Wikipedia:Reference desk/Archives/Language/2012 January 21

= January 21 =

/eɪ/ and /ɛ/
A while ago I flipped through two French-English dictionaries, both of which were geared to British speakers. What surprised me was that the English pronunciation guides do not use /ɛ/ and transcribe all instances of close-mid to open-mid vowels as /e/. Is this merely an anomaly or laziness on the part of the lexographers, or does British English really go without /ɛ/? I find the second option hard to believe, as my dialect relies on this difference for several important word pairs (late/let, wait/wet, abate/abet, phase/fez, and date/debt to name a few). Inter change  able | talk to me  19:35, 21 January 2012 (UTC)


 * The distinction is there, but it's a judgement call how to transcribe it. John Wells uses ⟨e⟩ for the let vowel, most other phoneticians use ⟨ɛ⟩. The pair can be transcribed ⟨ɛ e⟩, ⟨e eː⟩, ⟨ɛ eː⟩, ⟨e eɪ⟩, ⟨ɛ ɛɪ⟩, ⟨ɛ eɪ⟩, ⟨ɛ ej⟩, ⟨e ej⟩, etc. Depends on whether you think length is primary, or if it's height, diphthongization, etc. — kwami (talk) 19:41, 21 January 2012 (UTC)
 * I would prefer the sixth of those options (and I've changed the title of this question to reflect that), because the sixth is the convention that Wikipedia uses. Inter  change  able | talk to me  19:57, 21 January 2012 (UTC)
 * Note that to English speakers unfamiliar with phonetics /eɪ/ is "long A" - it is 1) "long" and 2) more associated with , and hence with /æ/, than with  and /ɛ/. The fact that there exist distinct open and close vowels /e/ and /ɛ/ will not generally be appreciated (and nor will the fact that /eɪ/ is a diphthong). Writers of foreign pronunciation guides for non-specialist English speakers often need to choose the least worst options. --ColinFine (talk) 00:53, 22 January 2012 (UTC)


 * I'm an AmE speaker, but I'm fairly certain BrE has /ɛ/. I'm also fairly certain it doesn't have monophthong /e/ (only diphthong /eɪ/), which might be why the dictionaries used /e/... it's easier to type than /ɛ/, and /e/ isn't being used anyway. (We do the same thing a lot in AmE, for instance transcribing /ɹ/ as /r/ because  is easier to type--it's on a normal keyboard--and AmE doesn't have a real /r/ anyway.) That's just speculation though. r ʨ anaɢ (talk) 01:11, 22 January 2012 (UTC)
 * Most UK accents do indeed have /ɛ/. Yorkshire (where I live) also has /eː/ ( which combination the "IPA (English)" character set offered by this editor does not include ), but I can't think of instances of short /e/ - in Bradford "take" is not /teːk/ but /tɛk/. But there's no reason why a dictionary's own indication scheme should follow IPA. --ColinFine (talk) 20:47, 22 January 2012 (UTC)
 * I'm not sure of what you're saying, because the dictionary was using the IPA. I'm sorry if I didn't make that clear. It distinguished the sounds as /eɪ/ for itself and /e/ for /ɛ/. Inter  change  able | talk to me  21:02, 22 January 2012 (UTC)
 * The person who selected the transcription conventions used in that dictionary presumably would have said that a "broad transcription" was being used instead of a "narrow transcription", without any intention to indicate the non-existence of the [ɛ] sound... AnonMoos (talk) 03:31, 23 January 2012 (UTC)
 * Just a few days ago I heard a Scottish speaker on the radio who said the names of letters j, m, and f. He clearly made the distinction [dʒeɪ] but [ef] and [em]. I think that clears things up. Inter  change  able | talk to me  16:41, 26 January 2012 (UTC)

Apostrophes and soft signs
What's the point of using apostrophes (or anything else) to indicate soft signs when transliterating from the Cyrillic alphabet? Is it simply to avoid generation loss when transliterating back into Cyrillic? Neither the soft sign article nor Scientific transliteration of Cyrillic explains why we don't ignore Ь all the time instead of just some of the time. Nyttend (talk) 20:28, 21 January 2012 (UTC)

light sounds are pretty common, so anyone even a little familiar with how russian actually sounds - regardless of whether they can read cyrillic - can benefit. 80.98.112.4 (talk) 23:26, 21 January 2012 (UTC)
 * There are minimal pairs, such as "брат" (brother) vs "брать" (to take). But generally it's part of one of the general issues of transliteration - do you aim to be faithful to the spelling or the pronunciation, do you preserve purely graphical distinctions, do you preserve phonemic distinctions that are difficult for English speakers to appreciate, etc. In principle it depends on your purposes for transliteration; but since there is value in standardisation, non-ad hoc transliteration schemes try to be general purpose, which may involve compromises. --ColinFine (talk) 01:00, 22 January 2012 (UTC)


 * Most of the time it substantially changes the pronunciation, and often it also changes the meaning. Situations where it can be ignored are mostly exceptional and indicate the laziness of the person who's doing the transliterating, the presumption that the reader will know the correct way to pronounce the word (you're supposed to know that the last 'r' in Tver is soft, but in Vladimir it's not), or that the soft sign may be inferred from the transliteration without an apostrophe (e.g. Михаил Касьянов is transliterated Mikhail Kasyanov rather than Kas'yanov, because Касьянов is a common last name, but Касянов is very rare.)--Itinerant1 (talk) 11:29, 22 January 2012 (UTC)

It is true that in practice they often ignore the ь in transliterations: e.g. Perm, Yaroslavl, Bolshoi Theatre and many others. I'd say it's a matter of some differences between "scientific" and "conventional"/"simplified" transliterations of some languages. For example, the "scientific" transliterations use diacritics to indicate long vowels (Japanese, Arabic, Hindi), tones (Chinese), retroflex consonants (Hindi) or pharyngealised consonants (Arabic), or two different types of apostrophes to indicate Arabic hamza and ʿayn - all these things might be irrelevant for common English speakers, but are quite substantial for those who speak or study those languages. --Theurgist (talk) 10:13, 23 January 2012 (UTC)