Time for another Languagehat Poll! There are two English words cist, both ultimately from Latin cista ‘basket’; one, meaning “A wicker receptacle used in ancient Rome for carrying sacred utensils in a procession,” is directly from Latin, the other, “A stone-lined grave, especially a tomb consisting of a pit lined with stones and often having a lid of stone or wood,” comes to us via Welsh, specifically the phrase cist faen ‘stone chest’ (where the second word is the lenited form of maen ‘stone,’ which is apparently related to Old Irish mag ‘field,’ and if you’ve ever seen a field in the west of Ireland you won’t be surprised at that semantic shift). Now, in Welsh the letter c is pronounced /k/, and the second word (‘stone-lined grave’) can be pronounced /kist/ as in Welsh, but the first pronunciation given in dictionaries is the /sist/ you’d expect from the spelling in English. I’m torn as to which pronunciation to mentally adopt as I read the manuscript on archaeology I’m editing; on the one hand, /kist/ is etymologically accurate and distinguishes it from its homograph, but on the other hand it’s very unintuitive and misleading to anyone not familiar with the word (not to mention that it sounds like kissed). But if it’s the pronunciation used by actual archaeologists (and others who use the word in speech), I’m happy to go along with them. Anybody know?


  1. Both and AHD5 both provide both /kɪst/ and /sɪst/ as valid pronunciations of the second cist, but ODO and Collins list only /sɪst/. So it looks like this is another of those places where Americans sometimes (if not always) use the foreign pronunciation.

  2. Oops, saved too soon. The OED2 (1889) lists both pronunciations as well, and gives the alternative spelling kist. The second sense is the older in English, landing in 1804, the first sense not until 1847.

  3. The UK TV archaeology programme ‘Time Team’ say ‘kissed’ and I think they are reasonably academically respectable (within the limits of TV of course).

  4. That’s good enough for me — “kissed” it is!

  5. Oops again: one both too many. I hate capitalizing domain names, but then I hate starting sentences with a lowercase letter.

  6. Fun fact: in Swedish, the word for coffin or chest is actually “kista”, very similar to the Welsh word. The initial consonant is neither /s/ nor /k/, it’s one of them fricative sibilants /ʃ/, like the “sh” in “ship”.

    As for how you would pronounce it in English, well, I think /kist/ sounds nicer than /sist/ (after all, “kissed” is a much more pleasant word than “cyst”), but I’d probably go with /sist/. After all, the closest English relative to the word is probably “cistern”, right? And you pronounce that with a /s/.

  7. In the regional dialect of the province of Wielkopolska, where I live, and of Silesia, kista means ‘box, chest, trunk’. It is of course a loan from German (Kiste, cognate to English chest, borrowed from Latin before the palatalisation of /k/). It’s perhaps worth pointing out that not only cist₁ and cist₂, but also chest, have the same origin (starting with Greek → Latin), only different historical trajectories:

    (1) Latin → Modern English
    (2) Latin → Brittonic >…> Welsh → Modern English
    (3) Latin → Common Germanic >…> Old English >…> Modern English

  8. Offhand, I cannot think of a single English word beginning with [ci-] in which the [c] is pronounced /k/. However, there are many, many that are pronounced /s/.

    There are a number of English [c+other vowels-] that are pronounced /k/, ‘cook, corn, cavern, cave, Celtic (except in basketball teams), cub, cuddle,’ etc.

  9. Trond Engen says:

    Same as Norwegian and Danish kiste. The opposition between the fricatives resulting from k- and sk- is maintained in Norwegian (but lost in some western dialects and among many younger Eastern speakers). Swedes (and I presume Norwegians who have the merger) tend to carry this into their foreign languages, hence Swedish insistence that the German ich-sound and the English sh-sound are the same (which of course is close enough for some German dialects but not for the standard language).

    Also English chest and German Kiste. The shift from “basket” to “coffin, box” seems to be shared by Celtic and Germanic.

  10. Trond Engen says:

    And we can see that it took me more than 15 minutes to hack down those few lines.

  11. ‘Kist’ in Scots is not only a noun (chest, box, case etc), but also a verb which means, among other things, “put or enclose in a coffin”.

  12. The shift from ‘woven basket’ to ‘container, casket’ took place already in Latin (or Etruscan, if it mediated between Greek and Latin). A typical Etruscan cista looked like this:

    Already in Greek, kístē had secondary meanings like ‘writing-case’ and ‘voting-urn’ (LSJ).

