths" differently than "math".

Yes, of course — /mæθ/ vs /mæθs/.

Does EngE have /pæðz/ for paths? I ask because it is common to pronounce paths that way in NAE -- for instance, I pronounce it as [pʰɛːθs] in isolation and [pʰɛːðz] before a vowel, not as *[pʰɛθs].

(I should have googled it -- RP has /pɑːðz/.)

For me, I think both are acceptible — /pɑːθs/ and /pɑːðz/. (I don’t say it often enough to know which is more usual for me.)

Note also that the singular is /pɑːθ/ for me.

Does anyone else have a long vowel in karate, as if it had /d/?. For instance I have karate [kʰəːˈʁˤɑːɾi(ː)]~[kʰʁ̩ˤːˈʁˤɑːɾi(ː)] but spotty [ˈspaɾi(ː)] (when I don't elide).

I have /kəˈɹɑːti/ [kʰəˈɻʷˤɑːtˢi].

[Travis B.]

Does EngE have /pæðz/ for paths? I ask because it is common to pronounce paths that way in NAE -- for instance, I pronounce it as [pʰɛːθs] in isolation and [pʰɛːðz] before a vowel, not as *[pʰɛθs].

(I should have googled it -- RP has /pɑːðz/.)

For me, I think both are acceptible — /pɑːθs/ and /pɑːðz/. (I don’t say it often enough to know which is more usual for me.)

Note also that the singular is /pɑːθ/ for me.

/pæθs/, which would for me be [pʰɛθs], feels like a spelling pronunciation, even though apparently it is cromulent GA. For the singular I have /pæθ/, which is typical GA; for me this is [pʰɛθ].

Does anyone else have a long vowel in karate, as if it had /d/?. For instance I have karate [kʰəːˈʁˤɑːɾi(ː)]~[kʰʁ̩ˤːˈʁˤɑːɾi(ː)] but spotty [ˈspaɾi(ː)] (when I don't elide).

I have /kəˈɹɑːti/ [kʰəˈɻʷˤɑːtˢi].

Were I to pronounce a /t/ as [tʰ] I would probably say [kʰəːˈʁˤɑˌtʰe̞(ː)].

[anteallach]

I’m pretty sure that English English can have paths with a short vowel and voiced fricatives, but I can’t pluck the plural from memory. Wiktionary describes it as Northern, but non-lengthening is quite common elsewhere.

I have the short vowel in both singular and plural and pronouncing the latter with voiced fricatives feels very weird to me. It seems to me that voicing of otherwise voiceless fricatives in plurals does not occur after short vowels IMD.

[Richard W]

I’m pretty sure that English English can have paths with a short vowel and voiced fricatives, but I can’t pluck the plural from memory. Wiktionary describes it as Northern, but non-lengthening is quite common elsewhere.

I have the short vowel in both singular and plural and pronouncing the latter with voiced fricatives feels very weird to me. It seems to me that voicing of otherwise voiceless fricatives in plurals does not occur after short vowels IMD.

I too have this rule, but pronounce the word with the lengthened vowel like Bradn. I don’t know how widespread the rule is in British English. Wells reports voicing as having recently becoming much commoner in British English, and didn’t note any phonological constraints on the voicing. I think I voice in monosyllables with long vowels more often than I used to.

[Darren]

I have the short vowel in both singular and plural and pronouncing the latter with voiced fricatives feels very weird to me. It seems to me that voicing of otherwise voiceless fricatives in plurals does not occur after short vowels IMD.

Except presumably in lexicalised cases like half/halves?

[Travis B.]

In the English I am familiar with (i.e. Inland North varieties and GA, particularly GA under their influence), alternation between fortis fricatives (i.e. preceding short vowels) in the singular and lenis fricatives (i.e. preceding long vowels, and voicing when followed by another vowel) in the plural is lexicalized and is largely not productive.

[anteallach]

I have the short vowel in both singular and plural and pronouncing the latter with voiced fricatives feels very weird to me. It seems to me that voicing of otherwise voiceless fricatives in plurals does not occur after short vowels IMD.

Except presumably in lexicalised cases like half/halves?

half is an exception to the normal BATH pattern and often has the long vowel in the North of England (probably because it has a different phonological history due to /l/-vocalisation) so that's complicated. I vary between vowels in the singular, but I think I always have /hɑːvz/ for the plural; again /havz/ feels odd.

I actually know someone who has singular /paθ/ and plural /pɑːðz/.

