Nearly all nasal consonants are nasal occlusives, in which air escapes through the nose but not through the mouth, as it is blocked (occluded) by the lips or tongue. The oral cavity still acts as a resonance chamber for the sound. Rarely, non-occlusive consonants may benasalized.
Most nasals arevoiced, and in fact, the nasal sounds[n] and[m] are among the most common sounds cross-linguistically. Voiceless nasals occur in a few languages such asBurmese,Welsh,Icelandic andGuaraní. (Compare oralstops, which block off the air completely, andfricatives, which obstruct the air with a narrow channel. Both stops and fricatives are more commonly voiceless than voiced, and are known asobstruents.)
In terms of acoustics, nasals aresonorants, which means that they do not significantly restrict the escape of air (as it can freely escape out the nose). However, nasals are alsoobstruents in their articulation because the flow of air through the mouth is blocked. This duality, a sonorant airflow through the nose along with an obstruction in the mouth, means that nasal occlusives behave both like sonorants and like obstruents. For example, nasals tend to pattern with other sonorants such as[r] and[l], but in many languages, they may develop from or into stops.
Acoustically, nasals have bands of energy at around 200 and 2,000 Hz.
TheNuosu language also contrasts six categories of nasals,/m,n,m̥,n̥,ɲ,ŋ/. They are represented in romanisation by <m, n, hm, hn, ny, ng>. Nuosu also contrasts prenasalised stops and affricates with their voiced, unvoiced, and aspirated versions.
/ɱ/ is the rarest voiced nasal to be phonemic, its mostly an allophone of other nasals before labiodentals and currently there is only 1 reported language,Kukuya, which distinguishes/m,ɱ,n,ɲ,ŋ/ and also a set of prenasalized consonants like/ᶬp̪fʰ,ᶬb̪v/.[2]Yuanmen used to have it phonemically before merging it with/m/.[3]
Catalan,Occitan, Spanish, and Italian have/m,n,ɲ/ asphonemes, and[ɱ,ŋ] as allophones. It may also be claimed that Catalan has phonemic/ŋ/, at least on the basis ofCentral Catalan forms such assang[saŋ], although the only minimal pairs involve foreignproper nouns.[4] Also, among many younger speakers ofRioplatense Spanish, the palatal nasal has been lost, replaced by a cluster[nj], as in Englishcanyon.[5]
InBrazilian Portuguese andAngolan Portuguese/ɲ/, written⟨nh⟩, is typically pronounced as[ȷ̃], anasal palatal approximant, a nasal glide (inPolish, this feature is also possible as an allophone). Semivowels inPortuguese often nasalize before and always after nasal vowels, resulting in[ȷ̃] and[w̃]. What would becoda nasal occlusives in otherWest Iberian languages is only slightly pronounced beforedental consonants. Outside this environment the nasality is spread over the vowel or become a nasal diphthong (mambembe[mɐ̃ˈbẽjbi], outside thefinal, only in Brazil, andmantém[mɐ̃ˈtẽj~mɐ̃ˈtɐ̃j] in all Portuguese dialects).
TheJapanesesyllabary kanaん, typically romanized asn and occasionallym, can manifest as one of several different nasal consonants depending on what consonant follows it; this allophone, colloquially written in IPA as/N/, is known as themoraic nasal, per the language's moraic structure.
Welsh has a set of voiceless nasals,/m̥,n̥,ŋ̊/, which occur predominantly as a result ofnasal mutation of their voiced counterparts (/m,n,ŋ/).
TheMapos Buang language of New Guinea has a phonemic uvular nasal, /ɴ/, which contrasts with a velar nasal. It is extremely rare for a language to have /ɴ/ as a phoneme. The/ŋ,ɴ/ distinction also occurs in a fewInuit languages likeIñupiaq.Chamdo languages likeLamo (Kyilwa dialect),Larong sMar (Tangre Chaya dialect),Drag-yab sMar (Razi dialect) have an extreme distinction of/m̥n̥ȵ̊ŋ̊ɴ̥mnȵŋɴ/, also one of the few languages to have a[ɴ̥].[6]
Yanyuwa is highly unusual in that it has a seven-way distinction between/m,n̪,n,ɳ,ṉ/ (palato-alveolar),/ŋ̟/ (front velar), and/ŋ̠/ (back velar). This may be the only language in existence that contrasts nasals at seven distinct points of articulation.[7]
Yélî Dnye also has an extreme contrast of/m,mʷ,mʲ,mʷʲ,n̪,n̪͡m,n̠,n̠͡m,n̠ʲ,ŋ,ŋʷ,ŋʲ,ŋ͡m/.[8][9][10]
The term 'nasal occlusive' (or 'nasal stop') is generally abbreviated tonasal. However, there are also nasalized fricatives, nasalized flaps,nasal glides, andnasal vowels, as in French, Portuguese, and Polish. In theIPA, nasal vowels and nasalized consonants are indicated by placing a tilde (~) over the vowel or consonant in question: Frenchsang[sɑ̃], Portuguesebom[bõ], Polishwąż[vɔ̃w̃ʂ].
