Talk:Swedish phonology

Why not phonemic status of the retroflexations?
The retroflexations [ʂ], [ʈ], [ɳ] and [ɖ], ought to be regarded as phonemes (/ʂ/, /ʈ/, /ɳ/ and /ɖ/), since they are crucial (in standard central Swedish as well as most regiolects of Swedish) to tell differences like in the following pairs. I.e. there is true phonemic splitting in the pairs below:

Some oppostions involving /ɳ/

vara /`vɑːra/ ('be'), varna /`vɑːɳa/ ('warn', 'tip off'), vana (/`vɑːna/) ('habit')

Tor /tuːr/ (boy's name), torn /tuːɳ/ ('rook', 'tower'), ton /tuːn/ ('tone')

kor /kuːr/ ('cow's'), korn /kuːɳ/ ('barley'; 'grain'), kon /kuːn/ ('cone')

hor /tuːr/ ('prostitution'), horn /huːɳ/ ('antler'), hon /huːn/ ('the sink')

Some oppostions involving /ʂ/

far /fɑːr/ ('father'), fars /fɑːʂ/ ('father's'), fas /fɑːs/ ('phase')

mor /muːr/ ('mother'), mors /muːʂ/ ('mother'), mos /muːs/ ('mashed potatoes')

kor /kuːr/ ('cows'), kors /kuːʂ/ ('cows'), kos /muːs/ ('cow's')

Some oppostions involving /ɖ/ and /ʈ/

mor /muːr/ ('mother'), mord /muːɖ/ ('murther'), mod /muːd/ ('courage')

torr /tɔrː/ ('dry'), torrt /tɔʈː/ ('dry'), Tott /tɔtː/ (an old cartoon figure 'Knoll och Tott')

vår /voːr/ ('our'), vårt /voːʈ/ ('our'), våt /voːt/ ('wet')

våra /`voːra/ ('our' pl.), vårta /`voːʈa/ ('wart'), våta /`voːta/ ('wet' pl.)

bor /buːr/ ('live', 'reside'), bord /buːɖ/ ('table'), bod /buːd/ ('shed')

Failing to distinguish mod ('courage') from mord ('murther'), unequivocally changes the meaning of a message... All the examples above show that there is true phonemic opposition between /ʂ/, /ʈ/, /ɳ/, /ɖ/ and the corresponding consonants, respectively. Also, it shows that the phonemic oppositions are not a marginal phenomenon, but a frequently recurring one.

Also, if comparing with the treatment of English some consonant clusters are always treated as individual phonemes (/tʃ/ and /dʒ/) despite that the difference between /t/ + /ʃ/ vs. /tʃ/ is arguable much smaller than in Swedish between /s/ and /ʂ/ and is not truly phonemic in i meaning-splitting sense.

Furthermore, the phonemic status of /ʂ/ is strengthened by the fact that in standard central Swedish (and actually in all of Sweden except the most southern provinces), dusch /dɵʂ/ ('shower') rhymes with kurs /kɵʂ/ ('course').

Sylvaticum (talk) 13:17, 27 July 2018 (UTC)


 * Moreover, the distinction between /n/ /ŋ/ and /g/ works in a highly similar way: tugga /`tɵga/ ('chew'), tunga /`tɵŋa/ ('tongue'), tunna /`tɵna/ ('barrel'). Since /ŋ/ is regarded as a distinct phoneme, /ʂ/, /ʈ/, /ɳ/ and /ɖ/ ought to be regarded as that as well. Sylvaticum (talk) 14:07, 25 August 2018 (UTC)

