Friday 18 March 2011

initial ps-, ts-, ks-

One or two people have written to me to suggest that I discuss the English pronunciation of tsunami. I have not done so, because the matter has already been extensively aired on Language Log, including here.

OK, then: let’s agree that if we articulate very carefully we can reproduce in English the Japanese affricate ts (which is the allophone of /t/ used before u), giving tsuˈnɑːmi. But once we have got into our stride we’re likely to simplify the initial affricate to a plain fricative, suˈnɑːmi.

(Fun fact: in the local speech of Kōchi 高知, on the island of Shikoku, this allophone is not used, and people say tunami.)

The same sort of thing happens in other borrowed words that have initial ts in the donor language. So tsetse (fly), from Tswana, becomes English ˈ(t)setsi. However with tsar (from Russian царь tsarʲ) and Zeitgeist (German pronunciation ˈtsaitgaist) the English outcome tends to be a voiced fricative, thus ˈzɑː, ˈzaɪtɡaɪst. In the latter word this is easily explained as a spelling pronunciation giving the letter Z its usual English value, and I suppose tsar may be similarly influenced by the alternative and older spelling czar, which the OED dates from 1549.

This voicing uncertainty made me think about other cases where we have borrowed from a foreign language words beginning with a consonant plus s. Notably, these are the Greek-derived words spelt in English with ps- and x-.

It is clear that words with Greek initial psi (Ψ, ψ, ps) end up in English with a voiceless s, while those with initial xi (Ξ, ξ, ks) end up with z. So on the one hand we have psalm, pseudonym, psoriasis, psychology etc with s-, but Xerxes, xenon, xenophobia, xylophone with z-.

So the bilabial produces a voiceless outcome, the velar produces a voiced outcome, and the alveolar can go either way. I do not know why.


  1. The Polish pronunciation of tsunami is a bit more interesting - an initial phonetic affricate that is a sequence of /t/ and /s/ and distinct from the phonological affricate /ts/ (written 'c') found, for example, in the word 'car' (tsar).

  2. Could the [z] in words starting with /ks/ be influenced by French pronunciation? At least presentday French has [gz] in those cases (I'm not sure about medieval French).

  3. Could be that they got rid of the plosive at different times, eg ps- -> s- when initial S was still automatically voiceless.

    (French has ks- initially, hasn't it?)

    Anyway -

    Schoolmaster: Why do we say sɑːm and saɪkɪ, not [makes grimaces, tries hard] pəsɑːm and pəsaɪkɪ? Yes, Jones?

    Schoolboy: Because the Greek had different mouths and we English don't pronounce unnatural words.

    Schoolmaster: Pcisely.

  4. It hadn't occurred to me that people might be pronouncing tsunami as suːˈnɑːmi rather than tsuː-. It obviously didn't occur to our choir director who, when explaining the pronunciation of a German piece the other day, simply said 'zu, as in tsunami' and moved on to the next word.

  5. In French, initial ps is pronounced ps (psychologue psikɔlɔg) and initial x is gz (xénophobe gzenɔfɔb), which is the same as the English but with the plosive preserved.

    So presumably English takes its lead from French. But that doesn't explain anything: why does French pronounce those letters like that? In Greek they were pronounced ps and ks. Could it have been Latin that introduced the voicing of the Xi?

  6. I'm quite sure French has at least both ks- and gz-. Does anyone know details?

    In theory, another explanation could be boundaries for intervocalic -ps- and -ks-, the result of which might have been taken over to initial sounds.

  7. I find this absolutely demented, i.e. the change from the affricate to the fricative. But it's everywhere. If only people pronounced it tsuˈnɑːmi, suˈnɑːmi sounds inexact and a bit 'illiterate'. Though it's legitimate, I know.

  8. Steve, I don't know how it hasn't occured to you, almost every newsreader and reporter pronounces it like that. Ugh!

  9. I just wish English had borrowed tsunami as /t(s)uˈnɑːmi/ instead of /(t)suˈnɑːmi/ to reflect the Japanese phonemes better, but of course that's not what an English speaker perceives when presented with /#ts-/. :-(

  10. Beatrice Portinari18 March 2011 at 10:42

    @ Jana: Spanish speakers feel exactly the opposite, so that it would sound a bit pedantic if you pronounced (in Spanish) the t in tsunami or the p in psique.

  11. I've followed the Language Log link and had a look at the comments. How disappointing to see all the fuss and snobbery evoked by the question. Has anything changed since the days of Robert Bridges' Society for Pure English? Must there be only one correct pronunciation for each word and must we condemn those who don't use it as illiterate, uneducated and repugnant?

  12. @Lipman - initial ks- in French doesn't ring any bells for me, but looking on Wiktionary I've found a few:
    * xylographier ksi.lɔ.ɡʁa.fje
    * xylophone ksi.lɔ.fɔn ou gzi.lɔ.fɔn
    * xérophage kse.ʁɔ.faʒ
    * xérotribie kse.ʁo.tʁ ou gze.ʁo.tʁ

    Only a few but they're definitely there. The rest nearly all have gz (in Wiktionary).

