Talk:thorn

I believe these pronunciations are overly pedantic and unhelpful. They have are mostly phonemic transcriptions with one or two phonetic elements thrown in for good measure. Please choose one or the other, or do the two separately. /r/ is the same phoneme in both languages, [ɻ] is only ever used in narrow (phonetic) transcriptions, and even then rarely. Phonetic transcriptions must be written in [square brackets] and only phonetic transcriptions /between slashes/. Also "r" is not the SAMPA equivalent of the IPA "ɻ". "r" is the same in both IPA and SAMPA 138.130.33.215 09:39, 6 Mar 2004 (UTC)

Okay, I changed it to be more simple, since these are slashes and not square brackets. (I was waffling between using /r/, which is unhelpful to a non-English speaker but not overly pedantic (as befits slashes), or /ɻ/, which is indeed overly pedantic but indicates pronunciation accurately for those unfamiliar with "American r".) As for it being the same in IPA and SAMPA, SAMPA isn't meant to be an equivalency for IPA—that's what X-SAMPA is for, and English pronunciations aren't given in X-SAMPA (unlike say, Finnish ones such as under jää, as Finnish doesn't have a SAMPA alphabet). —Muke Tever 20:13, 6 Mar 2004 (UTC)

Hi Muke. My advice is stick with /r/ for English in slashes. Within slashes one symbol per phoneme, if anybody wants to add narrow transcriptions then [ɻ] is fine but every sound should have a narrow transcription (I don't think any others would be needed in this word though). We don't need to indicate "American r" because it's not a contrasting phoneme - there are no two words in any dialect which can be confused by using an "Americn r" on one hand and some other "r" on the other hand. If a foreign speaker wants to say it she will be understood everwhere whatever "r" she uses, if she wants to pronounce it the American way she will know that Americans pronounce /r/ as [ɻ] in certain conditions. The thing that has puzzled me on Wiktionary has been that we give so much attention to "r" sounds but never to say "clear l" vs "dark l" or aspirated versus non-aspirated plosives - that would help foreign learners just as much!

As for IPA and SAMPA yes they are supposed to be equivalent, X-SAMPA just does a better job because the original SAMPA was only designed with European languages in mind. It doesn't matter if there is an "official" Finnish SAMPA alphabet - you just use the official SAMPA (or X-SAMPA) symbol for each IPA symbol. Of course because SAMPA and X-SAMPA are much younger than IPA they may not have caught up and some symbols you need may not yet have been added. There are plenty of linguists who feel that even the IPA alphabet doesn't have all the symbols for the language they're studying! (By the way I see no symbols in jää which are not part of SAMPA).

Sorry for ranting, I've been woken up without enough sleep today (: Hippietrail 23:50, 6 Mar 2004 (UTC)

X-SAMPA is an ASCII transcription of the IPA. That's what it's designed for, and every symbol has its equivalent (except for the IPA's "disordered speech" sounds, and XS's two or three non-IPA symbols). SAMPA is a set of alphabets for standard phonemic descriptions of certain languages. SAMPA symbols are not necessarily compatible across languages or with X-SAMPA; when I went over this at Wikipedia I gave the example of the several different ways SAMPA alphabets treat IPA /c/; another example is Arabic's emphatic consonants are represented in Arabic SAMPA with /`/, while in X-SAMPA the notation is /_e/, and /`/ stands for retroflex or rhotic: and indeed, /`/ was borrowed from X-SAMPA for American English SAMPA for /@`/ and /3`/, as opposed to the alternative notations /@r/ and /3r/. The Finnish transcription should be marked as X-SAMPA because, since there is no Finnish SAMPA alphabet, we can't say for sure what symbols would be used in it. [Not that we wouldnt have a pretty good idea; standard Finnish doesn't do much exciting or different in the way of phonology.]

As for the specific value of /r/, I think it's possible that it might be important to indicate it (my opinion isn't as strong as "must" or "should" yet). Dark l and light l are allophones, which generally means that misuse may mark one as having an accent but won't indicate a "wrong" word. But:

1. a rather common realization of /r/ is [ɾ] (e.g., Spanish /r/), which *is* a sound in American English [1] that contrasts with /ɻ/. A minimal pair between American [ɻ] and this [ɾ] is "sparring" vs "spotting".
2. ɻ may be judged a complex symbol for a simple sound, but it is an accurate description of the sound, and it's not particularly any more complex than θ or ʒ or ɝ. It may be bulky in X-SAMPA as /r\`/, but American English SAMPA conveniently defines American r simply as /r/, so this isn't an issue.
3. Personally I'm too used to having IPA /r/ having its regular value, a trill, which just feels too far distant from American r to be used for it.

