My sister-in-law gave me a bottle of sriracha sauce, which made me happy because I’d been curious to try it ever since I first heard about it (probably within the last couple of years). I once lived in Thailand, after all, and I like spicy food (though my tolerance has diminished over the years). It turned out to be not that spicy, and (as I wrote her) it had an interesting flavor and left a pleasant buzz on my palate.

But we’re not here to talk about sauce but about the word itself, which is very odd indeed. The spelling is misleading; it’s pronounced (per Wikipedia — it’s too new to be in any of my dictionaries) [sǐː rāː.t͡ɕʰāː] in Thai and /sᵻˈrɑːtʃə/ in English, which makes sense, because it’s “named after the coastal city of Si Racha, in Chonburi Province of eastern Thailand, where it may have been first produced for dishes served at local seafood restaurants.” So why is that -r- there? Because it’s from the Sanskrit honorific sri (familiar from Sri Lanka), which winds up in Thai, lacking the consonant cluster /sr/, as /si/. You would think, since the name of the town is written without the -r-, that the name of the sauce could be as well, but for reasons unknown to me it is not, meaning that to the already mind-boggling English repertoire of spelling/pronunciation mismatches is added the new, and unique, sr- pronounced s-. Even I, a fan of irregularities and of unpredictability, think this is absurd.


  1. Ryan Weller says:

    In Thai, loan words from Sanskrit and Pali are spelled to represent the original spelling as closely as possible. The correct pronunciation is then as close as Thai phonology allows based on that spelling. This is the policy of the Royal Institute of Thailand and is taught as a rule in monolingual Thai grammars. It’s also more or less the way English handles loan words. Thus, Sri/Si is spelled with a silent ร in Thai to represent the source orthography, and Roman transliterations will vary based on whether they favor the spoken or written language.

    Note that proper names of Thai people in English-language sources usually adhere to the spelling. For example, in “Bhumibol,” the name of the late king, both the “Bh” and “b” are pronounced as aspirated “p”‘s, while the “l” is pronounced as an “n.”

  2. Ransom says:

    Wikipedia’s omission of the early R in the English pronunciation does not reflect my speech. I don’t trust it to reflect the speech of my hot-sauce-loving peers either. I just read it how it looks on the bottle, and I expect most English speakers do likewise.

    Maybe we should double-check the pronunciation with a straw poll of the Boston Celtics. =)

  3. Wikipedia’s omission of the early R in the English pronunciation does not reflect my speech. I don’t trust it to reflect the speech of my hot-sauce-loving peers either. I just read it how it looks on the bottle, and I expect most English speakers do likewise.

    Ah, that makes sense, except that /sr/ is pretty marginal in English, so I’d expect a certain amount of avoidance. So yeah, straw poll: how do those of you who use the word say it?

  4. Greg Pandatshang says:

    Wikipedia gives Sri Racha as an alternate spelling of the city’s name. One might assume it was once the more common variant.

    Funny, I was just thinking about the customary spelling of the name of the late, lamented King Bhumibol, which matches neither the pronunciation nor the Sanskrit etymons. Unable to wrap my brain around the Thai script, I can only assume that “Bhumibol” and “Sri Racha” are based on transliterations of the native written form. And that, I assume, is based on how the words were once pronounced, much as French orthography represents an intermediate pronunciation between the Vulgar Latin and the modem French. In the Thai case, perhaps the spelling of loanwords is not based on any ordinary register, but on what would be expecteded from an educated person who could handle some foreign Indic sounds.

  5. I’ve also had to explain the sauce’s Thai provenance to my parents (who’ve recently taken a liking to it), because it’s a fixture of Vietnamese restaurants here – and there’s not much of a local Thai community in comparison.

    @Ransom: I would find the /ɹ/ hard to say without retracting the /s/ to /ʃ/ (like I do in Sri Lanka), and maybe adding a secondary stress as well. Ever since learning of the city that it’s named for, I’ve confidently said it without.