  13. Indeed, a search of 30,000 words with British and American pronunciations finds none with ci- pronounced /kɪ/.

  14. Needless to say, that does not deter me; I like outliers.

  15. It sometimes happensstem-finally. There are neo-Latin pedants who pronounce foci and loci to rhyme with okey-dokey. Familiarity with Gaelic and Welsh proper names (Ciarán, Cilcennin) also provides some precedent for a “hard” pronunciation of ci, at least in Britain.

  16. “Indeed, a search of 30,000 words with British and American pronunciations finds none with ci- pronounced /kɪ/.”

    Maybe there is some English aversion to [c] + certain front vowels. Other than ‘Celt,’ I can’t think of any [ce-]s either.

  17. David Eddyshaw says:

    The Welsh spelling with c before i e y is in fact modern; Middle Welsh used k.
    The change is due to the introduction of printing: the printers, tooled up as they were for printing English, didn’t have enough k’s to cope with traditional Welsh orthography, so they used c throughout instead.

  18. GeorgeW, there’s soccer and the British spelling sceptic. Also arced, arcing, synced, syncing. (Via Wikipedia.)

  19. We discussed sceptic here.

  20. “GeorgeW, there’s soccer and the British spelling sceptic. Also arced, arcing, synced, syncing. (Via Wikipedia.)”

    Yes. The constraint seems to be word initial.

  21. Bob Gillham says:

    Backin the day on the welsh borders it would have been sist for sure, and we pronounced Kikero sisero….

  22. “Celtic (except in basketball teams)” – and soccer teams. The Glasgow team’s name is pronounced with an “s”.

    Other names from the Irish tradition beginning with Ci- include Cian, although curiously Killian (Cillian in the original Irish) is normally spelt with a K – I have nephews with those names.

  23. Stephen Bruce says:

    Part of me wants to start pronouncing Celtic /sɛltɪk/, sceptic /sɛptɪk/, and respelling the other words as socker, arcked, arcking, syncked, and syncking (or maybe synched and synching). Alas, I am not so brave. But /kɪst/ I like as the unnaturalized pronunciation of a foreign word.

  24. But /kɪst/ I like as the unnaturalized pronunciation of a foreign word.


  25. I just remembered that I have a nice book about English orthography (“The English Writing System” by Vivian Cook) and decided to see what he has to say about orthographic [c]. He gives the following rule citing Venezky, “The Structure of English Orthography.”

    [c] = /tʃ/ in ‘cello’
    [c] = /ø/ in ‘czar’
    [c] = /s/ before [i,y,e] like ‘cell’
    [c] = /k/ otherwise

    So, an English reader would naturally try to pronounce [cist] as /sist/ unless taught otherwise (like Celt).

  26. Indeed, Celt is also unique in the pronunciation list I mentioned above, the Combined Anglo-American Pronunciation Reference. Note that I am only checking initial ci-, ce-, not medial.

  27. The similar case of “soft” and “hard” g has many more exceptions, though not many where g is soft where the rule would indicate it should be hard, as in margarine and gaol.

  28. Stefan Holm says:

    Oskar: with the risk of sounding fussy, the initial sound in Swedish kista isn’t /ʃ/ but /ç/ – as in English ‘chair’ without the initial ‘t’. (But Trond is right that a merger appears among some younger Swedes). And to archeologically fine-tune the burial fashions a little:

    dös ↔ dolmen
    hällkista ↔ cist
    gånggrift ↔ passage grave / chambered barrow.

  29. marie-lucie says:

    /ç/ – as in English ‘chair’ without the initial ‘t’

    This may be the Swedish pronunciation of English. [ç] is the phonetic symbol for German “ch” as in ich . Some English speakers use it as the first sound in “huge”.

  30. “This may be the Swedish pronunciation of English. [ç] is the phonetic symbol for German “ch” as in ich . Some English speakers use it as the first sound in “huge”.”

    Indeed. “hue” and “hew” =/= “shoe.”

  31. Stefan Holm says:

    Every time fricatives are discussed it seems to end up in a can of worms. In my own language we traditionally count for three: /ʃ/, /ç/ and /x/ (except of course for the plain /f/, /s/ and /v/). But then phoneticians turn up and show that they have recorded 14 (only in Swedish)! And you can’t argue with an oscilloscope, can you? I suppose that the ways you can raise the tip, mid or rear part of your tounge form a continuum represented in both dialects and individually in different phonetic environments.

