According to Wikipedia,

Liaison (French pronunciation: ​[ljɛ.zɔ̃]) is the pronunciation of a latent word-final consonant immediately before a following vowel sound.

Scandinavian languages like Swedish, Norwegian, and Danish tend to drop final consonant sounds in speech, but do they experience liaisons when spoken?

For instance in Swedish: Jag is typically pronounced as Ja, but would Jag är be pronounced as Ja gär?

Does this happen in all Scandinavian languages? Is there any literature regarding this?

  • Are you telling us that you know as a matter of fact that [g] is pronounced in the phrase, in Swedish? Or are you asking "and is <jag är> pronounced [jɑ ar]?". – user6726 Jun 8 '16 at 15:23
  • @user6726 The latter. I'm a beginning learner of Swedish, so most of my exposure to the language has been ceremoniously hyper-enunciated. I'm interested to know if liaisons occur in Swedish (and other Scandinavian languages). – erip Jun 8 '16 at 15:45

As far as I know, the unpronounced citation consonants of are not ever recovered phrasally before a vowel. There is the puzzle of knowing which orthographic consonants are unpronounced (Norwegian [rø] "red", Swedish [rød], but note that even before a vowel suffix, the is gone in Norwegian ([røeste] "reddest"). I cannot parse Danish consonants, and I'd suggest consulting Basbøl's book if you want to know how that stuff is pronounced.

  • I agree. I can think of no instances of liaison in any Scandinavian language, either—including Danish. The only possible example I can think of is that han ‘he’ and hun ‘she’ nearly always lose their final /n/ before a consonant in rapid speech in Danish, but only facultatively before vowels. In slower or more careful speech, though, they don’t lose their nasal at all. – Janus Bahs Jacquet Jun 14 '16 at 22:43

Liaison exists in Danish but only in few words by inflections and word formation.


mand [ma̝nˀ] "man" > mandig [ma̝ndi] "manly"

valg [va̝lˀ] "choice" > valget [va̝ljð̠ˠ̞] "the choice"

valg is also sometimes heard as [ˈvalˀj], implying the j is not yet totally latent.

Such rules have a certain universality, i.e. < nd > + -ig and the latent j which is pronounced followed by such an ending.

It is also worth to mention that a [k] can appear after [ŋ], because this may exist in other languages but often ignored. An experiment below shows that the [k] is latent. Some participants were asked to add a verbalization suffix -ere, which starts with a vowel, to a word ending with < ng > [ŋ], to form a new word. Most participants add a [k], so [pɛŋˀ] becomes [pɛŋkeːˀɐ] with the ending. This case is pretty as similar as the mentioned example of < nd >.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.