  6. anders says:

    I always heard it said as /sᵻˈɹatʃə/ and I hadn’t even noticed the first r in the spelling

  7. kevincure says:

    1) Mostly I am amazed (as a 33 year old) that someone like our host had never had Sriracha – for folks my age, it is about as common as tabasco or salsa or soy sauce. I can’t remember the last time I had an Asian-y rice-based dish *without* it.

    2) I would have pronounced it without the r had I been asked, as would my friends, though I wouldn’t look at a “r” pronunciation with any surprise.

  8. Greg Pandatshang says:

    I say /ˌsriˈɹatʃə/, which I would guess is more common among LH readers than among the general public. I have often noticed other people, the not so linguistically woke element, saying /sᵻˈɹatʃə/, which I assumed is due to /sr/-avoidance. Most people seem to do okay with “Sri Lanka”, however. Ironically, I idiosyncratically got in the habit of saying ʃri in that case, less for /sr/-avoidance than as a hypercorrect Sanskritism.

  9. All of the Vietnamese places around here have it, and a lot of Chinese–noodle joints, barbeque merchants and such. I had it at a Korean restaurant yesterday. Strangely the one place you don’t see it much is Thai restaurants. I have an idea that in Thailand it’s somewhat local to Si Racha. However Mr. Tran, founder of the Huy Fong company, learned his trade in Vietnam.

    Apparently the PTO considers Sriracha a generic name, so Mr. Tran couldn’t get a trademark. That allows Heinz and all kinds of other companies to use the name. I think we’ll stick to the Huy Fong brand though.

    I’ve always heard it called “Sriracha”, but people who are uncomfortable saying “Sri” say “rooster sauce” after the design on the bottle.

    I don’t know how LH managed to avoid it all this time. I’ve been enjoying it for some thirty-odd years. Perhaps it is a regional taste.

  10. Японский городовой says:

    In Russian the siracha : shriracha: sriracha rate is close to 4 : 3 : 2 (according to Yandex), which reflects the English preference to skipping /r/ altogether or at least converting /sr/ to a more familiar /ʃr/ from Sri Lanka despite there’s no such limitation in Russian. Anyhow, most of the articles on top of the search results page mention how funny the word is — and indeed, ‘sri’ is the imperative of a vulgar Russian word meaning ‘to take a shit’ — and you can’t get away from a vulgar connotation no matter what your preference is…

  11. Lucy Kemnitzer says:

    I’ve never heard it without the R and was surprised to read that the normal pronunciation has no R. Also I’ve only heard it with ʃri, never with s.

    And yes, Sriracha is very mild and oddly sweet. When I first tasted it I was disappointed. I suppose I’ve gotten used to it.

  12. Yeah, I don’t think I’ve ever heard it spoken, but mentally I pronounced it with the “r”, same as “Sri Lanka.”

    In Japanese, it’s “shirachaa” (LHH).

  13. In Korean, the normative rendition into the Korean alphabet would be 시라차 siracha, as there is an explicit rule that Thai sr and thr should be rendered as ㅅ s in the Loanword Transcription Rules for Thai.

    However, the sauce seems to be better known as 스리라차 seuriracha based on the spelling pronunciation of the usual romanization.

    Most Koreans are not familiar with Thai pronunciation, and the Loanword Transcription Rules for Thai are even less likely to be followed than similar rules for more well known languages like English. The Thai tennis player Paradorn Srichapan was usually rendered even in Korean newspapers as 파라돈 스리차판 Paradon Seurichapan although it should be 파라돈 시차판 Paradon Sichapan according to the rules.

  14. January First-of-May says:

    I think I have /ʃri/ (or something similar) in “Sri Lanka” and /sri/ in “sriracha”, but that’s mostly because I know the former has a Ш in Russian and I have no idea what the Russian for the latter is (having not encountered it often enough outside of English online contexts).

    As it happens, Russian phonotactics have no problem with initial /sr/, but it happens to appear in a common taboo root, so doesn’t come up very often otherwise (though there are exceptions: средний “average”, среда “environment, Wednesday” – those two are related, by the way – and IIRC a few others).
    As mentioned, I have no idea how the Russian name for the sauce is spelled, but it probably isn’t a straight transliteration from the English because that would just make it a target for bad jokes. My guess (without googling) would be шрирача, but it can easily be something else entirely.

  15. David Marjanović says:

    In my experience, most of it in the US, every single “Asian” restaurant (and then some) has a bottle on every table. Possibly excepting the Thai ones, I can’t remember.

    I hadn’t even noticed the first r in the spelling

    I have no idea how people even do that.

  16. It would never occur to me to put anything from a bottle on my restaurant meal, so I never pay any attention to what might be on the table. (I’ll salt food in some circumstances, as when trying to eat cooked vegetables with no sauce on them.)

  17. Lars (the original one) says:

    Depends on the restaurant — malt vinegar for your (fish and) chips and soy sauce for your sushi are self-serve in the restaurants I eat those things in. I might also take some HP sauce with my hash if it needs it, though now we’re entering lazy cook territory — but attitudes to that noble condiment do vary so it’s often omitted.

    Some fake-Texan burger places here let people put BBQ sauce on everything too, but I think that’s a mistake.

    You have a point that a lot of people labor under the misconception that hot sauce makes sense when poured over a dish — it should have been added at a much earlier stage. At least the fad of adding sliced jalapeños to everything is fading here.

  18. Andrew Dunbar says:

    I haven’t known this word for that many years, butt have always pronounced the /sri/ and I think I’ve heard it pronounced that way by others. It would not have occurred to me to pronounce it with /ʃri/

    Also I can’t recall which restaurants I’ve seen it in and not seen it in, or in which countries. I recall reading its convoluted story recently, which I have since forgotten (-:

    And I’m sure I’ve been confusing it to some degree with salsa Valentina in Mexico. I have a feeling I first learned of Sri Racha sauce when somebody brought a bottle of it to a bar I used to frequent in Mexico and referred to it as “rooster sauce”….

  19. SFReader says:

    Holy King sauce?

  20. @SFReader: Etymologically, yes. It would be श्री राज śrī rāja in Devanagari.

    @Ryan Weller: The correct pronunciation is then as close as Thai phonology allows based on that spelling.

    This is basically correct, but a more accurate explanation would be that the Indic loanwords in Thai went through phonological simplifications long ago, complicated by the fact that different words entered the language in different times and some of these went through the medium of Khmer. Today, there is no simple and consistent rule for figuring out what the Thai pronunciation is based on the Sanskritized official orthography, so that even Thai speakers may be unsure about unfamiliar words.

    Lao orthography is much more phonetic because it makes less of an attempt to capture the Indic etymology, but Khmer orthography also poses lots of problems for the learner because for the Indic loanwords the usual pronunciation rules go out the window.

    @Greg Pandatshang: “Bhumibol” is indeed probably best understood as a transliteration of the Thai spelling where the Indic consonant values are followed. “Sri Racha” could also be understood as a transliteration, but where the usual Thai sound values for the letters are followed—if the Indic values were followed, it would be “Sri Raja”.

  21. I usually say /ʃri/ by association with Sri Lanka, but I hear and accept /sri/ and /sir/. Personally I’ve moved on to Sambal Oelek, also produced by Huy Fong, although I’m not confident of my pronunciation, which is oolek.

  22. @david: Oelek is just the Dutch-style pre-reform spelling for what would be ulek in modern Indonesian spelling. But note that the final e represents a schwa sound; listen to the Indonesian pronunciation of sambal ulek here.

  23. I have no idea how people even do that.

    Anglophones are accustomed from the very beginning of their attempts to read to the idea that some letters are not to be pronounced.

  24. David Marjanović says:

    That doesn’t explain not seeing the letter!

  25. I have been aware of it for a few years, have occasionally heard its name, and have occasionally used the sauce, and even have a bottle in the house now, but I rarely have occasion to say its name. I just assumed the r should be pronounced, but have never been surprised when it wasn’t.

    What is this about hot sauce never “making sense” when poured over a dish?

  26. Lars (the original one) says:

    Either your cook makes the food strong enough that it isn’t needed, or intends flavors that the hot sauce will kill, or you need to take over the cooking.

