Voiceless velar fricative
IPA Number140
Audio sample
Entity (decimal)x
Unicode (hex)U+0078
Braille⠭ (braille pattern dots-1346)

The voiceless velar fricative is a type of consonantal sound used in some spoken languages. It was part of the consonant inventory of Old English and can still be found in some dialects of English, most notably in Scottish English, e.g. in loch, broch or saugh (willow).

The symbol in the International Phonetic Alphabet that represents this sound is ⟨x⟩, the Latin letter x. It is also used in broad transcription instead of the symbol ⟨χ⟩, the Greek chi, for the voiceless uvular fricative.

There is also a voiceless post-velar fricative (also called pre-uvular) in some languages, which can be transcribed as [x̠] or [χ̟]. For voiceless pre-velar fricative (also called post-palatal), see voiceless palatal fricative.

Some scholars also posit the voiceless velar approximant distinct from the fricative, used in some spoken languages. The symbol in the International Phonetic Alphabet that represents this sound is ⟨ɰ̊⟩, but this symbol is not suitable in case of the voiceless velar approximant that is unspecified for rounding (the sound represented by the symbol ⟨ɰ̊⟩ is specified as unrounded), which is best transcribed as ⟨⟩, ⟨ɣ̞̊⟩ or ⟨ɣ̊˕⟩ - see voiced velar approximant. The velar approximant can in many cases be considered the semivocalic equivalent of the voiceless variant of the close back unrounded vowelɯ̊⟩.


Features of the voiceless velar fricative:


IPA Description
x plain velar fricative
xʷʼ ejective labialised
x̜ʷ semi-labialised
x̹ʷ strongly labialised
xʲʼ ejective palatalised


The voiceless velar fricative and its labialized variety are postulated to have occurred in Proto-Germanic, the ancestor of the Germanic languages, as the reflex of the Proto-Indo-European voiceless palatal and velar stops and the labialized voiceless velar stop. Thus Proto-Indo-European *r̥nom "horn" and *ód "what" became Proto-Germanic *hurnan and *hwat, where *h and *hw were likely [x] and [xʷ]. This sound change is part of Grimm's law.

In Modern Greek, the voiceless velar fricative (with its allophone, the voiceless palatal fricative [ç], occurring before front vowels) originated from the Ancient Greek voiceless aspirated stop /kʰ/ in a sound change that lenited Greek aspirated stops into fricatives.

