Help talk:IPA/Norwegian

Problems with basing IPA on the Oslo dialect.

I see way too many problems with having the Oslo dialect as a representative of Norwegian language as a whole. I simply suggest the retroflexes to be written out with "rd", "rs" etc. And when I see that "nynorsk" is supposed to pronounced as [nʏːnɔʁsk], it just creates confusion for the reader. Nynorsk was created with the principle that Hochdeutsch had. But the IPA for Germany-related articles aren't based on the Berlin dialect. I suggest Sunnmøre to be the "basis" for the pronunciation (without the retroflexes and palatalization of kk and gg, though). –– ♫ Mara/Freya ♫ 21:04, 19 June 2015 (UTC)

Trying to define one pronounciation for Norwegian is a mine field. The country hasn't managed to settle on one standard way of writing the language, with some still holding on to unoffical standards as well. I think the reason why this article uses whatever it uses is because that's what books on the subject use. (Particularly Das Aussprachewörterbuch mention in the bibliography?) This is at least the impression I got back when Norwegian and Swedish were presented on the same page (although often which different columns for each language). Ters (talk) 07:09, 20 June 2015 (UTC)

Consider this quote:

One can therefore say that Bokmål has a spoken realization that one might call an unofficial standard spoken Norwegian. It is in fact often referred to as Standard Østnorsk (‘Standard East Norwegian’). In addition to being spoken by a fair proportion of the members of the Norwegian speech society, a fairly conservative version of it is the most commonly used stage language in Norwegian theatres, including those outside Oslo. It is also a variety often used in news readings on radio and TV, and it is the variety most commonly taught to foreigners.

— Gjert Kristoffersen, The Phonology of Norwegian, page 7

Peter238 (talk) 16:23, 22 June 2015 (UTC)

I think think we should use broad transcription where the dialects differs. E.g. we transcribe norsk as /nɔrsk/, because: the /n/, /ɔ/ and /k/ are found in all (?) normal pronunciations of norsk (cf. consensus sequence in molecular biology), whereas /rs/ is broad, and the reader would find on this page that it could be pronounced as e.g. (?) /ʂ/ or /ʁs/. --Njardarlogar (talk) 12:37, 22 March 2016 (UTC)

Watch out: [ʂ, ʁs] are phonetic realizations of the sequence /rs/, and so they require square brackets (read phoneme and allophone). Mr KEBAB (talk) 09:47, 16 September 2016 (UTC)

[v] v [ʋ]

So I've noticed that about half the transcriptions linking to this guide use [v] and about half [ʋ]. Norwegian phonology lists it as [ʋ], and I suspect the choice of [v] is a vestige of a search for harmony with Swedish from back when they shared a common guide. I don't have a particular feeling about which one we should use here, but if we stick with [v] someone should slog through the transcriptions and remove the [ʋ]s to minimize reader confusion. — ˈzɪzɨvə (talk) 21:34, 15 September 2016 (UTC)

I vote for ⟨ʋ⟩ - that's what we use on Norwegian phonology. Mr KEBAB (talk) 09:46, 16 September 2016 (UTC)

Actually, I think this guide needs to have two sections: one with a set of (rather broad) phonetic symbols, as on Help:IPA for Alemannic German (for vowels, I suggest the simplest choice [ɑ ɑː æ æː e eː i iː ɔ ɔː ø øː u uː ʉ ʉː y yː] and non-syllabic [i̯ ʉ̯ y̑] for diphthongs), and a section below it dedicated specifically to the Standard Oslo accent which, as it was pointed out above, is only an unofficial standard of pronunciation. In my opinion, we should always provide Oslo pronunciation alongside the regional one. It's not an official standard, yes, but it's a very useful point of reference, and, for learners of Norwegian as a foreign language, Standard Eastern Norwegian is the de facto standard that they're learning.

If we're using broad transcription, then we should use [v] for the voiced labiodental continuant, and limit our transcription of rhotics to [r] (representing any coronal rhotic) and [ʁ] (representing any dorsal rhotic). Of course, retroflex assimilations still need to be transcribed.