    And the strange thing is the split between ks- and gz- occurs even within a single prefix: xylographier ksilɔɡʁafje but xylophage gzilɔfaʒ, both from the prefix xylo-/Greek ξύλον.

  13. My Robert Micro — a small but serious monolingual dictionary — lists only

    xénophobe [gzenɔfɔb]

    xérès [xeʀɛs; kseʀɛs; gzeʀɛs]

    xylène [ksilɛn; gzilɛn]

    xylo- [gzilɔ] ou [ksilɔ]
    ▶ [gzilɔfɔn; ksilɔfɔn]

    I can't find any statement of principle behind the ordering of different variants. It may or may not be significant.

    I also own a copy of the Larouuse dictionnaire d'ancien français. There is not a single entry spelled with initial X.

  14. More from le Robert micro

    tsar [dzaʀ]
    tsarévitch [dzaʀevitʃ]
    tsarine [dzaʀin]

    tsé-tsé [tsetse]

    tsigane[tsigan] ou tzigane[tzigan]

  15. @David:

    xérès I've heard as gzeʀɛs, and kseʀɛs wouldn't surprise me now, in light of the above discussion. But xeʀɛs? French hasn't even got x (although I suppose that would be a reasonable approximation of the original Spanish 'χeɾes).

  16. Murray gave [ps-] preference over [s-] in ps- words in NED. This survived in OED2, interleaved with the ps- words added by the supplements, for which only [s-] is specified.

    French also preserves initial [p] in pn- words.

    I wonder whether initial [k] in chthonic is commoner among fans of HP Lovecraft.

  17. why do we never get [bz-] for ψ- but generally do get [gz-] for ξ-?

    I realise it shifts the question somewhere else, but surely the immediate reason is that we get [gz] for intervocalic Latin -x-, but we don't get [bz] for intervocalic Latin -ps-.

  18. I realise it shifts the question somewhere else

    The OED suggests that the deciding factor is whether the succeeding vowel is stressed or not:

    The pronunciation of the prefix ex- followed by a vowel or h varies according as it bears the stress or not, the general rule being that ˈex— = /ɛks/ and ex—ˈ = /ɪgz/ , as exile /ˈɛksaɪl/ , exact /ɪgˈzækt/ , exhort /ɪgˈzɔːt/ ; but there is considerable variety in individual words and individual usage: ... . The same general principle governs the pronunciation of anxious /ˈæŋkʃəs/ , anxiety /æŋgˈzaɪɪtɪ/ , luxury /ˈlʌksjʊərɪ/ /ˈlʌkʃərɪ/ , luxurious /lʌgˈzjʊərɪəs/ /lʌgˈʒʊərɪəs/ Alexander /ælɪgˈzɑːndə(r)/ /-æ-/ , Alexandrine /ælɪgˈzɑːndrɪn/ /-æ-/ ; but here also individual usage varies.

  19. On initial x- the OED writes:

    The third value /z/ , arising from a reduction of /gz/ , is given in all cases to initial x, as Xerxes /ˈz3ːksiːz/ ; this value is shown in many instances in the 17th and 18th centuries by the spelling with z, as Zanthian, zebeck, Zerez, and instances are not uncommon in the 19th century of zantho- and zylo- for xantho- and xylo-; early examples are Zanctus Xanthus (Lydg. Troy Bk. ii. 731 rubric, 15th cent.), zyphe xiph n. (1572). Cf. Santippa Xanthippe (Chaucer), Cerses Xerxes (Wyntoun Chron., S.T.S., III. 54). A similar reduction of x took place in French:

    x, if he be the fyrste letter of a worde, as xenotróphe, xylobalsóme, whiche they sounde but s, sayenge senotrophe, sylobalsome, for they can nat gyve x, whiche is also a greke letter, is true sownde. (Palsgr. Esclarc. (1530) i. xxv.)

  20. Of course there must be a correct pronunciation. You can't pronounce words the way you like it or think is correct. Certainly, a word can have two, three, four or five pronunciations, but then it all becomes mockery and legitimizes atrocities and crimes against language and reason. If everyone were to speak the way they want to, we'd be living on the seventh floor of the Tower of Babel. Then again, perhaps we're already there – just listen to the dreadful pronunciations on TV or radio. Or the lack of proper coaching in drama schools.

  21. Paul Carley:

    The Society for Pure English was not what you think it was. Most members were literary figures, but there were also philologists, including Henry Bradley, the only English editor of the OED before 1993, when John Simpson began to work on OED3. Their prospectus, one of five "tracts" reprinted by Project Gutenberg, listed among the purposes of the society:

    The maintenance of the traditional English pronunciation of Latin, Greek, and other borrowed words;

    The creation of new words from native stock, especially by men of letters rather than of science;

    The maintenance of traditional dialects;

    Opposition to the use by schoolteachers of bogus rules of grammar, notably the split infinitive.

  22. Strange. I've evidently learned to pronounce "Xavier" as /zavje/ and "xénophobie" as /zenɔfɔbi/ and I don't think that I could have just missed this all this time.