Language Word IPA Meaning Notes
Abaza хьзы /xzë [xʲzə] 'name'
Adyghe хы /xë [xəː] 'six'
Afrikaans groot [χrʊət] 'big'
Albanian gjuha [ɟuxɑ] 'language' Allophone of /h/. See Albanian phonology
Aleut Atkan dialect alax [ɑlɑx] 'two'
Arabic Modern Standard ﻀراء [xadˤraːʔ] 'green' (f.) May be velar, post-velar or uvular, depending on dialect.[1] See Arabic phonology
Armenian խրոխտ/xëroxt [χəˈɾoχt] 'brave'
Assamese মীয়া/ohomia [ɔxɔmia] 'Assamese'
Assyrian ܚܡܫܐ xemša [xεmʃa] 'five'
Avar чeхь / čex [tʃex] 'belly'
Azerbaijani x / хош/ﻮش [xoʃ] 'pleasant'
Basque Some speakers[2] jan [xän] 'to eat' Either velar or post-velar.[2] For other speakers it's [j ~ ʝ ~ ɟ].[3]
Brahui[4] [xan] 'eye' Corresponds to /x/ in Kurukh and /q/ in Malto.
Breton hor c'hi [hor xiː] 'our dog'
Bulgarian тихо / tiho [ˈt̪ixo] 'quietly' Described as having "only slight friction" ([x̞]).[5]
Catalan kharja [ˈxɑɾ(d)ʑɐ] 'kharjah' Found in loanwords and interjections. See Catalan phonology
Chechen хан / xan [xɑːn] 'time'
Chinese Mandarin / hé [xɤ˧˥] 'river' See Standard Chinese phonology
Czech chlap [xlap] 'guy' See Czech phonology
Danish Southern Jutlandic kage [ˈkʰaːx] 'cake' See Sønderjysk dialect
Dutch Standard Belgian[6][7] acht [ɑxt] 'eight' May be post-palatal [ç̠] instead. In dialects spoken above the rivers Rhine, Meuse and Waal the corresponding sound is a postvelar-uvular fricative trill [ʀ̝̊˖].[7] See Dutch phonology and Hard and soft G in Dutch
Southern Netherlands accents[7][8]
English Scottish loch [ɫɔx] 'loch' Younger speakers may merge this sound with /k/.[9][10] See Scottish English phonology
Irish lough [lɑx] 'lough' Occurs only in Gaelic borrowings. See Irish English phonology
Scouse[11] book [bʉːx] 'book' A syllable-final allophone of /k/ (lenition).
Esperanto monaĥo [moˈnaxo] 'monk' See Esperanto phonology
Estonian jah [jɑx] 'yes' Allophone of /h/. See Estonian phonology
Eyak duxł [tʊxɬ] 'traps'
Finnish kahvi [ˈkɑxʋi] 'coffee' Allophone of /h/. See Finnish phonology
French jota [xɔta] 'jota' Occurs only in loanwords (from Spanish, Arabic, Chinese, etc.). See French phonology
Georgian[12] ჯო / joxi [ˈdʒɔxi] 'stick'
German Buch [buːx] 'book' See Standard German phonology
Greek τέχνη / ch [ˈte̞xni] 'art' See Modern Greek phonology
Hebrew Biblical מִיכָאֵל/Michael [mixaʔel] 'Michael' See Biblical Hebrew phonology
Hindustani Hindi ख़ुशी/xuší [xʊʃiː] 'happiness' Occurs only in loanwords. May be replaced in Hindi with /kʰ/. See Hindustani phonology
Urdu ﻮشی/xuşi
Hungarian sahhal [ʃɒxːɒl] 'with a shah' See Hungarian phonology
Icelandic október [ˈɔxtoːupɛr̥] 'October' See Icelandic phonology
Indonesian khas [xas] 'typical' Occurs in Arabic loanwords. Often pronounced as [h] or [k] by some Indonesians. See Indonesian phonology
Irish deoch [dʲɔ̝̈x] 'drink' See Irish phonology
Japanese マッハ / mahha [maxːa] 'Mach' Allophone of /h/.[13] See Japanese phonology
Kabardian хы / khy [xəː] 'sea'
Kazakh ханзада / hanzada [xanzada] 'prince'
Korean 흥정 / hûńjóń [xɯŋd͡ʑʌ̹ŋ] 'bargaining' Allophone of /h/ before /ɯ/. See Korean phonology
Kurdish xanî [xɑːˈniː] 'house' See Kurdish phonology
Kurukh[14] कुड़ुख़ [kuɽux] 'Kurukh' Corresponds to /x/ in Brahui and /q/ in Malto.
Limburgish[15][16] loch [lɔx] 'air' The example word is from the Maastrichtian dialect. See Maastrichtian dialect phonology and Hard and soft G in Dutch
Lishan Didan Urmi Dialect חלבא / xalwa [xalwɑ] 'milk' Generally post-velar
Lithuanian choras [ˈxɔrɐs̪] 'choir' Occurs only in loanwords (usually international words)
Lojban xatra [xatra] 'letter'
Macedonian Охрид / Ohrid [ˈɔxrit] 'Ohrid' See Macedonian phonology
Malay اير / akhir [axir] 'last', 'end' Occurs in Arabic loanwords. Often pronounced as [h] or [k]. See Malay phonology
Manx aashagh [ˈɛːʒax] 'easy'
Nepali आँखा/axa [ä̃xä] 'eye' Allophone of /kʰ/. See Nepali phonology
Norwegian Urban East[17] hat [xɑːt] 'hate' Possible allophone of /h/ near back vowels; can be voiced [ɣ] between two voiced sounds.[17] See Norwegian phonology
Norwegian Brekke dialect[18] seg [sɛɰ̊] 'oneself' Also described as an approximant. Coda allophone of /ɡ/; reported to occur only in this word.[18] See Norwegian phonology
Kaldfarnes dialect[18]
Sørkjosen dialect[18] [sæɰ̊]
Undheim dialect[18] [seɰ̊]
Bryne dialect[18] sterkaste,
[ˈstæɰ̊kɑstə] 'strongest' Also described as an approximant. Allophone of /r/ when it is in contact with voiceless consonants. Exact distribution may differ between dialects. In the Hafrsfjord dialect, [ɰ̊] may also occur in the word seg. See Norwegian phonology
Hafrsfjord dialect[18]
Raundalen dialect[18]
Stanghelle dialect[18]
Fyllingsdalen dialect[18] [ˈstæɰ̊kestɛ]
Pashto اخته / axta [ax.t̪a] 'occupied' See Pashto phonology
Persian دُخـتَر / doxtär [dox'tær] 'daughter' See Persian phonology
Polish[19] chleb [xlɛp] 'bread' Also (in great majority of dialects) represented orthographically by ⟨h⟩. See Polish phonology
Portuguese Fluminense arte [ˈaxtɕi] 'art' In free variation with [χ], [ʁ], [ħ] and [h] before voiceless consonants
General Brazilian[20] rosa [ˈxɔzɐ] 'rose' Some dialects. An allophone of /ʁ/. See Portuguese phonology
Punjabi Gurmukhi ਖ਼ਬਰ/xabar [xəbəɾ] 'news'
Shahmukhi ﺒر/xabar
Romanian hram [xräm] 'patronal feast of a church' Allophone of /h/. See Romanian phonology
Russian[21] хороший / horošij [xɐˈr̠ʷo̞ʂɨ̞j] 'good' See Russian phonology
Scottish Gaelic[22] drochaid [ˈt̪ɾɔxɪtʲ] 'bridge' See Scottish Gaelic phonology
Serbo-Croatian храст / hrast [xrâːst] 'oak' See Serbo-Croatian phonology
Slovak chlap [xɫäp] 'guy'
Slovene Standard pohlep [poˈxlɛ̂p] 'greed' See Slovene phonology
Some dialects bog [ˈbôːx] 'god' Allophone of /ɣ/ before voiceless obstruents or pause. See Slovene phonology
Somali khad [xad] 'ink' Also occurs allophone of /q/ in Arabic loan words. See Somali phonology
Spanish[23] Latin American[24] ojo [ˈo̞xo̞] 'eye' May be glottal instead;[24] in northern and central Spain it is often post-velar[24][25][26] or uvular /χ/.[26][27] See Spanish phonology
Southern Spain[24]
Spanish Standard European[28] Predrag [ˈpɾe̞ð̞ɾäɣ̞̊] 'Predrag' Also described as an approximant. Allophone of /ɡ/ before a pause.[28] See Spanish phonology
Sylheti ꠛꠞ/xobor [xɔ́bɔɾ] 'news'
Tagalog bakit [baxit] 'why' Allophone of /k/ in intervocalic positions. See Tagalog phonology
Toda[29] pax [pax] 'smoke'
Turkish[30] ıhlamur [ɯxlamuɾ] 'linden' Allophone of /h/.[30] See Turkish phonology
Turkmen hile [xiːle] 'cunning' (noun)
Tyap kham [xam] 1. 'calabash'; 2. 'prostitute'
Xhosa rhoxisa [xɔkǁiːsa] 'to cancel'
Ukrainian хлопець / hlopeć [ˈxɫɔ̝pɛt͡sʲ] 'boy' See Ukrainian phonology
Uzbek[31] oxirgi [ɒxirgi] 'last' Post-velar. Occurs in environments different from word-initially and pre-consonantally, otherwise it is pre-velar.[31]
Welsh chwaer [χwai̯r] 'sister'
Vietnamese[32] không [xəwŋ͡m˧] 'no', 'not', 'zero' See Vietnamese phonology
Yaghan xan [xan] 'here'
Yi / he [xɤ˧] 'good'
Zapotec Tilquiapan[33] mejor [mɘxoɾ] 'better' Used primarily in loanwords from Spanish