If there will be no objections, I will start the process soon (and, of course, will do the same on Norwegian phonology once I'm done here). Mr KEBAB (talk) 10:39, 26 September 2016 (UTC)

Tone units

We seem to be lacking means to transcribe tone units. Take Møre og Romsdal - the transcription says [ˈmøːrə ɔ ˈrumsdɑːl], yet the recording is more like (narrow transcription) [ˈmǿːɾ̩ɔˌɾumsdɑ̀ːl] (with /rə/ turned into a syllabic tap), said practically as one word (Møråromsdal).

Transcribing the tone units as [²møːrəɔˌrumsdɑːl] is probably a bad idea, since there will always be someone who will feel the need to 'correct' the transcription to [²møːrə ɔ ˌrumsdɑːl] (which in this case doesn't look very bad, but there likely are tone units that are longer than three words, in case of which such a transcription would be ambiguous).

I propose using [‿], which we use to mark French liaison, so that e.g. Møre og Romsdal would be transcribed [²møːrə‿ɔ‿ˌrumsdɑːl]. Mr KEBAB (talk) 18:56, 6 October 2016 (UTC)

Is it really a good idea to transcribe colloquial contractions? Whether Møre og Romsdal is pronounced as one word or three seems to me somewhat like "fish and chips" versus "fish 'n chips". If the name of that dish was written using IPA, which I know nothing of, would it be according to the former or the latter? Ters (talk) 19:40, 6 October 2016 (UTC)

Move discussion in progress

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:17, 15 July 2017 (UTC)

We're trying too hard to be politically correct

I've just replaced ⟨i, y, u, e, ø, ɔː, ʉi, ei, øy, ɔy, æi, æʉ, ɑi⟩ with ⟨ɪ, ʏ, ʊ, ɛ, œ, oː, ʉɪ, ɛɪ, œʏ, ɔʏ, æɪ, æʉ, ɑɪ⟩. The latter set is more helpful to our readers because it allows them to match Norwegian vowels with the Swedish and German ones. AFAICS, there's no reason to use a more broad transcription here. ⟨ɛɪ, œʏ⟩ look close to Dutch /ɛi, œy/ and our readers can safely pronounce them the same. The same applies to Norwegian [æɪ], which is a lot like /æɪ/ in Australian English (though the latter has various allophones). As you can see, I've also changed the second elements of most of the diphthongs from ⟨i, y⟩ to ⟨ɪ, ʏ⟩ so that they match the way we transcribe the short /ɪ, ʏ/.

Of course, by Norwegian I mean Urban East Norwegian (UEN) because that's the accent we use as a model, in accordance with Norwegian pronunciation dictionaries and books teaching Norwegian pronunciations. Because of that, there should be a clear distinction between UEN and other dialects on Wikipedia, and every dialectal pronunciation should be preceded by a UEN one (that is, should we even include pronunciations from dialects other than UEN - Wikipedia isn't a pronunciation dictionary of Norwegian). Because when that's not the case (and as of now, it's not), it renders WP as a borderline useless source of Norwegian pronunciations. No native mixes dialects in a way the totality of our transcriptions suggests, they stick to their own dialect in speech. What we're doing now is a violation of WP:NPOV to me. Our readers are more likely than not to expect our transcriptions to reflect the way UEN speakers pronounce any given word.

I propose that:

• We mandatorily transcribe Norwegian words in either Urban East Norwegian and the relevant local dialect or just UEN but never just a non-UEN dialect. We should clearly label each transcription.
• We rename this guide to Help:IPA/Urban East Norwegian and use IPA-no only for UEN pronunciations. Should we do so, we should remove [ʁ] etc. from the guide.
• We use IPA-all for pronunciations from other dialects (or, alternatively create Help:IPA/Norwegian dialects akin to Help:IPA/Alemannic German and create an IPA-no-dial template or a similar one)

We should also start using the superscript 1 for syllables with the first toneme. The primary stress mark shouldn't be used at all because it's ambiguous.