    While it's certainly possible that my French is affected by English I wonder if Manitoban French might be simplifying the cluster too and it just rubbed off on me. I'll have to look into that to be sure.

    To Pete and Lipman, a sure example of /ks-/ in French is the Greek lettername xi. I can't imagine it ever being pronounced /gzi/.

    As for crazy clusters in English, my all-time favourite is "Tlingit".

  23. @John Cowan

    I know what the society was. I have the original tracts here and am long familiar with Bridges' works and opinions. For me the purposes of the society that you quote don't support your argument, but weaken it. Our approaches to the matter are fundamentally different, I fear.

    I am also aware that from the very beginning the society included among its members most of the biggest names in English philology. However, for me this doesn't prove the society's credentials, but shows that these scholars, despite their learning, were products of their time.

  24. Interestingly, Léon Warnant's Dictionnaire de la prononciation française favours /ks/ in most cases.

    Warnant gives only /ks/ for xénon, xénophile, xénophilie, xénophobe, xénophobie, xérophtalmie, xérus, xiphidion, xiphoïde, xylène, xylidine, xylocope, xylofer, xylographe, xylographie, xylol, xylophage, xylophone, xyste, and xystique. For the letter 'x' he gives /iks/ and /ksœ/ (for consonant letters he seems to be giving the consonant value plus /œ/ in addition to the traditional letter names).

    Warnant gives only /gz/ for xanthie, xanthine, xantholin, and xanthophylle, while giving /ke-ʀɛs/ for xérès and /kis(-)tʀ(ə)/ for xystre.

    In his section on proper names, /gz/ does come up more often; it is the only possibility given for Xanthe, Xanthi, Xanthippe, Xanthippos, Xavier, Xénocrate, Xénophane, Xénophon, and Xerxès. By contrast /ks/ is the only choice given for Xénia (the Russian grand duchess), Xertigny, Xuthos, and Xylander.

    The rest are as follows:
    Xaintrailles sɛ̃-tʀɑː(-)j(ə)
    Xalapa xa-la-pa et hga-
    Xenia (aux États-Unis) ze-nja
    Xenil xenil et hge-
    Xérès ke-ʀɛs
    Ximénès gzi-me-nɛs et ki-
    Xingu ʃiŋ-gu

    So Warnant gives /x/ or /hg/ (!) for a couple of Spanish-derived names and /k/ for another couple. In this context it's not surprising that le Robert Micro gives /xeʀɛs/ as a possibility for xérès, although Warnant prefers the more frenchified /k/ in this case. The /k/ for the Greek-derived xystre is more puzzling.

  25. I think we have enough in the way of general rules.

    1. A very general rule makes combinations of stop consonant + sibilant assimilate to both voiced or both voiceless.

    2. A specific rule for /ks/ (assuming there is no morpheme boundary between the two consonants) adds voicing
    • in the presence of a following stressed vowel (whether primary word-stress or secondary)
    • in the absence of immediately preceding consonant or stressed vowel

    [This rule may need improvement, correction even.]

    3. Another general rule reduces the combination to a simple sibilant at the beginning of a word.

    Psalm, pseudonym, psoriasis, psychology etc are explained by Rules 1 & 3.
    Xerxes, xenon, xenophobia, xylophone are explained by Rule 2 operating before Rule 3.

    The fact that they are all derived from Geek words with ψ- and ξ- explains the underlying phonological forms and why English accommodates them, but is otherwise irrelevant.

    Rule 3 explains the usual pronunciation of tsunami without appeal to Greek — or Japanese — spelling.

  26. @ Thomas W. Some #ts words get /t/ in English
    Tsawwassen (a suburb of Vancouver B.C.)

  27. David Marjanović20 March 2011 at 00:10

    And again one of my comments has disappeared overnight. What is going on???

    If everyone were to speak the way they want to, we'd be living on the seventh floor of the Tower of Babel. Then again, perhaps we're already there

    Of course we are, and we have always been, ever since the origin of language or longer. Live with it.

  28. I've been doing the electronic equivalent of thumbing through the OED — expecting to find evidence for an underlying /ks/ realised as [gz] when roots such as xen and xanth are prefixed so as to make the /ks/ intervocalic.

    Bang went the theory. What I found instead were words like axenic eɪˈzɛnɪk and euxanthic juːˈzænθɪk.

    Another bout of thumbing failed to find any words with spelling that reflected initial /ks/ in a source language. Only x- spelling seems to be tolerated. Nearly all words with that spelling are from Greek. The only other words with a x- spellings are a couple of words taken from (old) Spanish, where the pronunciation was never ks.

    It seems as if the afterlife of x as a successor to Greek ξ seems to have some sort of psychological reality in English — among the small number of people who use technical words.

  29. What I found instead were words like axenic eɪˈzɛnɪk and euxanthic juːˈzænθɪk.

    However apsychical has the reassuring pronunciation æpˈsaɪkɪkəl.


Note: only a member of this blog may post a comment.