Anusvara (Sanskrit: अनुस्वारः anusvāra) is the diacritic used to mark a type of nasal sound used in a number of Indic scripts. Depending on the location of the anusvara in the word and the language in which it is used, its exact pronunciation can vary greatly.
In the context of Sanskrit, anusvara may also refer also to the nasal sound itself.
In Vedic Sanskrit, the anusvāra (lit. "after-sound") is a sound that occurs as an allophone of /m/, at a morpheme boundary, or /n/ morpheme-internally, if either of them is preceded by a vowel and followed by a fricative (/ś/, /ṣ/, /s/ or /h/). The material in the various ancient phonetic treatises points to two possible phonetic descriptions. The anusvara could amount to an effect on the preceding vowel, consisting in nasalisation (different from the anunāsika) accompanied by lengthening. Alternatively, the anusvara could have been realised as a nasal fricative having the same place of articulation as the following consonant. The discrepancies have been attributed to differences of analysis of the same pronunciation or to dialectal or diachronic variation.
In the later language its use gradually expanded to other contexts. The anusvāra first began to be used before /r/ under certain conditions. Later, in Classical Sanskrit, its use had extended before /l/ and /y/, replacing earlier [l̃] and [ỹ]. It began to be used to replace the homorganic nasal before a plosive: Pāṇini gave such an optional use in sandhi word-finally, and later treatises also prescribed it at morpheme junctions and in intra-morphemic position.
In the Devanagari script, anusvara is represented with a dot (bindu) above the letter (e.g. मं). In the International Alphabet of Sanskrit Transliteration (IAST), the corresponding symbol is ṃ (m with an underdot). Some transcriptions render notation of phonetic variants used in some Vedic shakhas with variant transcription (ṁ).
In Marathi the anusvara is pronounced as a nasal that is homorganic to the following consonant (with the same place of articulation). For example, it is pronounced as the dental nasal न् before dental consonants, as the bilabial nasal म् before bilabial consonants, etc.
In Standard Hindi, the anusvāra is traditionally defined as representing a nasal consonant homorganic to a following plosive, in contrast to the candrabindu (anunāsika), which indicates vowel nasalization. In practice, however, the two are often used interchangeably.
Word-finally it is realized as nasalization of the preceding vowel: kuāṃ [kʊ̃ãː], "a well". It results in vowel nasalization also medially between a short vowel and a non-obstruent (kuṃvar [kʊ̃ʋər] "a youth", gaṃṛāsā [ɡə̃ɽaːsaː] "a long-handled axe") and, in native words, between a long vowel and a voiceless plosive (dāṃt [dãːt] "tooth", sāṃp [sãːp] "a snake", pūṃch [pũːtʃʰ] "tail").
It is pronounced as a homorganic nasal, with the preceding vowel becoming nasalized allophonically, in the following cases: between a long vowel and a voiced plosive (tāṃbā [taːmbaː] "copper", cāṃdī [tʃaːndiː] "silver"), between a long vowel and a voiceless plosive in loanwords (dāṃt [daːnt] "repressed", baiṃk [bæːŋk] "a bank", khazāṃcī [kʰəzaːɲtʃiː]), and between a short vowel and an obstruent (saṃbhāl- [səmbʱaːl] "to support", saṃdūk [sənduːk] "a chest").
The last rule has two sets of exceptions where the anusvāra effects only a nasalization of the preceding short vowel. Words from the first set are morphologically derived from words with a long nasalized vowel (baṃṭ- [bə̃ʈ], "to be divided" from bāṃṭ- [bãʈ], "to divide"; siṃcāī [sɪ̃tʃai], "irrigation" from sīṃc- [sĩːtʃ], "to irrigate"). In suchs cases, the vowel is sometimes denasalized ([bəʈ], [sɪtʃai] instead of [bə̃ʈ-], [sɪ̃tʃai]). The second set is composed of a few words like (pahuṃc- [pahʊ̃tʃ], "to arrive" and haṃs- [hə̃s], "to laugh").[note 1]
In Nepali, chandrabindu and anusvara have the same pronunciation similarly to Hindi. Therefore, there is a great deal of variation regarding which occurs in any given position. Many words containing anusvara thus have alternative spellings with chandrabindu instead of anusvara and vice versa.
