In Sanskrit phonology, visarga () is the name of the voiceless glottal fricative, written in Devanagari as . It was also called, equivalently, ' by earlier grammarians. The word visarga () literally means "sending forth, discharge".
Visarga is an allophone of and in pausa (at the end of an utterance). Since is a common inflectional suffix (of nominative singular, second person singular, etc.), visarga appears frequently in Sanskrit texts. In the traditional order of Sanskrit sounds, visarga and anusvÃÂra appear between vowels and stop consonants.
The precise pronunciation of visarga in Vedic texts may vary between à ÂÃÂkhÃÂs. Some pronounce a slight echo of the preceding vowel after the aspiration: will be pronounced , and will be pronounced . Visarga is not to be confused with colon.
The visarga is commonly found in writing, resembling the punctuation mark of colon or as two tiny circles one above the other. This form is retained by most Indic scripts.
According to Sanskrit phonologists, the visarga has two optional allophones: the (, the velar visarga) and the (, the labial visarga). The former may be pronounced before and , while the latter may be pronounced before and . Generally, and are not rendered distinctively from the visarga, though the glyphs and have occasionally been used in Devanagari for that purpose; sometimes, the (, ) is used for both. Similar glyphs are found in Brahmi, Kannada, Tibetan, Sharada, and Lantsa. Examples:
In the Bengali script, the visarga is written as and is known as bisarga (à ¦¬à ¦¿à ¦¸à ¦°à §Âà ¦Â). In Bengali it represents a weak post-vocalic release rather than a clearly articulated /h/ sound. Although it originates from Sanskrit final r- and s-sounds, modern Bengali preserves it primarily as an orthographic element, with its phonetic value varying according to position and environment.
When occurring medially within a word, the bisarga no longer produces a distinct aspiration; instead it tends to reinforce or lengthen the following consonant. Examples include or , where the bisarga historically reflects Sanskrit , but in Bengali the release is replaced by consonant strengthening or an increased sibilant effect.
When occurring finally at the end of a word, the bisarga marks an etymological aspiration inherited from Sanskrit, but in contemporary Bengali it is realised only as a very faint breath or is silent. Forms such as or illustrate this weakened contrast. Overall, Bengali pronunciation treats the visarga not as an independent consonant but as a modifier whose audible effect ranges from minimal to null.
In sandhi environments, the bisarga undergoes several regular transformations that depend on the nature of the following sound. These changes reflect its Sanskritic origins and are largely systematic in Bengali spelling, even when their phonetic effect is weak.
Before vowels, the bisarga commonly yields or produces an -glide, preserving the historical continuity of r-final forms (e.g. ). Before sibilants and coronal or palatal stops, it assimilates to the following consonant as the appropriate sibilantâÂÂ, , or âÂÂa process that accounts for forms such as .
When followed by voiced consonants or semivowels, r-jÃÂt bisarga typically appears as , while s-jÃÂt bisarga may produce an -transition; thus surfaces as . If the following consonant is , the bisarga disappears entirely and the preceding vowel lengthens, as seen in . After the vowels à ¦ or à ¦ and before velar or labial stops, the bisarga may appear as , producing forms such as .
Some compounds retain the written bisarga without any phonological modification, as in . In other cases, the bisarga may be lost without affecting adjacent sounds, as in from . As a whole, visarga sandhi in Bengali reflects historical Sanskritic patterns that are preserved orthographically even though their phonetic realisation is limited in modern speech.
In the Burmese script, the visarga (variously called shay ga pauk, wizza nalone pauk, or shay zi and represented with two dots to the right of the letter as ), when joined to a letter, creates the high tone.
Motoori Norinaga invented a mark for visarga which he used in a book about Indian orthography.
In the Javanese script, the visarga, known as the wignyan (), is represented by two curls to the right of a syllable as : the first curl is short and circular, and the second curl is long. It adds a /-h/ after a vowel.
In the Kannada script, the visarga (which is called visarga) is represented with two small circles to the right of a letter à ²Â. It adds an aḥ sound to the end of the letter.
This script also has separate symbols for ardhavisarga absent in most other scripts, jihvamuliya, , and upadhmaniya, .
In the Khmer script, the visarga (known as the reÃÂhmà Âkh (; "shining face")) indicates an aspirated sound added after a syllable. It is represented with two small circles at the right of a letter as , and it should not be confused with the similar-looking yà ÂkôlpÃÂntà  (; "pair of dots"), which indicates a short vowel followed by a glottal stop like their equivalent visarga marks in the Thai and Lao scripts.
In the Lao script, the visarga is represented with two small curled circles to the right of a letter as . As in the neighboring related Thai script, it indicates a glottal stop after the vowel.
In the Malayalam script, the visarga is represented with two small circles to the right of a letter as . It indicates a /h/ after a letter. Visarga is considered as a vowel in Malayalam, where its independent form is written as à ´ à ´Â. Unlike other languages, visarga need not necessarily occur at the end of a word. Examples include à ´¦à µÂà ´Âà ´Âà ´Â, à ´®à ´¨à ´Âà ´ªà µÂà ´°à ´¯à ´¾à ´¸à ´Â, à ´ªà µÂà ´¨à ´Âà ´¸à µÂà ´·à µÂà ´Âà ´¿.
In the Odia script, the Bisarga is represented with a vertical infinity sign to the right of a letter as . It indicates the post-vocalic voiceless glottal fricative aḥ [h] sound after the letter, Unlike other languages, the bisarga can appear in middle of words, such as à ¬¨à ¬¿à ¬Âà ¬¶à ÂÂà ¬¬à ¬¾à ¬¸, à ¬¨à ¬¿à ¬Âà ¬¸à ÂÂà ¬¬, à ¬¨à ¬¿à ¬Âà ¬¸à ¬¨à ÂÂà ¬¦à ÂÂà ¬¹, à ¬¨à ¬¿à ¬Âà ¬¶à ÂÂà ¬· etc. All words with Bisarga aren't borrowed from Sanskrit.
In the Sinhala script, visarga is represented with two small circle to the right of a letter as à ¶Â.
In the Tamil script, similar to visarga (which is called ÃÂyuta eḻuttu (), ÃÂytam (), muppÃÂl puḷḷi, taá¹Âinilai, aḵkÃÂá¹Âam, ak), transliterated as ḵ, is represented with three small circles to the right of a letter as . Its used to transcribe an archaic sound inherited from the proto Dravidian *H that has either become silent or geminates the next letter in unlearnt speech, or pronounced as or in careful speech. Like Sanskrit, it cannot add on to any letter and add aspiration to them. It should be always placed between a single short vowel and a hard consonant (, , , , , ) for example (aḵtu), (eḵku). The ÃÂytam in modern Tamil is used to transcribe foreign phones like à ®Âà ®ªà ¯ (à ®Âp) for [f], à ®Âà ® (à ®Âj) for [z], à ®Âà ®¸ (à ®Âs) for [z, ÃÂ] and à ®Âà ® (à ®Âk) for [x], similar to a nuqta.
In the Telugu script, there are two visargas. One is represented with two small circles to the right of a letter . It brings an "ah" sound to the end of the letter.
In the Thai script, the visarga (known as the visanchani () or nom nang thangkhu ()) is represented with two small curled circles to the right of a letter as . It represents a glottal stop that follows the affected vowel.
Unicode encodes visarga and visarga-like characters for a variety of scripts: