Voiced velar approximant

Features of the voiced velar approximant:

• Its manner of articulation is approximant, which means it is produced by narrowing the vocal tract at the place of articulation, but not enough to produce a turbulent airstream. The most common type of this approximant is glide or semivowel. The term glide emphasizes the characteristic of movement (or 'glide') of [ɰ] from the [ɯ] vowel position to a following vowel position. The term semivowel emphasizes that, although the sound is vocalic in nature, it is not 'syllabic' (it does not form the nucleus of a syllable). For a description of the approximant consonant variant used e.g. in Spanish, see below.
• Its place of articulation is velar, which means it is articulated with the back of the tongue (the dorsum) at the soft palate.
• Its phonation is voiced, which means the vocal cords vibrate during the articulation.
• It is an oral consonant, which means air is allowed to escape through the mouth only.
• It is a central consonant, which means it is produced by directing the airstream along the center of the tongue, rather than to the sides.
• The airstream mechanism is pulmonic, which means it is articulated by pushing air solely with the intercostal muscles and diaphragm, as in most sounds.

Some languages have a velar approximant that is produced with the body of the tongue bunched up at the velum and simultaneous pharyngealization. This gives rise to a type of retroflex resonance resembling [ɻ].[19] The extension to the IPA recommends the use of the "centralized" diacritic combined with the IPA sign for the alveolar approximant (as in ⟨ɹ̈⟩) to distinguish the bunched realization from the prototypical apical [ɹ], which may be specified as ⟨ɹ̺⟩. Typically, the diacritic is omitted, so that the sound is transcribed simply with ⟨ɹ⟩ or ⟨ɻ⟩ (in broader transcriptions: ⟨r⟩), as if it were a coronal consonant.

In ordinary IPA, this sound can be transcribed with ⟨ɰˤ⟩ or ⟨ɣ̞ˤ⟩, depending on whether it can be labialized (see below). This captures pharyngealization (uncommon in velar approximants) but not bunching of the tongue, which can be transcribed by adding a superscript retroflex approximant: ⟨ɰˤʵ⟩, ⟨ɣ̞ˤʵ⟩ (see International Phonetic Alphabet § Diacritics and prosodic notation).

In Dutch, this type of r is called Gooise r [ˌɣoːisə ˈʔɛr] 'Gooi r'. It is named after het Gooi, a region of the Netherlands where Hilversum (the main centre for television and radio broadcasting) is located.

Some languages have a voiced velar approximant that is unspecified for rounding, and therefore cannot be considered the semivocalic equivalent of either [ɯ] or its rounded counterpart [u]. Examples of such languages are Catalan, Galician and Spanish, in which the approximant consonant (not semivowel) unspecified for rounding appears as an allophone of /ɡ/.[8]

Eugenio Martínez Celdrán describes the voiced velar approximant consonant as follows:[22]

As for the symbol ⟨ɰ⟩, it is quite evidently inappropriate for representing the Spanish voiced velar approximant consonant. Many authors have pointed out the fact that [ɰ] is not rounded; for example, Pullum & Ladusaw (1986:98) state that 'the sound in question can be described as a semi-vowel (glide) with the properties "high", "back", and "unrounded"'. They even establish an interesting parallelism: 'the sound can be regarded as an unrounded [w]'. It is evident, then, that ⟨ɰ⟩ is not an adequate symbol for Spanish. First of all, because it has never been taken into consideration that there is a diphthong in words like paga 'pay', vago 'lazy', lego 'lay', etc., and, secondly, because this sound is rounded when it precedes rounded vowels. Besides, it would be utterly wrong to transcribe the word jugo 'juice' with ⟨ɰ⟩ *[ˈχuɰo], because the pronunciation of that consonant between two rounded vowels is completely rounded whereas [ɰ] is not. [...]

The symbol I have always proposed is ⟨ɣ̞⟩, the correlate to the other central approximants in Spanish, [β̞ ð̞] (Martínez Celdrán 1991, 1996:47). This coincides with Ball & Rahilly (1999:90), whose example for the three approximants is the Spanish word abogado 'lawyer'[...]. Ball & Rahilly too criticise in a footnote the confusion between these symbols: 'The difference between an approximant version of the voiced velar fricative [ɣ̞], and the velar semi-vowel [ɰ] is that the latter requires spread lips, and must have a slightly more open articulatory channel so that it becomes [ɯ] if prolonged' (p. 189, fn. 1).

There is a parallel problem with transcribing the palatal approximant.

The symbol ⟨ɣ̞⟩ may not display properly in all browsers. In that case, ⟨ɣ˕⟩ should be substituted. In broader transcriptions,[23] the lowering diacritic may be omitted altogether, so that the symbol is rendered ⟨ɣ⟩, i.e. as if it represented the corresponding fricative.

The velar bunched approximant is typically written with ⟨ɹ⟩ or ⟨ɻ⟩ (in broader transcriptions: ⟨r⟩), none of which capture its place of articulation. The prototypical [ɹ], [ɻ] and [r] are all unspecified for rounding like [ɣ̞].

• Index of phonetics articles

• List of languages with [ɰ] on PHOIBLE