Other Indic script languages
In the Bengali script, the anusvara diacritic (অনুস্বার onushshar in Bengali) is written as a circle above a slanted line (ং), and represents /ŋ/. It is used in the name of the Bengali language বাংলা [baŋla]. It has merged in pronunciation with the letter ঙ ungô in Bengali. Although the anusvara is a consonant in Bengali phonology, it is nevertheless treated in the written system as a diacritic in that it is always directly adjacent to the preceding consonant, even when consonants are spaced, apart in titles or banners: বাং-লা-দে-শ bang-la-de-sh, not বা-ং-লা-দে-শ ba-ng-la-de-sh for বাংলাদেশ Bangladesh It is never pronounced with the inherent vowel "ô", and it cannot take a vowel sign (instead, the consonant ঙ ungô is used pre-vocalically).
In the Burmese script, the anusvara (အောက်မြစ် auk myit IPA: [aʊʔ mjɪʔ]) is represented as a dot underneath a nasalised final to indicate a creaky tone (with a shortened vowel). Burmese also uses a dot above to indicate the /-ɴ/ nasalized ending (called "Myanmar Sign Anusvara" in Unicode), called သေးသေးတင် thay thay tin (IPA: [θé ðé tɪ̀ɴ])
In the Sinhala script, the anusvara is not a diacritic but an independent grapheme. It has circular shape ( ං), which is why it is called binduva in Sinhala, which means "dot". The anusvara represents /ŋ/ at the end of a syllable. It is used in the name of the Sinhala language සිංහල [ˈsiŋɦələ]. It has merged in pronunciation with the letter ඞ ṅa in Sinhala.
The Telugu script has full-zero (anusvāra) ం , half-zero (arthanusvāra) and visarga to convey various shades of nasal sounds. Anusvara is represented as a circle shape after a letter: క - ka and కం - kam.
The equivalent of the anusvara in the Thai alphabet is the nikkhahit, which is used when rendering Sanskrit and Pali texts. It is written as an open circle above the consonant (for example อํ) and its pronunciation depends on the following sound: if it is a consonant then the nikkhahit is pronounced as a homorganic nasal, and if it is at the end of a word it is pronounced as the velar nasal ŋ.
Anunasika (anunāsika) is a form of vowel nasalization, often represented by an anusvara. It is a form of open mouthed nasalization, akin to the nasalization of vowels followed by "n" or "m" in Parisian French. When "n" or "m" follow a vowel, the "n" or "m" becomes silent and causes the preceding vowel to become nasal (pronounced with the soft palate extended downward so as to allow part or all of the air to leave through the nostrils). Anunasika is sometimes called a subdot because of its IAST representation.
In Burmese, the anunasika, called သေးသေးတင် (IPA: [θé ðé tɪ̀ɴ]) and represented as (ံ), creates the /-ɴ/ nasalized ending when it is attached as a dot above a letter. The anunasika represents the -m final in Pali.
- Ohala (1983, p. 90) lists five more such words: dhaṃs- "to sink", phaṃs- "to be stuck", haṃslī "a necklace", haṃsiyā "a sickle" and haṃsī "laughter".
- Monier-Williams 1899.
- Allen 1953, p. 40
- Whitney, cited in Emeneau 1946, p. 91
- Varma 1961, pp. 148–55
- Emeneau 1946, p. 91
- Allen 1953, p. 41
- The following rules are from Ohala (1983, pp. 87–90)
- Chenchiah, P.; Rao, Raja Bhujanga (1988). A History of Telugu Literature. Asian Educational Services. p. 18. ISBN 81-206-0313-3.
- Allen, W.S. (1953), Phonetics in ancient India, OUP
- Emeneau, M. B. (1946). "The Nasal Phonemes of Sanskrit". Language. 22 (2): 86–93. JSTOR 410341. doi:10.2307/410341.
- Monier-Williams, Monier (1899). A Sanskrit-English dictionary : etymologically and philologically arranged with special reference to cognate Indo-European languages. Oxford: Clarendon Press.
- Ohala, Manjari (1983), Aspects of Hindi Phonology, Motilal Banarsidass, ISBN 0-89581-670-9
- Varma, Siddheshwar (1961) . Critical studies in the phonetic observations of Indian grammarians. James G. Forlong Fund. Delhi: Munshi Ram Manohar Lal.