Ladefoged and Maddieson (1996) distinguish purely nasal consonants, the nasal occlusives such asm n ng in which the airflow is purely nasal, from partial nasal consonants such asprenasalized consonants and nasalpre-stopped consonants, which are nasal for only part of their duration, as well as fromnasalized consonants, which have simultaneous oral and nasal airflow.[11] In some languages, such asPortuguese, a nasal consonant may have occlusive and non-occlusiveallophones. In general, therefore, a nasal consonant may be:
A nasal trill[r̃] has been described from some dialects of Romanian, and is posited as an intermediate historical step inrhotacism. However, the phonetic variation of the sound is considerable, and it is not clear how frequently it is actually trilled.[12] Some languages contrast/r,r̃/ likeToro-tegu Dogon (contrasts/w,r,j,w̃,r̃,j̃/)[13] andInor.[14] A nasal lateral has been reported for some languages,Nzema contrasts/l,l̃/,[15]Nemi contrasts/w,w̥,h,w̃,w̥̃,h̃/.Ganza[16] contrasts/ʔ,ʔ̃/.
A few languages, perhaps 2%,[17] contain no phonemically distinctive nasals. This led Ferguson (1963) to assume that all languages have at least one primary nasal occlusive. However, there are exceptions.
When a language is claimed to lack nasals altogether, as with severalNiger–Congo languages[note 1] or thePirahã language of the Amazon, nasal and non-nasal or prenasalized consonants usually alternateallophonically, and it is a theoretical claim on the part of the individual linguist that the nasal is not the basic form of the consonant. In the case of some Niger–Congo languages, for example, nasals occur before only nasal vowels. Since nasal vowels are phonemic, it simplifies the picture somewhat to assume that nasalization in occlusives is allophonic. There is then a second step in claiming that nasal vowels nasalize oral occlusives, rather than oral vowels denasalizing nasal occlusives, that is, whether[mã,mba] are phonemically/mbã,mba/ without full nasals, or/mã,ma/ without prenasalized stops. Postulating underlying oral or prenasalized stops rather than true nasals helps to explain the apparent instability of nasal correspondences throughout Niger–Congo compared with, for example, Indo-European.[18]Proto-Mande has also been reconstructed with a system in which nasals are allophones of oral stops and approximants.[19][20]
This analysis comes at the expense, in some languages, of postulating either a single nasal consonant that can only be syllabic, or a larger set of nasal vowels than oral vowels, both typologically odd situations. The way such a situation could develop is illustrated by aJukunoid language,Wukari. Wukari allows oral vowels in syllables likeba, mba and nasal vowels inbã, mã, suggesting that nasals become prenasalized stops before oral vowels. Historically, however, *mb became **mm before nasal vowels, and then reduced to *m, leaving the current asymmetric distribution.[21]
In older speakers of theTlingit language,[l] and[n] are allophones. Tlingit is usually described as having an unusual, perhaps unique lack of/l/ despite having fivelateral obstruents; the older generation could be argued to have/l/ but at the expense of having no nasals.[citation needed]
Several of languages surroundingPuget Sound, such asQuileute (Chimakuan family),Lushootseed (Salishan family), andMakah (Wakashan family), are truly without any nasalization whatsoever, in consonants or vowels, except in special speech registers such asbaby talk or the archaic speech of mythological figures (and perhaps not even that in the case of Quileute). This is anareal feature, only a few hundred years old, where nasals became voiced stops ([m] became[b],[n] became[d],[ɳ] became[ɖ],[ɲ] became[ɟ],[ŋ] became[g],[ŋʷ] became[gʷ],[ɴ] became[ɢ], etc.) after colonial contact. For example, "Snohomish" is currently pronouncedsdohobish, but was transcribed with nasals in the first English-language records.[citation needed]
The only other places in the world where this is known to occur are in Melanesia. In the central dialect of theRotokas language of Bougainville Island, nasals are only used when imitating foreign accents. (A second dialect has a series of nasals.) TheLakes Plain languages of West Irian are similar.
The unconditioned loss of nasals, as in Puget Sound, is unusual. Currently inKorean,/m/ and/n/ are shifting to[b] and[d], but only word-initially. This started out in nonstandard dialects and was restricted to the beginning of prosodic units (a common position forfortition), but has expanded to many speakers of the standard language to the beginnings of common words even within prosodic units.[22]
^Maddieson, Ian. 2008. Absence of Common Consonants. In: Haspelmath, Martin & Dryer, Matthew S. & Gil, David & Comrie, Bernard (eds.) The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter 18. Available online athttp://wals.info/feature/18Archived 2009-06-01 at theWayback Machine. Accessed on 2008-09-15.