[Richard W]

I have the short vowel in both singular and plural and pronouncing the latter with voiced fricatives feels very weird to me. It seems to me that voicing of otherwise voiceless fricatives in plurals does not occur after short vowels IMD.

Except presumably in lexicalised cases like half/halves?

half has a long vowel, so this rule allows either. The lexicon may dictate. In particular, spelling guides the literate. Having said that, I still find it hard to utter [roofs] with voiceless fricatives.

I haven’t assessed the rule for /f/; orthography guides people like me; though pre-literate acquisitions do persist. (I have been accused of speaking like a book, and beating didn’t rid me of RP.) I’ve only looked at it for the voiceless dental fricative, which I did take a long time to distinguish from /f/. I only had the voiced dental fricative word-initially, fronting it to /v/ intervocally and finally.

[Richard W]

In the English I am familiar with (i.e. Inland North varieties and GA, particularly GA under their influence), alternation between fortis fricatives (i.e. preceding short vowels) in the singular and lenis fricatives (i.e. preceding long vowels, and voicing when followed by another vowel) in the plural is lexicalized and is largely not productive.

This just confirms the deep difference between the British and American vowel systems.

[Travis B.]

In the English I am familiar with (i.e. Inland North varieties and GA, particularly GA under their influence), alternation between fortis fricatives (i.e. preceding short vowels) in the singular and lenis fricatives (i.e. preceding long vowels, and voicing when followed by another vowel) in the plural is lexicalized and is largely not productive.

This just confirms the deep difference between the British and American vowel systems.

Ignoring the non-rhotic NAE varieties (because I don't know enough about their phonologies to speak of them), the key difference between NAE and EngE varieties w.r.t. their vowel systems is that the former has lost historical phonemic vowel length while the latter not only retains it but adds onto it through non-rhoticism. Both, however, have vowel length allophony, but how this applies to their vowels in the details depends on the exact variety. (E.g. my own variety has turned it into a full-blown new systematic vowel length contrast while some varieties do not preserve a contrast before, say, unstressed intervocalic /t d/ and some varieties only have vowel length allophony for, say, historical long vowels.)

[Richard W]

Except presumably in lexicalised cases like half/halves?

A good example of the rule not working for <lf> is shelf/shelves. On the other hand, introspection says it works (for me) for health and wealth. It also works for monosyllabic ordinals like tenth.

[Richard W]

Is ‘phonemic vowel length’ a well-defined term? One definition might be that long v. short is a pervasive contrast. Another seems to be that it distinguish at least one pair. My, British, English seems to be in the latter category, with the low yield contrast of bed and baird amongst morphemes. Even then, one could argue that the latter is affected by a phonetic prohibition against long monophthongs. However, length remains a feature in British English.

[Estav]

Is ‘phonemic vowel length’ a well-defined term? One definition might be that long v. short is a pervasive contrast. Another seems to be that it distinguish at least one pair. My, British, English seems to be in the latter category, with the low yield contrast of bed and baird amongst morphemes. Even then, one could argue that the latter is affected by a phonetic prohibition against long monophthongs. However, length remains a feature in British English.

Even in situations where there is not even one case of a contrast marked purely by length, it is possible to argue for a short/long contrast based on stuff like phonotactics: e.g. in a language that has [ə ɪ ʊ] and [ɑː iː uː], we might see that a monosyllabic word can consist of the latter but not the former (which can be analyzed as a requirement that the minimal word be at least two moras long) or that the latter get changed to the former in closed syllables (which can be analyzed as a process of shortening extra-heavy syllables).

[Travis B.]

Is ‘phonemic vowel length’ a well-defined term? One definition might be that long v. short is a pervasive contrast. Another seems to be that it distinguish at least one pair. My, British, English seems to be in the latter category, with the low yield contrast of bed and baird amongst morphemes. Even then, one could argue that the latter is affected by a phonetic prohibition against long monophthongs. However, length remains a feature in British English.

By phonemic vowel length I mean that vowel length is inherent in vowel phonemes. Take, for instance, SSBE TRAP versus PALM/START. While they are not the same in quality, the former is normally short and the latter is normally long in an arbitrary fashion.