    Bland food is a crime, but so is refusing to taste good food.

  27. January First-of-May says:

    I have no idea how people even do that.

    I can do worse – I’ve sometimes imagined entirely extraneous letters in words that never had them.
    One definite example that comes to mind is “diactritic” (instead of “diacritic”).

    (Other words I regularly misspell despite all my attempts to the contrary include “relevation”, “recurrest”, “substract” and “geniune”, instead of correct “revelation”, “resurrect”, “subtract” and “genuine” respectively.
    But those can at least be explained by cross-contamination from, respectively, “relevance”, “recurrent”, “substrate” and “genius” – while I can’t think of any plausible explanation for “diactritic”.)

  28. An interesting example of spelling coinage is the Tibetan word for ‘English’, ¯ʔinʥi in Lhasa Tibetan, a loan so modern as to be neither in Jäschke nor Das. This is spelt dbyin.ji., which would in theory correspond to some pronunciation such as *dbjinʥi in “Old Tibetan”. It is unlikely, to say the least, that the Tibetan coiner of this written form ever heard such a pronunciation, yet he chose this complex spelling in preference to the simpler ˝in.ji. which would have been equally valid. Curiously, the Amdo Khake dialect pronunciation of the initial syllable of the word is given as ɣjen- (ɣj- being the usual Amdo reflex of dbj- but not of ˝-). This pronunciation must be derived from a reading or spelling-style pronunciation of the written form, incidentally falsely appearing to legitimise the spelling as that of a genuinely “old” syllable.
    — Philip Denwood. 2007. The language history of Tibetan. In Roland Bielmeier & Felix Haller (eds.), Linguistics of the Himalayas and Beyond


    Amaranth owes its -th spelling (and pronunciation) to folk etymology. But why did it occur to someone to spell anthem, author, authority in this way, and what made educated people adjust their pronunciation to such artificial spellings? Anthony is already going the same way.

  29. I can’t think of any plausible explanation for “diactritic”.

    And yet a Google search reveals that it’s quite a common misspelling. One can also find examples of French diactritique, Polish diaktrytyczny, Russian диактритический, German and Dutch diaktritisch, and even the various declensional forms of Greek διακτριτικός.

  30. “named after the coastal city of Si Racha, in Chonburi Province of eastern Thailand, where it may have been first produced for dishes served at local seafood restaurants.”

    Supposedly it was invented by a Sino-Viet guy after being rescued by fishermen from Sri Racha at sea. he was on a boat escaping Vietnam – one of the Boat People. He named it after their city in gratitude and now people there make it too, not just in Van Nuys, and export it, so that was concrete gratitude.

    You see it everywhere in Vietnamese restaurants in the US, especially the West Coast, because it’s a Vietnamese-American product. I would be very surprised to see it in Vietnam and it’s no mystery why you don’t see it in Thai restaurants here.

    As for pronunciation, the one most common in the Seattle are is “rooster sauce”, based on the label.

  31. Jim: As for pronunciation, the one most common in the Seattle are is “rooster sauce”, based on the label.

    It should be spelt wrorcester.

  32. My day is made.

  33. Piotr: In all these cases, the OED attributes the t > th spelling change, and the following (mostly 18C) spelling pronunciation, to contamination from a specific Greek word: polyanthus for amaranth, authentic for author and friends, *ant(i)-hymn for anthem (which is ultimately < antiphon), and anthos ‘flower’ for Anthony (note tantony pig ‘St. Anthony’s pig, runt pig’ which preserves the older spelling and pronunciation).

  34. For some reason, BrEng Anthony seems to have undergone the spelling change but not the sound change, still taking /t/.

  35. Lars (the original one). “Bland food is a crime…” I do not like bland food, but I am also ‘allergic’ to any pepper or hot spices because my parents did not use them and I so I never got accustomed to them. So I am very sensitive about hot spices and can detect levels that other people hardly notice.

