17

Looking at the IPA, many different types of sounds are given symbols based of of the Latin R,r: approximants, trills, taps/flaps; both coronal and uvular segments.

Sometimes, these sounds are historically related, for example the French uvular approximant /ʁ/ replaced the earlier alveolar trill /r/.

My question is, why do these sounds pattern together despite being so phonetically different? Why do they often historically interchange and why do foreign speakers perceive foreign rhotics as comparable to their own, even if they are vastly different in terms of articulation?

  • 6
    It may be of interest that r sounds sometimes pattern partly like vowels as in at least Croatian and Slovenian. Also voiced velar fricatives might be an r in some languages but not in others such as Arabic غ and Georgian ღ, both of which have another sound which functions as an r. – hippietrail Jan 13 '12 at 07:47
  • 5
    In contemporary Polish and (historically) in Germanic languages, voiced s(h)ibilants like z and ʒ can also pattern like r or even alternate with them. So I'd agree with @Askalon below that it's hard to characterize segments as "rhotic" on a purely phonetic basis. – Mark Beadles Jan 13 '12 at 14:15
  • 2
    Huh, I guess I'll need a new question then: what defines a rhotic, if such a class can be said to exist?

    I am not entirely convinced that there is no phonetic distinction between, for example, English flapped /d/ and Spanish /r/. Has any study looked at this, either formant wise or by seeing whether speakers of either language can distinguish a Spanish (i.e. pronounced by a Spanish native speaker) /ɾ/ from an English one?

     –  Jan 17 '12 at 20:59
  • 1
    @Knitter There is a slight articulatory difference between Spanish /r/ and English flapped /t/ or /d/. One's a flap and one's a tap (unfortunately can't remember which is which), but most of the time a distinction isn't made between the two because they're very similar. So you're right that there is a phonetic distinction between them, but it's not because one's rhotic and the other isn't, it's because they could be considered different segments. – Askalon Jan 18 '12 at 03:21
  • @hippietrail: And American English, for that matter. – Mechanical snail Sep 21 '12 at 00:20

4 Answers4

6

There isn't any phonetic basis that captures all the various R sounds. A rhotic is considered such based on phonological factors, i.e. rhotics pattern in similar ways across languages.

Note that the same segment could be a rhotic in one language and non-rhotic in another. For example, as you may be familiar with, in North American dialects an alveolar stop often becomes an alveolar flap invervocalically after a stressed syllable (e.g. "ladder"). An alveolar flap/tap is also used in many other languages as an R. So that seems to be some evidence that there's no phonetic factor that will make a segment inherently rhotic.

Rhotics are a relatively poorly understood group of segments.

Askalon
  • 368
  • 2
  • 4
5

To add to Askalon's answer, there are many examples of phones that are rhotic in one language but not in another (here the phone in [] is an allophone of the // in each language):

  • [ɾ] is rhotic in e.g. Spanish /ɾ/, but not in American English /t/
  • [χ] in French /ʁ/, but not in Spanish /x/
  • [x] in Brazilian Portuguese, but not in German /x/
  • [ɣ] in Haitian Creole, but not in Japanese /ɡ/
  • [d̠ɹ̝] in Japanese /ɽ/, but not in English /dʒ/
  • [ʐ] in Mandarin /ɻ/, but not in Russian /ʐ/
  • [ʂ] in Scandinavian languages /ɾs/, but not in Russian /ʂ/
  • [ʋ] in some dialects of English /R/, but not in Finnish /ʋ/
  • [ɹ] in various languages, but not in Spanish /s/
  • [ʁ] in French /ʁ/, but not in Kabardian /ʁ/
  • [ʀ] in Portuguese /ʀ/, but not in Lakota /ʁ/ apparently

Wikipedia says Brazilian Portuguese has [r ɻ̝̊ ç x ɣ χ ʁ ʀ ħ h ɦ] as allophones of /ʁ/, most of which are not rhotic in many languages.

Community
  • 1
Mechanical snail
  • 3,254
  • 1
  • 20
  • 41
4

A phonetician can probably give a better explanation, but I used to worry about this, too, and the conclusion I came to was that when you analyze any language, there's always a few consonants that crowd the borderline with vowels, and form lots of consonant clusters, and weird vowel diphthongs, and they always turn out to be sonorants and laterals like [r ʀ ɾ ɹ l ɭ ɫ], etc.

There may be only one phoneme in a given language, like Mandarin or Japanese, that shares these allophones and patterning; or there may be as many as 4 or more (usually 2 lateral, 2 retroflex/2 palatal, 2 velarized, or some such). But there are always some.

The ones in one language might count as normal stops in some other language, but in the one they're sonorants and pattern the same as sonorants in any other language, mutatis mutandis.

jlawler
  • 10,042
  • 1
  • 31
  • 55
  • 2
    I'm not saying I disagree, but I'd be interested in knowing what these patterns are for various langauges. For instance, you mention Japanese, but in Japanese, the [ɾ] doesn't seem to me to form consonant clusters or weird vowel diphthongs. As for "crowd the borderline with vowels", not sure exactly what you mean, but cannot see any difference from many other Japanese consonants in this respect. – dainichi Aug 01 '12 at 00:31
  • It doesn't happen in Japanese because there are no consonant clusters in Japanese. And the Japanese /r/ has several allophones, encompassing allophones of English /r/ and /l/. – jlawler Aug 09 '12 at 18:35
  • 1
    The Japanese /r/ is a postalveolar flap, although it can vary between central and lateral. So as far as I can see, it encompasses none of the allophones of English /l/ and /r/, except for maybe the alveolar flap. – dainichi Aug 09 '12 at 23:25
  • And regarding consonant clusters, depending on how you interpret vowel devoicing, you can get 3-consonant clusters like [tsk], but /r/ is never part of them, since vowel devoicing doesn't happen around /r/. – dainichi Aug 09 '12 at 23:32
  • Of course; but those aren't phonemic clusters, only the result of devoicing. – jlawler Aug 10 '12 at 00:09
  • 2
    OK, then. But I still don't see how sonorants and laterals are special in Japanese. – dainichi Aug 10 '12 at 00:17
  • 1
    Probably not all that special; on the other hand, Japanese has an extremely limited repertory of possible syllables, so it's likely to be an exception to a lot of generalizations. – jlawler Aug 10 '12 at 00:22
-1

This is because the anatomic phonetic theory is a complete fail. People perceive sounds not dependent on how they are articulated. It is possible to pronounce uvular and alveolar "r" in such way that a listener hardly could determine how the sounds were pronounced.

Anixx
  • 6,643
  • 1
  • 26
  • 38
  • 2
    I presume by 'anatomic phonetic theory' you are referring to articulatory phonetics? If so, it is correct that articulatory phonetics does not explain perception; but it does not try to, that job is left to the field of auditory phonetics. – Gaston Ümlaut Sep 22 '12 at 02:29
  • -1: This doesn't answer the question because even if articulatory phonetics is completely bogus, there are phones that would be perceived as rhotic and non-rhotic in two different languages. – Mechanical snail Sep 22 '12 at 03:56