In phonetics and phonology, a sonorant or resonant is a speech sound that is produced with continuous, non-turbulent airflow in the vocal tract; these are the manners of articulation that are most often voiced in the world's languages. Vowels are sonorants, as are consonants like /m/ and /l/: approximants, nasals, flaps or taps, and most[clarification needed]trills.
In older usage, only the term resonant was used with this meaning, and sonorant was a narrower term, referring to all resonants except vowels and semivowels.
Whereas obstruents are frequently voiceless, sonorants are almost always voiced. A typical sonorant consonant inventory found in many languages comprises the following: two nasals /m/, /n/, two semivowels /w/, /j/, and two liquids /l/, /r/.
In the sonority hierarchy, all sounds higher than fricatives are sonorants. They can therefore form the nucleus of a syllable in languages that place that distinction at that level of sonority; see Syllable for details.
Among consonants pronounced in the back of the mouth or in the throat, the distinction between an approximant and a voiced fricative is so blurred that no language is known to contrast them. Thus, uvular, pharyngeal, and glottal fricatives never contrast with approximants.
Voiceless sonorants are rare; they occur as phonemes in only about 5% of the world's languages. They tend to be extremely quiet and difficult to recognise, even for those people whose language have them.
In every case of a voiceless sonorant occurring, there is a contrasting voiced sonorant. In other words, whenever a language contains a phoneme such as /r?/, it also contains a corresponding voiced phoneme such as /r/).
Voiceless sonorants are most common around the Pacific Ocean (in Oceania, East Asia, and North and South America) and in certain language families (such as Austronesian, Sino-Tibetan, Na-Dene and Eskimo-Aleut).
Another European language with voiceless sonorants is Icelandic, with [l? r? n? m? ] for the corresponding voiced sonorants [l r n m ? ?].
Voiceless [r? l? ?] and possibly [m? n?] are hypothesized to have occurred in various dialects of Ancient Greek. The Attic dialect of the Classical period likely had [r?] as the regular allophone of /r/ at the beginning of words and possibly when it was doubled inside words. Hence, many English words from Ancient Greek roots have rh initially and rrh medially: rhetoric, diarrhea.
Old Irish had one of the most complex sonorant systems recorded in linguistics, with 12 coronal sonorants alone. Coronal laterals, nasals, and rhotics had a fortis–lenis and a palatalization contrast: /N, n, N?, n?, R, r, R?, r?, L, l, L?, l?/. There were also /?, , m/ and /m?/, making 16 sonorant phonemes in total.