Before you object, please make sure that you properly differentiate between the way dialects are treated in Norway and the way Wikipedia and language teaching works. They're different things. Kbb2 (ex. Mr KEBAB) (talk) 13:41, 24 October 2018 (UTC)

Using rolling r instead of uvular r for southwestern Norwegian cities and town is just misleading and nonsensical 178.232.196.246 (talk) 02:05, 8 January 2024 (UTC)

Pinging @Aeusoes1, Nardog, and Xyzzyva: as well as @McLennonSon and Ters:. Kbb2 (ex. Mr KEBAB) (talk) 14:07, 24 October 2018 (UTC)

Since there are no objections I will start the process soon. There's an additional argument for insisting on transcribing words as they are pronounced by speakers of UEN: because it's an unofficial standard of pronunciation of Bokmål. It corresponds rather closely to the way the majority of Norwegians write their language.

I propose two additional things: that we move the majority of Norwegian phonology to Urban East Norwegian and make the former about all Norwegian dialects while restricting the latter just to UEN (which, of course, is very surprising given the title of that article :P)

The other thing is transcribing tonemes with their proper IPA signs. Because we won't be trying to accomodate any other dialect but UEN (and because there's no other dialect that's as close to the written language of the majority of Norwegians), IMO we should write [ˈbœ̀nːər, ˈbœ̂nːər] instead of [ˈbœnːər, ²bœnːər]. This will also eliminate the seeming special treatment of words with the second toneme, which some scholars (I think Kristoffersen is one of them, but I could be mistaken) consider to be the unmarked one. (EDIT: I meant Wetterlin (2010). She says that Kristoffersen is among scholars who consider accent 2 to be marked.)

UEN has a huge advantage over Swedish in that the phonetic realization of tonemes seems to be much more consistent and straightforward than it is the case in Central Standard Swedish. I still think that transcribing words with the first toneme unmarked is a mistake. This means that the reader can never be sure whether a polysyllabic word that isn't stressed on the final syllable and which is transcribed with a simple stress mark has accent 1, whether the transcriber didn't care about transcribing tonemes (or thought that they were unimportant, when in reality this is a very important information for learners of Norwegian who want to sound native), didn't know they exist or whether the word really has accent 1.

I think we can deal with local pronunciations this way: Urban East Norwegian: [XXX], local: [YYY]. The former would use the IPA-no template, the latter the IPA-all one. I think in order not to spam WP articles with multiple transcriptions of Norwegian words they should be moved into a footnote. Kbb2 (ex. Mr KEBAB) (talk) 13:32, 1 February 2019 (UTC)

I object. For any towns/cities/regions of interest, the most objectively correct pronunciation would be the local one, and not the eastern Norwegian pronunciation.

Eastern Norwegian is hardly any more accurate than how a british person would pronounce it 178.232.196.246 (talk) 02:09, 8 January 2024 (UTC)

Also, if there are no objections, I'm going to restore the old transcription of vowels (that is, ⟨i, y, ʉ, u, e, ø, ɔ, æ, ɑ⟩ for the short vowels and ⟨iː, yː, ʉː, uː, eː, øː, ɔː, æː, ɑː⟩ for the long ones). The research that I've done shows that length is the only important factor in distinguishing between the two sets. Any other phonetic differences between them are secondary if not tertiary and natives tend not to hear them very well. Sources do vary on how exactly these two sets differ, but the agreement seems to be that length is the most important. Also, on Kristoffersen's vowel chart there's a difference between the short [æ, ɑ] and the long [æː, ɑː], yet he writes these pairs with the same symbol (as all other sources do, AFAIK). This is slightly inconsistent.

Sources in general are rather inconsistent in the way they use those lax symbols in Norwegian, and the biggest inconsistency is in the case of the short close central vowel. No source writes it with ⟨ʉ̞⟩ (Kristoffersen suggests that this transcription is good, yet he doesn't use it) and two sources use the non-IPA symbol ⟨ᵿ⟩. I'm aware of the shortcomings of the official IPA in this case, but still...

