The Wrong Schuppen.

Via Laudator Temporis Acti, John Woods (translator of Thomas Mann) on his line of work:

Literary translating is about literature, it’s about being very, very skilled in your native language. Not only do you need to know the second language, and I mean know it well, but you have to love and work with and understand your own language. You have to know how to make it sing and dance. Being bilingual certainly is a leg up, but that doesn’t create a literary text. There are a lot of people out there who are bilingual. They think, “Well, I ought to be able to translate something.” And the publishing houses know that there’s this vast pool of workers out there, and so they farm the stuff out in piecework. The results are haphazard. Some great and wonderful translations are produced, and some that are not so great and wonderful.


Every German publishing house has editors who go over every translation, comparing it word for word, with the original. I don’t know of any publishing house that does that anymore in this country.


I’ve made lots of horrible mistakes. One in fact that Helen Wolff didn’t catch, in a small book by Günter Grass called Show Your Tongue, about a year he spent in Calcutta, India. And somewhere he uses the word “Schuppen,” which can mean the “scales of a fish” or “dandruff.” And in a moment of average inattention, I chose dandruff — and it should have been the scales of a fish. That sort of thing happens, it just happens. You’re working fast, you go back and check, but it never registers, and suddenly the translation is there forever in black and white.

One of the things you learn, particularly in this job, is that there is no such thing as perfection in this existence. And you learn to live with that. It will never be perfect. Any translation can be made better both aesthetically and in terms of accuracy, and that’s why you correct four and five times yourself, and that’s why somebody else should look at it too. Because it will never be as good as it truly ought to be.

Refreshing modesty! And hurray for the German publishers who takes such pains to make sure translations are accurate.


  1. Every German publishing house has editors who go over every translation, comparing it word for word, with the original.

    True as of 1997. But today?

  2. Oh dear. Good question, and I’m afraid of the answer.

  3. “… they farm the stuff out in piecework,” looks like interference from piecemeal. The statement about piecework is (presumably) not wrong, but I don’t think it really has the implication that Woods wants to convey.

  4. And now I know why the Hebrew word קַשְׂקַשִּׂים kaskasim, ‘fish scales’, is also used for ‘dandruff’.

  5. The statement about piecework is (presumably) not wrong, but I don’t think it really has the implication that Woods wants to convey

    No, I think it does, in the sense of work paid per unit produced and often done at home. It’s a term that comes up all the time in conversations among translators, since we tend to be paid by the word or by the standard line or page (a certain number of characters), which sometimes makes about as much sense as being paid by the ounce.

    And I also share Vanya’s doubts about “every German publishing house.” The same was true of Italian publishers once upon a time, but just this morning some Italian translators I know were complaining online about how rarely they can count on it nowadays. Especially now that the editing is almost never done in-house.

    But what he said in 1997 about publishers relying on bilingual translators without worrying about their writing skills in English is also a little outdated. Nowadays there are a ton of MFA programs in translation that often have more overlap with creative writing departments than with language departments. If anything, I feel like people may have gone a little overboard with de-emphasizing the importance of L2 skills.

  6. David Marjanović says

    Every German publishing house has editors who go over every translation, comparing it word for word, [sic] with the original.

    …perhaps for literature…? Definitely not for popular books about science, and not in 1997 either. Not one of the translators for those knew that freshwater isn’t literally fresh water.

    Schuppe f., pl. -n, is also used for the scales of lizards, bird feet, rat tails, butterfly wings and so on.

  7. Stu Clayton says

    Eine Schuppe is a fish scale, or scale of many another critter. This word can be used to mean a flake of dandruff, but the German word for dandruff is a plurale tantum like “heebie-jeebies”: Schuppen, usually occurring without die [west of Wien, at any rate]. Er hat Schuppen.

    On the other hand [der] Schuppen is “shed”, as for bicycles.

    This might sound confusing, but only if you are not on board with the fact that “gender” and “declination” in German are essential features of words. They are not ballast bolted onto dictionary lemmata.

    A fully functional sword is not a razor blade with a handle.

  8. Stu Clayton says

    A fully functional sword needs context as well, such as Uma Thurman.

  9. David Eddyshaw says

    Western Oti-Volta languages use the same word for “scale” (of a fish) as “bark” (of a tree), which is one of those things that makes sense in retrospect, I suppose. So do other Oti-Volta branches, even though the actual words used are mostly not cognate with the WOV forms.

    Sadly, I know no WOV words for “dandruff.” And, unaccountably, the word does not feature in Swadesh lists.

  10. Trond Engen says

    Schuppe(n) is weird. The geminate p in High German is one thing. That I can’t think of close cognates is another.