See also


  1. ^ Watson (2002), pp. 17, 19–20, 35–36 and 38.
  2. ^ a b Hualde & Ortiz de Urbina (2003), pp. 16 and 26.
  3. ^ Hualde & Ortiz de Urbina (2003), p. 16.
  4. ^ Bhadriraju Krishnamurti (2003), p. 100.
  5. ^ Ternes, Elmer; Vladimirova-Buhtz, Tatjana (1999). "Bulgarian". Handbook of the International Phonetic Association. Cambridge University Press. p. 55. ISBN 978-0-521-63751-0.
  6. ^ Verhoeven (2005:243)
  7. ^ a b c Collins & Mees (2003:191)
  8. ^ Gussenhoven (1999:74)
  9. ^ "Annexe 4: Linguistic Variables". Archived from the original on 2021-02-25. Retrieved 2014-12-03.
  10. ^ "University of Essex :: Department of Language and Linguistics :: Welcome". Essex.ac.uk. Retrieved 2013-08-01.
  11. ^ Wells (1982:373)
  12. ^ Shosted & Chikovani (2006), p. 255.
  13. ^ Okada, Hideo (December 1991). "Japanese". Journal of the International Phonetic Association. 21 (2): 94–96. doi:10.1017/S002510030000445X. S2CID 242782215. Retrieved 14 July 2022.
  14. ^ Bhadriraju Krishnamurti (2003), p. 74.
  15. ^ Gussenhoven & Aarts (1999:159)
  16. ^ Peters (2006:119)
  17. ^ a b Vanvik (1979), p. 40.
  18. ^ a b c d e f g h i j "Nordavinden og sola: Opptak og transkripsjoner av norske dialekter". Archived from the original on 4 March 2016. Retrieved 13 April 2015.
  19. ^ Jassem (2003), p. 103.
  20. ^ Barbosa & Albano (2004), pp. 5–6.
  21. ^ Padgett (2003), p. 42.
  22. ^ Oftedal, M. (1956) The Gaelic of Leurbost. Oslo. Norsk Tidskrift for Sprogvidenskap.
  23. ^ Martínez-Celdrán, Fernández-Planas & Carrera-Sabaté (2003), p. 255.
  24. ^ a b c d Chen (2007), p. 13.
  25. ^ Hamond (2001:?), cited in Scipione & Sayahi (2005:128)
  26. ^ a b Lyons (1981), p. 76.
  27. ^ Harris & Vincent (1988), p. 83.
  28. ^ a b Wetzels & Mascaró (2001), p. 224.
  29. ^ Bhadriraju Krishnamurti (2003), p. 149.
  30. ^ a b Göksel & Kerslake (2005:6)
  31. ^ a b Sjoberg (1963), pp. 11–12.
  32. ^ Thompson (1959), pp. 458–461.
  33. ^ Merrill (2008), p. 109.