Diphthongs should be written with ⟨æj, æw, øj⟩ etc. per Kristoffersen. There are multiple good reasons for it (check the source). Kbb2 (ex. Mr KEBAB) (talk) 09:13, 6 February 2019 (UTC)

I think that it says a lot that the source which uses the lax symbols most consistently (Kvifte & Gude-Husken (2005)) describes most short-long pairs as very similar to each other, but when it comes to the /ɔ–ɔː/ pair (in which the short vowel is supposedly much more open, the /ø–øː/ pair is similar in that regard) they use the same symbol. I think it's of no benefit to anyone to write the short vowels with different symbols. Kbb2 (ex. Mr KEBAB) (talk) 10:11, 6 February 2019 (UTC)

Symbols for the 'retroflexes'

I think that using ⟨ɖ, ɭ, ɳ, ɽ, ʂ, ʈ⟩ for these sounds isn't very appropriate, for multiple reasons:

• These sounds are never properly retroflex. [ɖ, ɳ, ʈ] vary from alveolar to postalveolar and differ from the main allophones of the laminal /d, n, t/ in that they're apical. I propose writing these with ⟨d̠, n̠, t̠⟩, with the retracted diacritic indicating that they're articulated in a more posterior position in comparison with the main allophones of /d, n, t/.
• The 'retroflex' [ɭ] seems to have merged with [l] (which used to be laminal (denti-)alveolar) into an apical (post)alveolar [l]. Capturing this detail here could lead to rather absurd transcriptions such as [ˈɭɔːt] for låt (which you can find in Kristoffersen's Phonology of Norwegian. But he uses them inconsistently). In other words, we'd have to turn many instances of what we write with ⟨l⟩ into ⟨ɭ⟩. But the formerly laminal [l̻] is mandatorily preserved after laminal nasals and stops as well as after back vowels (but not always after the close ones). In the latter case, it is velarized [ɫ̻] and it contrasts with a clear [l] (phonemically /rl/) in words like Karl [ˈkɑːl] and forlag [²fɔlːɑːɡ]. I propose that we transcribe the non-velarized variants (both apical and laminal) with ⟨l⟩ and the velarized laminal variant with ⟨ɫ⟩. The latter clearly belongs to the /l/ phoneme, whereas the former can be either /l/ or /rl/ phonemically.
• However, both [ɽ] and [ɾ] are apical and both are alveolar! The former is a flap and the latter is a tap. This contradicts what we can read in the lead of alveolar flap. However, since there's no way to distinguish alveolar flaps and taps in the official IPA, any convention we adopt will by definition be an ad-hoc one. I propose that we distinguish them in an ad-hoc way using ⟨ɾ̠⟩ for the flap and ⟨ɾ⟩ for the tap. The flap sounds retroflex but it's not.
• [ʂ] clearly isn't properly retroflex as it's pronounced with a flat tongue, like [ʃ] in French, Russian and Polish (we do use the symbol ⟨ʂ⟩ in transcriptions of the last two languages on WP. In my opinion this is a mistake and many sources write this sound with ⟨ʃ⟩). The same goes for Mandarin /ʂ/ which should be written with ⟨ʃ⟩. The confusion here seems to be that the former official IPA name of the sound represented by ⟨ʃ⟩ is voiceless palato-alveolar fricative. But that's the problem: it's not palato-alveolar in French, Russian or Polish. It's just (post)alveolar with no palatalization (why on earth would the IPA provide a symbol for [ɕ] if [ʃ] were slightly palatalized by definition? Such a contrast seems to be impossible to maintain in any language). The current official name of the sound represented by ⟨ʃ⟩ is, much more appropriately, voiceless postalveolar fricative. Besides - some sources already write the Norwegian sound with ⟨ʃ⟩. We can just follow them.