There are many reasons, but one of them is that not all Standard Swedish speakers realize the retroflex consonants as such, pronouncing them instead as /r/ + the affected consonant. In those cases, the pronunciation of /r/ is highly variable - this is why, for example, the table does not indicate [ʐ] as a phoneme despite being the pronunciation of /r/ in central Sweden because not all Swedish speakers pronounce /r/ as [ʐ], and that is so mentioned in the article. Likewise, while it's possible to give [ʂ] phonemic status because it occurs outside of the /r/+/s/ position, but that is already taken care of within the article as one of the "light" sounds that can occur for /ɧ/; as a result, [ʂ] is an allophone of two separate consonants. The point is that in the analysis backed up (for now - there's still other areas where citations are needed) by academic literature and professional linguists who have analyzed Swedish and has been chosen for the article, the retroflex consonants are allophones of existing consonants (in this case, a rhotic consonant followed by a coronal [dental/alveolar] consonant), which for those speakers who pronounce it are predictable from context and only occur if /r/ is also alveolar (and even then not in all areas, as Standard Swedish in Finland demonstrates, as well as the area around Gothenburg where there are no retroflexes but some of its effects remain; e.g. in your [ʈ]/[ɖ] examples the latter two of each pair are merged with [t]/[d] but are distinguishable in context despite apparent homonymity at first glance), and this is so mentioned in the article under the "Sonorants" section as occurring in certain areas of the Swedish-speaking world and in certain positions - a position also taken in the Norwegian phonology article, where Urban East Norwegian and some Norwegian regional varieties also possess retroflexes in similar contexts as Standard Swedish. The best example of that in English is not the affricates, but the /t/-/d/ merger in North America to a flapped [ɾ] (and even then not it all areas of North America) when intervocalic; its occurrence does not create a phoneme */ɾ/ in opposition to /r/ [ɻʷ] due to a series of complex factors. — Preceding unsigned comment added by 100.10.82.67 (talk) 15:23, 30 November 2019 (UTC)
 * In addition to previous discussion on this talk page, both above this and in archive, there's also an overview of Swedish phonology published in 2014 by Tomas Riad which mentions the retroflexes, not promoting them to phonemes but instead mentions it as part of a sandhi rule on pp. 73-79. Riad also does the opposite of the consensus elsewhere, by promoting /ʂ/ as a phoneme and treats the various allophones of [ɧ] within that phoneme.  Note, however that [ʂ] from /r/+/s/ is treated as a sandhi-related allophone - something already mentioned in the article with different source citations.  — Preceding unsigned comment added by 100.10.82.67 (talk) 16:10, 30 November 2019 (UTC)
 * He doesn't assign -the-retroflex-rs a phonemic status. ⟨ʂ⟩ is his transcription of what is (perhaps) more commonly written with ⟨ɧ⟩ in the literature. Kbb2 (ex. Mr KEBAB) (talk) 08:00, 22 April 2020 (UTC)

/d/ into /r/ or /ɾ/
I feel like I'm blind, as I can't find any mentions of /d/ turning into rhotics in certain words, other than a mention in the article about Dental and alveolar taps and flaps.

Anyway, sentences like "Vad sa du?" often get pronounced (pardon my IPA) in colloquial speech, enough to warrant a mention on LearningSwedish.se.

So, eh... what to say about it?

OliviaEljest (talk) 04:17, 9 August 2019 (UTC)

I think the ɵ sound should be in the close-mid category
As the title says, I think the ɵ sound should be in the close-mid category and not the open-mid.

I think this chart speaks for itself:

The ɵ sound is almost right in the middle but a little closer to close vowels, and certainly not open-mid as @Fdom5997 insists on. Pol Cəl (talk) 07:56, 4 November 2022 (UTC)


 * Phonemically, yes it is close-mid, but phonetically (and as also described in the chart) it is slightly lowered, as more open. Fdom5997 (talk) 08:27, 4 November 2022 (UTC)
 * Actually in the chart it's slightly raised* not slightly lowered. Take a ruler out and try to measure it, it's CLEARLY close-mid both phonetically and phonemically, as long as we're looking at the same chart File:Swedish monophthongs chart.svg Pol Cəl (talk) 14:58, 15 November 2022 (UTC)
 * It's not phonemically close-mid but close. It's the short counterpart of . Phonetically, it is as low as and, being true-mid, right in-between close-mid and open-mid. Sol505000 (talk) 14:39, 18 November 2022 (UTC)

It could be either close, close-mid or true-mid both phonemically and phonetically, but for sure it is not open-mid as @Fdom5997 says. However, the vowel chart shows phonetically a more close-mid or true-mid position. Pol Cəl (talk) 07:21, 21 December 2022 (UTC)


 * If the sound is phonetically lowered (as addressed per source), then yes it is technically *closer* to being considered open-mid, but it is not considered to be close-mid, however; it is still phonologically transcribed as a close-mid, central, rounded vowel. Fdom5997 (talk) 20:13, 21 December 2022 (UTC)

It is NOT technically open mid, because to assert that you'd have to look at the diagrams, and it's clearly closer to being close than open. At best, it's true-mid but definitely not open-mid. It might have been more open historically, but that doesn't affect the reality of the sound in the present. Ask any competent linguist and they're going to tell you the same thing. Pol Cəl (talk) 16:44, 31 December 2022 (UTC)

kv and kn
The article states, right at the beginning, that "Swedish also features consonant clusters not seen in other Germanic languages such as the "kv" consonant cluster and the "kn" consonant cluster." As far as I can tell, this is simply wrong. If no-one can convince me otherwise, I will delete the sentence. See, amongst others: etc.
 * Norwegian: kvinne, knapp
 * Danish: kvinde, knap
 * Faroese: kvinna, knappur
 * Icelandic: kvinna, knappur
 * German: quälen /ˈkvɛːlən/, Knecht
 * Dutch and North Frisian: knecht
 * Saterland Frisian: Knächt
 * West Frisian: knjocht

Tynnoel (talk) 18:49, 30 March 2023 (UTC)

Stress and pitch accent
I am somewhat bewildered by the signs marking stress and pitch accent in this article. They do not seem to be IPA. Is the author not satisfied with IPA's standards? 109.104.19.6 (talk) 06:45, 18 April 2023 (UTC)