    But over the years I have had meals in eg Michelin 3, 2 and 1 star restaurants, and many others – I lived in France for more than 20 years – with great enjoyment, and you couldn’t call them bland, but there was no overt hot spice taste. And I once ate at Gaylords Indian restaurant on W 59 th, (long gone, I believe) where the menu had starred items that were guaranteed no spice- and it was delicious. In the UK, however, it is assumed that everyone wants hot spicy Indian dinners, which are the most popular meal eaten out. I’m just in a tiny minority, it appears.

  36. Greg Pandatshang says:

    @Jongseong Park, it’s those non-Sanskrit vowel values in bol that have confused me, however.

  37. @Lazar For some reason, BrEng Anthony seems to have undergone the spelling change but not the sound change, still taking /t/.

    Ahem, speaking as an Anthony …

    Antony is a Latin name (Marcus Antonius) with no ‘h’. Some interfering busybody in the C19th thought it ought to have a better pedigree; so invented an etymology from Greek ‘Anthos’ (a flower — for example see

    So the /t/ pronunciation is correct and the inserted ‘h’ is bogus. I am occasionally compelled to give a ‘spelling pronunciation’ with /θ/ for those wanting to write it down. But I resent it.

  38. Greg Pandatshang says:


    Indeed, Wiktionary lists an Old Tibetan pronunciation for དབྱིན་ཇི: Old Tibetan: /*dbʲin.d͡ʑi/, which is ridiculous.

    But I thought Standard Tibetan distinguished a /ji/ initial from a /ʔi/~/i/ initial, so based on the spelling, I’d expect dbyin-ji to be realised as ¯jinʥi rather than ¯ʔinʥi (contra the spelling “inji” as a slang loan back into English). If that’s what was intended, then the prefix is necessary to convey high tone: it was either dbyin-ji or g.yin-ji.

    Another case that comes to mind where dialectical variation makes a hash of an attempted phonetic spelling is “Burma” (as well as “Myanmar”), which I, being an American, pronounce with an r-sound that was never intended,

  39. speedwell says:

    I just read January First-of-May’s post and now I have to lie down with a headache.

    By the way, for me “sriracha” is pronounced exactly as written, with “sree”; “Sri Lanka” and “Sri Ganesh” and similar with “shree”.

  40. SFReader says:

    There is a classical Tibetan description of world geography (Full description of Jambudvipa continent or Dzam-gling rgyas-bshad) compiled in early 19th century. The English are called there Emkiraisi and they are described in quite unflattering terms – “….compared to other Europeans, they are ill-mannered people as they like to drink so much”.

    Author, a prominent and rather high-ranking Tibetan lama Btsan po no mon han thought it useful to describe over 80 European countries, including even painstakingly transcribed list of departments of France and names of European colonies in America.

    Description of North America in his work:

    “If one goes in a northwesterly direction from the country of Isilanti [Iceland]
    crossing a large ocean by boat, there is the large continent called Shewirniya Amirika [< Russ. Severnaya Amerika, “North America”], i.e., Northern Jangling, which has a measurement of three thousand four hundred miles.
    In the northern part of that continent, there is a dé [region] called Su-bi-rayar [Superior],35 and some large dé, such as Mikchingkana [Michigan?], Ku-runa [Huron?], Si-ri-kra-ta, and Sa-ra-sa-sa (?). The people of those regions are large in stature, have light-yellow colored skin, and their facial features are like those of a Mongol. They (wear) coats of the skins of various kinds of beasts of prey, such as the tiger and leopard. They have no religious or social customs whatsoever. They are fierce and savage fools who make their livelihood with livestock, such as cattle and sheep.
    Among the regions mentioned above, Si-ri-kra-ta belongs under the control of the Emkiraisi [English], and the other regions are controlled by the people of various (other) countries.
    In the southern part of that continent, there is the large region called Karlin [Carolina], and there are many countries such as Lutsiyana [Louisiana], Miksika [Mexico], Phelorita [Florida], Kaliphirniya [California], and Yokhadan [Yucatan]."

    And the most touching piece about great man we all love and cherish:

    "At the time when Meparadza, that great learned man who was born in the town of Tsinaba [Genoa] of the country of glorious Shambhala, and whose other name was Kalampatsha [Columbus], i.e., “King of the Boots,” went to Jangling (i.e., the Americas), he first arrived at that island called Sakam [San Salvador Island]".

