Virama ( ्, (see #Names for other terms) is a Sanskrit phonological concept to suppress the inherent vowel that otherwise occurs with every consonant letter, commonly used as a generic term for a codepoint in Unicode, representing either halanta, hasanta or explicit virāma, a diacritic in many Brahmic scripts, including the Devanagari and Bengali scripts, or saṃyuktākṣara (Sanskrit: संयुक्ताक्षर) or implicit virama, a conjunct consonant or ligature. Unicode schemes of scripts writing Mainland Southeast Asia languages, such as that of Burmese script and of Tibetan script, generally do not gro
via Wikipedia infobox
Virama ( ्, (see #Names for other terms) is a Sanskrit phonological concept to suppress the inherent vowel that otherwise occurs with every consonant letter, commonly used as a generic term for a codepoint in Unicode, representing either halanta, hasanta or explicit virāma, a diacritic in many Brahmic scripts, including the Devanagari and Bengali scripts, or saṃyuktākṣara (Sanskrit: संयुक्ताक्षर) or implicit virama, a conjunct consonant or ligature. Unicode schemes of scripts writing Mainland Southeast Asia languages, such as that of Burmese script and of Tibetan script, generally do not group the two functions together.
== Names == The name is Sanskrit for "cessation, termination, end". As a Sanskrit word, it is used in place of several language-specific terms, such as: {| class="wikitable sortable" !Name in English books !Language !In native language !Form !Notes |- | rowspan="1" |halant |Hindi | |् | |- | rowspan="5" |halanta |Punjabi | |੍ | |- |Marathi | |् | |- |Nepali | |् | |- |Odia | |୍ | |- |Gujarati | |્ | |- | rowspan="3" |hosonto |Bengali | |্ | |- |Assamese | / |্ | |- |Sylheti | | ◌ ꠆ | |- |pollu |Telugu | |్ | |- |pulli |Tamil | |் | |- |chandrakkala |Malayalam | / |് |Unlike other virama diacritics, it is pronounced word-finally. |- |ardhakshara chihne |Kannada | / |್ | |- |hal kirima |Sinhalese | |් | |- |a that |Burmese | |် |lit. "nonexistence" |- |viream | rowspan="2" | Khmer | | ៑ | |- |toandeakheat | | ៍ | |- |karan, thanthakhat | rowspan="3" | Thai | / |◌์ |Thanthakhat is the name of the diacritic, while karan refers to the character that was marked. These two terms are often used interchangeably. It is used to mark as silent vowels or consonants that were originally pronounced, but have become silenced in Thai pronunciation (mostly from Sanskrit and Old Khmer). This diacritic is sometimes used in loanwords from European languages to mark final consonants in consonant clusters (e.g. want as วอนท์). |- |pinthu | |◌ฺ |Pinthu is akin to Sanskrit bindu, and means "point" or "dot". It is used to mark a syllable as closed, and it is only used in Thai script when writing Pali or Sanskrit. |- |nikkhahit |นฤคหิต / นิคหิต |◌ํ |Nikkhahit represents what was originally anusvāra in Sanskrit. Like pinthu, it is also only used when writing Pali or Sanskrit in Thai script. It marks a syllable as nasalized, realized in Thai as a nasal closed consonant following the vowel. |- | rowspan="3" |rahaam |Northern Thai (Lanna) | rowspan="3" | |◌᩺ | |- |Tai Khün |◌᩼ | |- |Tai Lue |◌᩼ | |- |wirama |Kawi | |◌𑽁 | |- |pangkon |Javanese | |◌꧀ | |- |adeg-adeg |Balinese | |◌᭄ | |- | rowspan="3" |pangolat |Mandailing | rowspan="3" | | rowspan="3" |◌᯲ | |- |Pakpak | |- |Toba | |- |penengen |Karo | | rowspan="2" |◌᯳ | |- |panongonan |Simalungun | | |- |pamaeh |Sundanese | |◌᮪ | |- |bunuhan |Rejang | |꥓ | |- |sukun |Dhivehi | | ް◌ | Derives from Arabic "sukun" |- |Srog med |Tibetan |Srog med |྄ |Only used when transcribing Sanskrit |}
Discovered by embedding cosine similarity (sentence-transformers MiniLM, 384-dim).