I'll stick to regular unadorned /r/ in IPA representations of English though, unless I've been remarkably convincing. —Muke Tever 06:23, 7 Mar 2004 (UTC)

Thanks for the thorough explanations! It seems I've been imagining SAMPA was more than it is. In fact I now have to agree that "plain old" SAMPA is not useful for Wiktionary since it'll differentiate overly between the "English" (which looks like RP but the SAMPA pages don't say) and "American English" and, worse, doesn't allow for Australian, New Zealand, South African, Indian, or West Indian Englishes. It seems to me we should use X-SAMPA instead on Wiktionary because it allows the 1:1 transcriptions with IPA and therefore allows greater consistency between all languages. If we do opt to keep SAMPA then we absolutely need to show both RP and US in their correct SAMPAs - and leave them out for other Englishes.

You do make a good argument with your sparring/spotting example and if I could find some US dictionaries which used the IPA in this way I'd be convinced. As it stands, all the dictionaries for all the Englishes I can find (England, Australia, Ireland) use much closer subsets of IPA than what you suggest for American English. The Macquarie dictionary in Australia only uses one "r" symbol even though we have both of the varieties that American English has although we don't have the contrasting pair because we don't have a cot/caught merger. Now Indian English does have a cot/caught merger and retroflex "r" but I don't know if it has a contrasting trill or flap and I don't know if there's a notable Indian English dictionary.

I understand how you are accustomed to /r/ representing a trill but I am firstly accustomed to it represting all the allophones (rhotic, flap or is it a tap) that Australian English has because it's been in every dictionary I've owned since childhood. And since I've become interested in linguistics and languages I've become accustomed to the fact that what one IPA symbol represents in one language doesn't always match what it represents in another - at least not in phonemic transcriptions. It's a hassle but not worth fighting unless famous print dictionaries change their pronunciation guides or we risk becoming too radical/pedantic to be useful. Hippietrail 12:32, 7 Mar 2004 (UTC)

Well, I have been trying to keep both RP and GenAm SAMPA on pages I edit; though when I don't know the RP for a word I usually leave it blank (any American word with an /æ/ in I can't even begin to guess what vowel RP has, for example). Different dialects might hafta make do with X-SAMPA, and we might do the major dialects in it also but note X-SAMPA gets kind of bulky sometimes.

As for sounds like [ɾ], phonemically this is an allophone of /d/, where (at least) American English has merged /t/ and /d/ intervocalically (except initially, initially in a stressed syllable, and before /ən/—which might be better presented /n̩/). The OED calls this sound /d/ also, which is phonemically okay but:

1. The sound is more generally [ɾ] than [d].
2. This should really be more a pronunciation guide than a phonemic breakdown. The latter is useful for native speakers, as e.g. the AHD can tell them that "water" is wôʹtər and they'll get the gist of it, but it's more helpful overall to say /ˈwɑɾɚ/.

Actually given (2) there, it might be that these pronunciations don't deserve to be in slashes. Admittedly, we don't want narrow überpedantic pronunciations (like aspirated stops, or nonphonemic length)—but I think "important" allophones like [ɾ] and [ʔ] should belong.

[ɻ], though, seems kind of borderline. Maybe we need some kind of wiktionary appendix page listing broad phones of standard Englishes languages (e.g. /r/, /p/) and their common narrow allophones ([ɻ], [pʰ])? —Muke Tever 07:56, 9 Mar 2004 (UTC)

Hi again Tever. I find all of this endlessly interesting and I've expended a lot of time and a bit of money too reading up on IPA, phonetic transcription, dictionaries' pronunciation guides, and phonetics generally. I think we need the Wiktionary users to decide what we most need out of the pronunciation section. Real dictionaries have "canonical transcriptions" but real dictionaries seldom need to describe more than two languages. Here we've had a muddle of "phonemic" and "phonetic" (to varying degrees) styles and now you're arguing for an "allophonic" transcription. If we already had an XML format and an ability for the reader to only see the secions she wanted, we could just put them all in and that would be great.

There are so many problems to solve. If we choose a symbol for "cot" and a symbol for "caught" which symbol would we use for languages with a cot-caught merger? Your example of which words have "a" as in "hat" or in "heart" is another perfect case.

Are we a descriptive dictionary? A foreign language dictionary? A foreign language learners dictionary? A pronouncing dictionary? All of the above? Different types of dictionaries have different needs from their pronunciation guides. We on the other hand seem to need to cover all bases. Hippietrail 08:24, 9 Mar 2004 (UTC)

The letter Thorn is not Latin

The letter thorn is definitely not from the latin alphabet, I would say it's borrowed from some rune alphabet. Kpbenz (talk) 18:19, 13 September 2012 (UTC)

Originally yes. But it is definitely part of the Latin alphabet now, just like J, K, U, W, Y, Z, and maybe even G, which were also not part of the original alphabet the Romans used. —CodeCat 18:25, 13 September 2012 (UTC)

Yes it definitely is part of the Latin alphabet, just presumably it wasn't at some point. Mglovesfun (talk) 18:55, 13 September 2012 (UTC)