Note that I do not mean a vowel length contrast per se. For instance, my dialect has a number of minimal pairs based on vowel length, e.g. matter [ˈmɛɾʁ̩ˤ(ː)]~[ˈmɛːʁˤ] versus madder [ˈmɛːɾʁ̩ˤ(ː)]~[ˈmɛːːʁˤ], but these are best analyzed as vowel length being a property of following consonants and not vowel phonemes because A) with the exception of flapping of unstressed intervocalic /t d/, which is obligatory, lack of neutralization or elision of triggering consonants is rarely "wrong" and B) it regularly applies across word boundaries (i.e. the length of a vowel in a final syllable lacking an obstruent in its coda will be triggered by the onset (or lack thereof) of the first syllable of the following word).

(I would prefer to not posit that vowel length is phonemic before and only before unstressed intervocalic historical /t d/ and rather would prefer an analysis in which there is underlying /t/ versus /d/ even if they are only distinguished by preceding vowel length for the reason that I would want to avoid an analysis in which vowel length is phonemic only in such a constrained set of cases.)

[Travis B.]

How would one analyze intramorphemic exceptions to affrication of /dr/ in English dialects? I ask because the regular outcome of intramorphemic historical /dr/ and in some cases /d/ + /r/ across morpheme boundaries (e.g. bedroom for me personally) is to merge with a hypothetical /dʒr/ (i.e. in the dialect here as [tʃɻʁ]; note that [ɻʁ] is coarticulated, I couldn't create a tie bar for it) in many varieties of English outside conservative ideolects. However, certain words resist this even within morphemes, such as federal, where /dr/ does not undergo affrication (and in the dialect here often becomes a geminate [ʁˤː] in quick speech).

The immediate analysis one might come to is that federal has /dər/, as implied by the spelling, and the /ə/ inhibits affrication. However, the problem with this is that the realization [ɾə˞] (in GA) or [ɾʁ̩ˤːʁˤ] (in the dialect here) only surfaces in markedly careful speech (I personally find it quite unnatural myself). This implies that it is not synchronically /dər/ and realizations with /dər/ are under the influence of orthography.

However, the alternative analysis requires an intramorphemic /dr/ that inexplicably resists affrication while most cases of intramorphemic historical /dr/ have shifted to /dʒr/ outside conservative speech in a productive fashion (e.g. I find it very hard to create new words with intramorphemic [dɻʁ], with the regular outcome of borrowed or coined /dr/ being [tʃɻʁ]).

So how would you guys analyze these cases?

[Richard W]

I think it pertinent that although I normally syncopate federal, I don’t syncopate federality. This keeps affrication in the adjective optional.

[vlad]

I ask because the regular outcome of intramorphemic historical /dr/ and in some cases /d/ + /r/ across morpheme boundaries (e.g. bedroom for me personally) is to merge with a hypothetical /dʒr/

I have an affricate for /d/ before /r/, but it's a distinct affricate from /dʒ/. E.g. bedroom is different from hedgerow. (Though loungeroom is unusually pronounced as though it were loundroom.)

Are /dr/ and /dʒr/ completely identical for you?

[Travis B.]

I ask because the regular outcome of intramorphemic historical /dr/ and in some cases /d/ + /r/ across morpheme boundaries (e.g. bedroom for me personally) is to merge with a hypothetical /dʒr/

I have an affricate for /d/ before /r/, but it's a distinct affricate from /dʒ/. E.g. bedroom is different from hedgerow. (Though loungeroom is unusually pronounced as though it were loundroom.)

Are /dr/ and /dʒr/ completely identical for you?

When I affricate /dr/ it merges with /dʒr/, i.e. in the vast majority of intramorphemic cases and in some words across morpheme boundaries; I even devoice it like my /dʒ/ (which is [tʃ]). For instance, I have [tʃɻʁ] in bedroom, loungeroom, and hedgerow.

Note I suspect this is not completely true for all people here, because in conservative speech in the dialect here intramorphemic /tr/ and /tʃr/ do not merge, with the former being [tʂ(ʰ)ɻʁ] or even [t(ʰ)ɻʁ] and the latter being [tʃ(ʰ)ɻʁ]. I lacked this merger as a little kid; I pronounced my own name with what was probably [tʂʰɻʁ] or [tʰɻʁ], and only merged them at some point during elementary school, having picked up the merger from other kids at school.

[Travis B.]

I should note that /str/ has three pronunciations in the dialect here, [stʂɻʁ], [sʲtʃɻʁ], and [ɕtɕɻʁ], from most to least conservative. I have the third pronunciation these days, but when I was younger I had one of the first two pronunciations. The older someone is the more likely they are to have more conservative pronunciations of /str/.