    I didn't know that Shambhala is actually Italy, but it does make sense….

  41. @Greg Pandatshang: it’s those non-Sanskrit vowel values in bol that have confused me, however.

    Yeah, the vowels on the other hand are following Thai pronunciation, where the inherent vowel is /o/ in closed syllables and you often have truncation of final vowels.

    Indeed, Wiktionary lists an Old Tibetan pronunciation for དབྱིན་ཇི: Old Tibetan: /*dbʲin.d͡ʑi/, which is ridiculous.

    The pronunciation template they use in Wiktionary automatically generates an Old Tibetan pronunciation based on the spelling, but of course this doesn’t make sense if it is a newer coinage not attested in Old Tibetan. I’ve just edited the page to suppress this behaviour for དབྱིན་ཇི.

  42. John: *ant(i)-hymn for anthem

    Anthem is partly a reflex of OE antef(e)n with the assimilation of [vn] > [mn] > [m] that often happened in such words (hræf(e)n ~ hræmn ~ hræm(m) ‘raven’). In a pedantic trisyllabic pronunciation OE speakers may have placed a tertiary stress on the medial syllable (since it was two moras away from the beginning of the word): [`ɑnˌtɛvn̥]. But less formally it was certainly [`ɑntəm], and this is what many ME spellings show: antem, antym etc. Convergence with Anglo-Norman anteme, antenne introduced a variant with stress on the second syllable. The more “careful” pronunciation also survived, reflected in spellings like antefen, antefne, antempne (and antimne, with partial assimilation). Late ME writers who were aware of the antiphone etymology began to introduce learned spellings like antephne, and it seems that there was some hesitation as to where the h properly belonged, since about 1400 we already find anthephne (ignoring Grassmann’s Law) and antheme (still as a rare variant of anteme). I can’t find any spellings sugesting the influence of “hymn”, which remained h-less throughout ME anyway: ymne, ympne, etc. The first attempts to reintroduce the Greek h- in this word date from the end of the 15th c.

  43. Lars (the original one) says:

    @Paul, I use bland as the opposite of well-seasoned, not of ‘hot’ — that would be ‘mild’ I think. And you rally behind my point, all the meals you describe would have been ruined by the application of hot sauce.

  44. David Marjanović says:

    It should be spelt wrorcester.

    I needed this today.

  45. “It should be spelt wrorcester.”

    No, that’s “what’shtishere sauce”.

    Speaking of spices and blandness and seasonings, apparently fish sauce was an item of trade out of Vietnam – “Annam” – into Britain in the early 19th(?) century. I can only wonder….

  46. The superfluous R in “Sriracha” reminds me of another curious spelling: that of the Indian state formerly known as “Orissa”. It insisted on changing its official English spelling to “Odisha” in 2010, even though the state’s name is pronounced with /s/ in the local language (which was also renamed, from “Oriya” to “Odiya”). The spelling in the local script uses a character historically associated with /ʃ/, but all sibilants have been merged to /s/ in the Oriya/Odiya language.

  47. The new name for the language is actually ‘Odia’, not ‘Odiya’. The r in ‘Orissa’ and ‘Oriya’ reflects the intervocalic flap allophone [ɽ] of the /ɖ/ that is rendered as d in the new official spellings ‘Odisha’ and ‘Odia’.

    I tried to argue that there was no need to change the officially approved Korean name for Orissa/Odisha, which was 오리사 Orisa, since this was just a spelling change with no difference in the pronunciation in the original language. But the language authorities went ahead and approved 오디샤 Odisya as the new name for the state.

  48. That’s very weird (both the original change to a nonpronounced “sh” and the Korean change reflecting it).

  49. I think it’s about switching from transcription to straight transliteration. In the first case, it is about suppressing an allophonic distinction not represented in writing: in the second, it’s about adopting a written-only distinction that does not reflect either a phonemic or an allophonic distinction.

Speak Your Mind