It's worth noting that none of these retroflex sounds except /ʂ/ are phonemic (but then again - phonemically, [ʂ] is both /ʂ/ and /rs/). /ɽ/ has a weak phonemic status and can be mostly counted as an allophone of /l/ or /r/. The situation of what Kristoffersen transcribes with [ɭ] is complicated and has been explained above. The retroflex symbols seem to be overused in the literature. We use them even more often on WP (see Russian, Polish, etc.) In the case of fricatives and affricates, it's clearly due to the anglocentric view which says that [ʃ, ʒ, tʃ, dʒ] are palatalized by definition ('palato-alveolar'). They aren't, and pronouncing Norwegian /ʃ/ with a slight palatalization wouldn't be a massive mispronunciation anyway (our native English-speaking readers have other things to worry about, such as correctly producing vowel length and tonemes).

These changes seem to overly complicate UEN transcriptions. Personally I don't think so, compared with the system we use now (with all the unnecessary lax symbols for the short vowels which differ in almost nothing but length from the long vowels.) Besides, if we start indicating tone using proper IPA, we're gonna be making extensive use of diacritics anyway. Writing ⟨d̠, n̠, t̠, ɾ̠⟩ instead of ⟨ɖ, ɳ, ʈ, ɽ⟩ doesn't seem like a massive problem to me and I'd argue that ⟨ʃ, l, ɫ⟩ are more easily recognizable than ⟨ʂ, l, ɭ⟩.

Pinging @Aeusoes1 and Nardog: per the discussion at Help talk:IPA/Hindi and Urdu that we've had back in September last year. I know that my position here is somewhat contradictory (given what I've said in that discussion), but I'm curious what you think. Kbb2 (ex. Mr KEBAB) (talk) 12:11, 6 February 2019 (UTC)

If I recall correctly, a lot instances of what the literature calls "retroflex" are not actually retroflex in the strictest terms. This means there's a little wiggle room already in what these symbols can mean. We try to avoid diacritics when we can and it seems as though a lot of the proposals here (other than parsing velarized and non-velarized laterals) would clutter transcriptions with a lot of diacritics for pretty minor phonetic details. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:29, 6 February 2019 (UTC)

@Aeusoes1: Hmm, I guess I didn't think the post through.

Ok, I can withdraw the proposal to substitute ⟨d̠, n̠, t̠, ɾ̠⟩ for ⟨ɖ, ɳ, ʈ, ɽ⟩. I admit that the last subtitution doesn't make a lot of sense as the symbol is still incorrect (the only difference between [ɽ] and [ɾ] is the manner of articulation (flap vs. tap)). ⟨ɖ, ɳ, ʈ⟩ aren't really incorrect per se since they can be apical postalveolar as well as just apical alveolar. Retroflexes in Hindi have basically the same allophonic range, which is something that I forgot.

So, am I understanding it correctly that you consider replacing ⟨l, ɭ⟩ with ⟨l, ɫ⟩ to be a good substitution?

What about using ⟨ʃ⟩ for what we now write with ⟨ʂ⟩? I have sources that use it for both /ʂ/ and /rs/. If we're ditching ⟨ɭ⟩ then we might as well replace ⟨ʂ⟩. It's the same [ʃ] that you can find in French, Lithuanian and Serbo-Croatian, minor phonetic differences (velarization, labialization, etc.) aside. It's basically a flat postalveolar sibilant without palatalization. Kbb2 (ex. Mr KEBAB) (talk) 20:30, 6 February 2019 (UTC)

mål - goal

Hi, as a native British English speaker (RP) that has been learning Norwegian for a year now I am highly doubtful about the correspondence listed here between the vowels in mål and goal... in fact, unless you are taking the vowel in goal from a rare/non-standard form of English they are completely different (and in that case, please list which form of English you are using because it's definitely not RP or GA). — Preceding unsigned comment added by 84.215.97.119 (talk) 09:23, 12 November 2019 (UTC)

