Ġ
This article needs additional citations for verification. (January 2023) |
Ġ | ġ |
G̣ | g̣ |
Ġ (
above the letter.Usage
Arabic
Ġ is used in some
Armenian
Ġ is used in the romanization of Classical or Eastern Armenian to represent the letter Ղ/ղ (ġat).
Chechen
Ġ is present in the
Inupiat
Ġ is used in some dialects of
Irish
Ġ was formerly used in Irish to represent the lenited form of G. The digraph gh is now used.[2]
Maltese
Ġ is the 7th letter of the Maltese alphabet, preceded by F and followed by G. It represents the voiced postalveolar affricate [dʒ].[3]
Old Czech
⟨ġ⟩ is sometimes (about 16th century) used to represent real [g], to distinguish it from the letter g which represented the consonant [j].
Old English
⟨Ġ⟩ is sometimes used in modern scholarly transcripts of Old English to represent [j] or [dʒ] (after ⟨n⟩), to distinguish it from ⟨g⟩ pronounced as /ɣ/, which is otherwise spelled identically. The digraph ⟨cg⟩ was also used to represent [dʒ].[4]
Ukrainian
⟨Ġ⟩ is used in some
Phonetic transcription
⟨ġ⟩ is sometimes used as a phonetic symbol transcribing [ɣ] or [ŋ].
Georgian
Ġ is used in the transliteration of Georgian to represent the letter ღ.
Computer encoding
ISO 8859-3 (Latin-3) includes Ġ at D5 and ġ at F5 for use in Maltese, and ISO 8859-14 (Latin-8) includes Ġ at B2 and ġ at B3 for use in Irish.
Appearance | Code points | Name |
---|---|---|
Ġ | U+0120 U+0047, U+0307 |
LATIN CAPITAL LETTER G WITH DOT ABOVE LATIN CAPITAL LETTER G + COMBINING DOT ABOVE |
ġ | U+0121 U+0067, U+0307 |
LATIN SMALL LETTER G WITH DOT ABOVE LATIN SMALL LETTER G + COMBINING DOT ABOVE |
OpenAI's GPT-2 uses 0xC4 0xA0 (Ġ) as the start of a word in its tokens.[5]
References
- ^ Koryakov, Yuri B. (2002). Atlas of Caucasian Languages (PDF). Moscow: Institute of Linguistics, Russian Academy of Sciences. pp. 6–7.
- ^ "Symbol Codes | Irish, Old Irish and Manx". Pennsylvania State University. 21 April 2017. Retrieved 10 January 2023.
- ISBN 978-1-57506-109-2. Retrieved 10 January 2023.
- ^ Daniel Paul O'Donnell. "The Pronunciation of Old English". University of Lethbridge Personal Web Sites. Retrieved 26 October 2022.
- ^ "Why \u0120 (Ġ) is in so many pairs? · Issue #80 · openai/GPT-2". GitHub.