Marked phonemes are those that require more effort during articulation or are "harder" to articulate. For example, the interdental fricatives /θ/ and /ð/ are considered to be marked. Marked phonemes are

  • Acquired later than unmarked phonemes in native acquisition.
  • Tend to pose more problems during Second Language Learning/Acquisition.
  • Are less frequent cross-linguistically.

English-speaking children frequently do not acquire /θ/ and /ð/ until age 5 or later.

Q: Are there marked or "hard" phonemes that are acquired even later in First Language Acquisition?

Q: Also, are there any phonemes that (are so marked that they) are not acquired at all by a substantial part of the population (say more than 5 %)?

EDIT: So far, most comments and one answer focus exclusively on the concept of the "marked phoneme". Let me make it clear that (1) I think this is a relatively well established concept - If you don't believe me google "markedness in phonology" and "marked phoneme". If you doubt the usefulness of this concept you're not the only one, but please let us move one from "I don't believe you" and "It's wrong", and please provide at least one academic reference why you think it's wrong or useless.

(2) More importantly, I couched my question in terms of markedness because I think it's a relatively uncontroversial concept, but it's really not central to my question. I'm asking whether there are phonemes that are acquired very late or never by a substantial number of speakers, whether you want to call them marked or articulatorily complex or challenging or hard. Answers should if possible be empirically sound such as quantitative claims from surveys, personal experience from teaching, child-rearing etc., with references where appropriate.

Note that I don't mean cases where there are differences between vernaculars and the standard languages and some speakers acquire the standard imperfectly (such as th-stopping in some English dialects, or the pharyngealised consonants present in Modern Standard Arabic but not in many Arabic dialects).

I have found literature on marked phonemes in a rage of languages, but not much on the age of acquisition of marked phonemes in different languages. Also, I have heard that the trilled /r/ present in Spanish and other languages is acquired very late and that some teachers think children often need help to properly acquire this sound - but this seemed rather unlikely to me if only for the reason that trilled /r/ appears to occur in a good number of languages and claims that children need "help" in acquiring their first language sound more like a difference between dialect and standard and teachers' desire to make their students use standard language.

  • 2
    Speaking only from my experience in parenting, I can tell you that my own children acquired interdental distinctions much earlier than resonants. My daughter was saying /y/ for /l/ at 5 or so, long after her dental fricatives were normative. All the kids in her cohort had delayed /r ~ l ~ y/ acquisition. I don't recall any /w/'s in the mix, but they would fit right in. Resonants are really variable phonetically.
    – jlawler
    Commented Aug 25, 2013 at 18:58
  • >usually do not acquire. The page you link to says "frequently". How did frequently turn into usually? To me, they're far from synonymous.
    – dainichi
    Commented Aug 27, 2013 at 6:53
  • 1
    I'm not convinced that the concept of a marked phoneme is established. The two references you give are to an unreferenced section of a Wikipedia article, and to a linguist-list posting which uses the phrase in the title but not in the body.
    – Colin Fine
    Commented Oct 3, 2013 at 20:49
  • Markedness is an extremely common concept in phonology. Just google it! You'll get references such as this, this, this, this, including criticism of it.
    – robert
    Commented Oct 3, 2013 at 21:35

2 Answers 2


Your definition of "marked phoneme" is wrong. Marked phonemes are those that are cross-linguistically relatively uncommon, and are prone to disappear during sound changes. This doesn't necessarily have anything much to do with how "hard" the sounds are to articulate.

It's true that sounds with complicated articulations are often simplified (e.g. the Proto-Indo-European sound /gwh/ with secondary labialization and breathy-voice is the most "complex" sound and the one most liable to disappear. But the markedness of most phonemes is due to the lack of auditory prominence. This is e.g. what makes /θ/ and /β/ marked in comparison with /f/, /s/, /v/ etc.

From an articulatory standpoint, /θ/ is easier than /s/; this is in fact the origin of a "lisp", which replaces /s/ with /θ/. Yet /s/ is one of the most stable, most common and least marked sounds cross-linguistically, due to the fact that it has extremely high acoustic prominence (that's why people say "pssss" or "sssst" to get someone's attention).

The hardest English sound to pronounce is almost certainly /r/, and it's often the last sound learned by children. To pronounce this sound you have to retroflex the front of the tongue, and bunch up (velarize) the back of the tongue, and round the lips, all at the same time. Foreigners have a hell of a time learning this sound. English /l/ is also difficult, and (as mentioned above) /s/. But all three sounds are highly stable and highly prominent. The Wikipedia article claiming that /θ/ is difficult to learn is certainly wrong; this impression comes from adult foreign-language speakers, who in general have difficulty with all sounds missing from their native languages.

  • Thanks for your answer! Could you please provide references linking "lack of auditory prominence" and markedness, and that /θ/ is easier to articulate than /s/, apart from lisping? Also for "the hardest English sound to pronounce is /r/", which, to be honest, seems rather subjective. For example, German-speaking learners of English do not tend to find this the most challenging English phoneme (from my experience teaching them).
    – robert
    Commented Oct 15, 2013 at 11:20
  • And see p. 78/88 of this chapter by Elizabeth Hume for the connection between markedness and ease of learning in phonology. It's certainly not something that everyone subscribes to 100%, but it's not "wrong" in the sense that no one believes in the validity of the link.
    – robert
    Commented Oct 15, 2013 at 11:23

Not exactly what you are asking, but in Spanish, /ʝ/ and /ʎ/ are considered different phonemes and are not distinguished at all by most speakers, not being really considered a question of standard vs. vernacular: http://en.wikipedia.org/wiki/Ye%C3%ADsmo

  • Thanks, so some dialects make a distinction and others don't, if I understand the Wikipedia article correctly?
    – robert
    Commented Sep 2, 2013 at 15:45
  • It is not so much a matter of dialects as of speakers. Yeísmo is not considered substandard or dialectal, it pervades all social dialects (but perhaps for some high cultural classes or University professors who are consciously against it) and it is no longer a hint of the origin of the speaker... I think it is rather that most speakers do not categorise the distinction between the sounds. It is also interesting that it affects other languages in contact with Spanish, such as Catalan or Galician(which prescriptively only presents /ʎ/, but most of its speakers produce something much nearer [ʝ])
    – Castellum
    Commented Sep 3, 2013 at 10:06
  • 2
    Maybe I'm missing something here, but the maps in the Wikipedia article indicate areas where the distinction is made and areas where it isn't made. So it's a regional distinction, i.e. a question of dialect (which doesn't exclude the possibility of overlapping regions or other factors co-determining the use of the feature). Or are you suggesting the Wikipedia article is wrong (which is of course a possibility)? What you call social dialects is not what I mean by dialect. I am talking about regional dialects, social dialects are also called sociolects.
    – robert
    Commented Sep 3, 2013 at 10:14

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.