Ok since nobody who knows what they are doing (I don't) has taken a look at this I have taken it into my own hands and changed 'mål - goal' to 'mål - maul'. I have to say many of the things on this page are questionable, such as specifying "Australian" for some vowel correspondences which are the same in RP and GA (it would therefore be sufficient to not specify the type of English because writing "Australian" would imply that this is a vowel found only in Australian and not more common types of English). I'm particularly thinking about 'ære - mad' and 'kai - price', at a glance the others seem ok. Also I find it bizarre that "Scottish save" is used twice when there are plenty of words common to Scottish and the more popular types of English that use this vowel. And I have to say that ʁ sounds really quite different to the fricative at the end of Scottish loch. 84.215.97.119 (talk) 01:30, 13 November 2019 (UTC)

Stress and tone

In accordance with the discussion at Help talk:IPA/Swedish#stress on monosyllabic names, I plan to revise the notation of stress and tone as e.g. [ˈbœ̀nːər] for accent 1 and [ˈbœ̂nːər] for accent 2. Nardog (talk) 16:58, 31 January 2020 (UTC)

Moved from User talk:Nardog

The problem with indiscriminately updating the tones on the entire English Wikipedia, is that ˈ for tone 1 is also used when tone is not specified. This leads to Rosenborg (station) having [ˈrùːsn̩bɔrɡ] with tone 1, which is correct, but also Röyksopp having [ˈrœ̀ʏksɔp] with tone 1, which it's not. I'm not sure what the best way to solve this is. — Tob c (talk) 14:12, 9 February 2020 (UTC)

@Tob c: Thanks for the feedback. What do you mean by "tone is not specified"? Are you saying Röyksopp has tone 2, or it varies? Nardog (talk) 23:50, 10 February 2020 (UTC)

@Nardog: A dangerously late reply to that: Yes, Röyksopp has tone 2. The result of this change was that all articles with unspecified tone ('), which was/is a lot, got changed to tone 1. And as a reply to @Kbb2: Even if it should have been specified, it's often not. It was ambiguous before too, but now the tone seems specified, when in reality it's not. It is also possibly a lot more time consuming to fix now. Tob c (talk) 20:45, 23 February 2021 (UTC)

@Tob c: It should've been specified in the first place as Norwegian is a language with pitch accent. You can search for hastemplate:"IPA-no" insource:/\{\{IPA-no[^\}]*PUTYOURSYMBOLHERE/ (replace PUTYOURSYMBOLHERE with a symbol you want to search for, e.g. ⟨ɑ̀⟩ - you can use this IPA keyboard to combine the tonal diacritics with vowel symbols) and go through articles that use them. If the tone is wrong (remember that transcriptions linking to this guide are Standard East Norwegian), replace the low tone diacritic ⟨ɑ̀⟩ with the falling tone one (⟨ɑ̂⟩). Kbb2 (ex. Mr KEBAB) (talk) 08:50, 11 February 2020 (UTC)

No, one should use [̀àèìòùỳ] for tone 1 and [̂âêîôûŷ] for tone 2. Nardog (talk) 09:04, 11 February 2020 (UTC)

@Nardog: That's what I meant though. What do you mean? Kbb2 (ex. Mr KEBAB) (talk) 09:07, 11 February 2020 (UTC)

/\{\{IPA-no[^\}]*[̀àèìòùỳ]/ will return all instances with tone 1 and /\{\{IPA-no[^\}]*[̂âêîôûŷ]/ all instances with tone 2. Also note that the tool you mentioned doesn't support precomposed characters, so simply replacing ⟨ɑ⟩ in ⟨ɑ̀⟩ with e.g. ⟨e⟩ won't work; you have to use ⟨è⟩, which is one character in the Latin-1 Supplement block. Nardog (talk) 09:20, 11 February 2020 (UTC)

@Nardog: Yeah, that sounds like a better solution. Thanks. Kbb2 (ex. Mr KEBAB) (talk) 09:26, 11 February 2020 (UTC)

Pronunciation of western Norwegian towns/cities

There are no correct way to pronounce Norwegian names and words. It is both inaccurate and misleading to use eastern Norwegian pronunciation for western Norwegian towns/cities and regions. I therefore think guidelines need to change. 178.232.196.246 (talk) 19:52, 5 January 2024 (UTC)