    To me both German ‘Ich’ and English ‘chair’ come with a palatal, i.e. a voiceless allophone of /j/ as in ‘yes’. In fact – to my biased Swedish ear – if I from ‘chess’ strip of the initial ‘t’ and vocalize the remaining intitial consonant I hear a ‘yes’. But I can’t get ‘shoe’ to fit in – to me the fricative here is produced with the tip of the tounge, either as a supradental, ‘above the teeth’ or a retroflex, ‘bent backwards’. (Or do you use the same fricative in ‘shoe’ and ‘chew’ or – as vocalized – in ‘you’)?

  32. But I can’t get ‘shoe’ to fit in

    English “sh” [ʃ] is usually articulated with the blade, as well as the tip, of the tongue. The front part of the tongue is bunched up, approaching the hard palate just behind the alveolar ridge. The place of articulation is exactly the same for the fricative [ʃ] and the affricate [ʧ]. Technically, these sounds are domed apico/lamino-postalveolars. If you articulate a postalveolar fricative with the tip of the tongue, and the surface of the tongue is flat or slightly curled back, the result is an apico-(or subapico-)postalveolar (“retroflex”) [ʂ]. If you use the front-to-middle part of the tongue, and the whole body of the tongue is raised towards the hard palate, the result is a lamino-prepalatal (a.k.a. “alveolo-palatal”) [ɕ]. If the mid-to-back part of the tongue (the dorsum) is involved, you get a dorso-palatal [ç].

    One of the hardest problems for English-speakers learning Polish phonetics is that Polish has a contrast between apical (slightly retroflex) postalveolars and laminal prepalatals (e.g. sz /ʂ/ : ś /ɕ/), while English /ʃ/ is articulatorily and acoustically intermediate between them. As a result, Enlish-speakers find it difficult either to realise or to hear the contrast. For the sake of completeness: our /x/ phoneme has a palatal allophone ([ç]) before /i/ or /j/, and this [ç] is distinct from /ɕ/.

    In Russian, ш (š) and ж (ž) are also apical, and even slightly “darker” than in Polsh, but, curiously, the affricate ч (č) is laminal and similar to Polish ć [ʨ] (though slightly less palatal, as if modified in the direction of English ch). The spelling щ (šč) used to represent the sequence [ɕʨ], but now its normal realisation is as a long lamino-prepalatal fricative [ɕː].

    The relevant articulatory parameters are the part of the tongue involved, the point/area of articulation, and the general shape of the tongue. These fricatives and affricates form a continuum rather than a set of sharply distinct articulations, and may easily shade into one another, hence their historical instability and synchronic variability.

  33. Stephen Bruce says:

    After reading the “SCEPTIC” post, I’ll amend my ideal pronunciation of skeptic to match my actual pronunciation, i.e. with a /k/. I normally like the Latin c, but being able to distinguish this word from septic is desirable, and in this we can be one up on the French.

    In my American accent [ʃ] and [ʧ] have an identical place of articulation, as Piotr described. But it doesn’t help that the comparisons we use are also subject to regional variation. For instance, I’ve heard the German ich-Laut varying from [x] to almost [ʃ].

    I had only ever seen [ɕʨ] as a possible pronunciation of щ in books (“fresh cheese”) until I first heard it from a Russian speaker from Ukraine.

  34. David Marjanović says:

    In fact – to my biased Swedish ear – if I from ‘chess’ strip of the initial ‘t’ and vocalize the remaining intitial consonant I hear a ‘yes’.

    That is indeed specifically Swedish phonology – though some rather standard English accents actually share it: /tj/ comes out as [t͡ɕ] or [t͡ʃ] rather than as [tj].

    For instance, I’ve heard the German ich-Laut varying from [x] to almost [ʃ].

    I doubt it’s [x] anywhere; but the Alemannic dialects simply lack the distinction of ich-Laut and ach-Laut and pronounce every /x/ as [χ].

    To the north of them, there are indeed accents that use [ɕ] or thereabouts; and this can trigger hypercorrectivisms, like Helmut Kohl pronouncing Menschen as [ˈmɛnçŋ̩].

    Finally, Turkish and Arabic lack [ç], so the sociolect of young people with a “migration background” substitutes /ʃ/ for every ich-Laut while leaving the ach-Laut intact.

  35. David Marjanović says:

    the sociolect of

    In Berlin and almost certainly other big cities in the northern half of Germany, I forgot to add. There doesn’t seem to be any such thing in Vienna, even though the density of Turkish signage in southeastern Vienna is no less than that in southeastern Berlin.

Speak Your Mind