I have an ASCII system for writing sounds mapped to IPA letters, but I can't figure out what the "l" sound is as in Danish "sidde". It's pronounced sort of like (if I were to try to write it out) "sihlthle". What IPA symbol is that? I don't see it here, and after talking with a Danish speaker, they said it's a unique sound that wasn't represented by some other common IPA stuff I used (represented in an ascii system).

  • 3
    fyi, there is a standardised ASCII equivalent to the IPA already which is X-SAMPA. For transferability of your project, it may be worthwhile using it, rather than a homebrewed system
    – Tristan
    Dec 15, 2020 at 10:29
  • Homebrewed system is much better and more user friendly than X-SAMPA. I have not met anyone who uses X-SAMPA or IPA for writing pronunciations. I have seen the results, so someone is doing it, but very small proportion of speakers know how to do it, let alone what it is or having the desire to learn it.
    – Lance
    Dec 15, 2020 at 16:18
  • 5
    @LancePollard IPA is absolutely not reserved for “the most hardcore people”. My seventh-grade French textbooks all had pronunciations in IPA. Wikipedia uses IPA (with help files, but still). Most dictionaries use IPA. Even tourist parlours often use IPA. It is by far the most common way of writing pronunciations. Also, it is there on the IPA for Danish, given as just plain /ð/, which is fine if you’re using (broad) IPA for Danish exclusively; it only becomes a problem when you also need to cater for an actual IPA /ð/ like the English one. (I agree that X-SAMPA is very user-unfriendly, though.) Dec 15, 2020 at 16:24
  • 3
    Again, to close-voters: how is this a language-specific grammar and usage question? It's a question about how to transcribe a particular sound in IPA.
    – Draconis
    Dec 15, 2020 at 19:37

2 Answers 2


Wikipedia says "/ð/ – the so-called 'soft d' (Danish: blødt d) – is a velarized laminal alveolar approximant [ð̠˕ˠ]", citing Basbøll (2005), Grønnum (2003), and Ladefoged & Maddieson (1996).

  • is the IPA right here? I'd expect a velarised laminal alveolar approximant to be [ð̻̞ˠ]. This seems to use the retracted rather than laminal diacritic and have the lowered diacritic (used to show an approximant) following rather than properly stacked
    – Tristan
    Dec 15, 2020 at 10:36
  • looking on the wiki talk page, it looks like the lack of stacking is a deliberate style choice on wikipedia (even though I believe it is not proper to the IPA) as noted by you it turns out, but the use of a retracted diacritic rather than a laminal one is odd given the description
    – Tristan
    Dec 15, 2020 at 10:40
  • @Tristan [ð] is dental so the retracted diacritic makes sense. [ð̠˕ˠ] is indeed the transcription Basbøll and Grønnum give (though with different diacritic-stacking practices). The IPA doesn't specify apicality/laminality in the values of the coronal symbols (see the Principles of the IPA, reprinted in IPA Handbook Appendix 1), so noting the laminality is not a must.
    – Nardog
    Dec 15, 2020 at 11:03
  • sure, my question isn't so much why this transcription, but the transcription and description do not match. If laminal is specified in the description it ought to be in the transcription too (although if it is also in the sources, I understand why wikipedia can't change that)
    – Tristan
    Dec 15, 2020 at 11:17
  • @Tristan I'm saying they do match. [ð] is dental, not alveolar, hence the retracted diacritic.
    – Nardog
    Dec 15, 2020 at 11:37

This gets in to the difference between phonetic and phonological descriptions

The phoneme in question is generally called /ð/. The issue is that the precise phonetic details of how this phoneme is produced, differ quite a lot from prototypical [ð] (which is why English-speakers generally hear it as closer to our /l/ than our /ð/)

As Nardog says in their answer (citing Basbøll (2005), Grønnum (2003), and Ladefoged & Maddieson (1996) via wikipedia), this phoneme is pronounced as a velarised laminal alveolar approximant so if you need a precise phonetic transcription, [ð̠̻̞ˠ] would be appropriate

Note: wikipedia gives the phonetic transcription as [ð̠˕ˠ] (with no laminal diacritic, despite this being specified in the description of the sound in words, and also rendered with the lowered diacritic following rather than stacked)

  • 1
    The version you give here is a velarised laminal dental approximant, which the soft d is not – you need the retracted symbol to make it alveolar. The alveolar place of articulation is essential, but the laminality isn’t really – that is, pronouncing it apically still produces something that sounds like a soft d, but pronouncing it dentally produces a sound that is audibly different and doesn’t sound like a natural soft d. Dec 15, 2020 at 13:14
  • good point at least in narrow transcription as here. More loosely, when I've seen ð̞ used it's been to refer to an alveolar approximant rather than a dental one, despite that strictly not being how its defined (likewise, see the way that [a] is often used as mid, with [æ] as fully open in the position [a] is officially defined as occupying). I've edited the retracted diacritic into my transcription, and changed my note on the wikipedia form to note only that the laminality is not specified (and the lowered diacritic is following not stacked)
    – Tristan
    Dec 15, 2020 at 13:54
  • not only English-speakers. As a German, I hear (and try to pronounce) it as a mix of /l/ and /d/ (and not close to anything English) Dec 15, 2020 at 20:18
  • 1
    @HagenvonEitzen In my experience, it is very common for speakers of many different languages to hear the soft d as an l or l-like sound. The problem is when they then also produce it as such, because that actually hinders intelligibility: Danes tend to hear the ‘l-like sound’ produced by such as people as an actual l, a different phoneme, so they’ll hear rude (window pane) as rulle (roll). Generally speaking, it’s better in terms of comprehension for non-native speakers to approximate a lazy [ð] than any kind of l-like sound. Dec 16, 2020 at 11:04
  • this sort of mismatched perceptions is responsible for a lot of confusions between languages tbh. Mandarin speakers tend to hear English /l/ & /ɹ/ as Mandarin /l/, whilst English speakers hear Mandarin /ʐ/ (pinyin <r>) as /ɹ/ meaning to English speakers it sounds like Mandarin-speakers just obstinately refuse to use /ɹ/ despite being able to produce it (I assume there is an inverse problem as well with English-speaking Mandarin students not being perceived as distinguishing /l/ & /ʐ/ because the students are making a distinction the Mandarin speakers just are not used to listening for)
    – Tristan
    Dec 16, 2020 at 13:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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