  11. @Trond Engen: I don’t know what is going on with the long=vowel-geminate sequence, but the cognates are easy to find, if you know that the word was apparently originally applied to scales by reference to their shapes. The native English cognate is shove[l], and we also have scoop mediated through Low German (indicating that the long vowel is original).

  12. David Marjanović says

    Scab would fit, except for the vowel…

    Edit: Long vowel? In Schuppen it’s short, and likewise in Schuppe.

  13. David Eddyshaw says

    As it happens, the WOV “bark/scale” words mean “scab” as well. (Coincidentally, the stem in Proto-WOV was *pag-.)

  14. Bark is self-evidently derived from *pag-, and scab merely requires s-mobile and a bit of transposition.

  15. David Eddyshaw says

    True, true …

  16. Stu Clayton says

    Americans have T-mobile, I believe. It shows German influence.

  17. David Marjanović says

    Ha! Mystery solved! In Middle Low German alone, we have the whole Kluge mess preserved: schōpe, schōve, schubbe, schuppe! So it’s effortlessly related to shave and shove and scab and its native doublet shab that I didn’t know…

    I’m not sure how to bring shovel, German Schaufel thanks to Bahder’s law, into this.

  18. David Eddyshaw says

    On scab/bark: “bark” as in “bark one’s shin” is evidently from the tree-bark word, though the connection never actually occurred to me before.

  19. Trond Engen says

    I thought of ‘shove’ etc. but couldn’t find the semantic bridge. Same for ‘ship’ etc.

    I think the geminate p would suggest spread from Low German. If it’s cognate with ‘shove’, we must be looking at a Kluge mess.

    Edit: Of course David sorted this out.

    I should have thought of ‘shave’, but it’s a pretty long way.

  20. Bathrobe says

    you have to love and work with and understand your own language. You have to know how to make it sing and dance.

    Ah, yes, I was never into that. The linguist, I suspect, is more into stripping it down and finding out what makes it tick than making it sing and dance.

  21. David Marjanović says

    That Schuppe is standard is basically due to Luther, who did pick a “Central and Low German” form here, but pp does occur in High, even Upper, German as the regularly shifted reflex of *bb: Rippe “rib”, Sippe “clan” as in sibling, Knappe “medieval valet” as part of a Kluge mess…

  22. David Eddyshaw says

    I must admit that if I’d come across a proposed “bark/scale” semantic shift in some Nostratic or Afroasiatic comparison I’d have been rather supercilious about it. Just goes to show …

    (Oti-Volta also consistently has “root/tendon”, which has never seemed a very natural association to me, either. But then, trees aren’t animate in my Weltanschauung.)

    Maybe “wife” comes from “ecstatic trembling” after all. What do I know?

  23. Trond Engen says

    David M.: pp does occur in High, even Upper, German as the regularly shifted reflex of *bb

    Of course it does. As does tt and kk <ck> in parallel circumstances.

  24. Does it ever happen that publishers bring out a revised edition/printing of a translated text purely/mainly to fix translation errors? It seems like fixing obvious typos is a thing for texts of all types, and nonfiction texts may sometimes be revised to fix factual errors without such more substantive rework as addressing subsequent research. In principle such patch releases should be easier in the age of epublishing; I know Amazon sometimes pushes revised texts to Kindle users.

  25. suomu/suomus


    From Proto-Finnic *soomu, from Proto-Uralic *śëme. Cognates include Estonian soomus, Veps somuz, Erzya сяв (sjav), Udmurt сьӧм (sʹöm) and Komi-Permyak сьӧм (sʹöm).

    IPA(key): /ˈsuo̯mu/, [ˈs̠uo̞̯mu]
    Rhymes: -uomu
    Syllabification: suo‧mu



    1. scale (one of the keratin pieces covering the skin of certain animals)
    2. (botany) scale (individual plate of a conifer cone)



    IPA(key): /ˈhilseˣ/, [ˈhils̠e̞(ʔ)]
    Rhymes: -ilse
    Syllabification: hil‧se



    1. dandruff


    Etymology 1

    From hilse +‎ -illa.

    IPA(key): /ˈhilsei̯lːæˣ/, [ˈhils̠e̞i̯lːæ(ʔ)]
    Rhymes: -ilseilːæ
    Syllabification: hil‧seil‧lä



    1. (intransitive) to scale off, peel, peel off

    AFAIK, suomuilla can also be used to mean this (‘to come off in flakes’).


    Etymology 1

    kesi +‎ -iä

    IPA(key): /ˈkesiæˣ/, [ˈke̞s̠iæ(ʔ)]
    Rhymes: -esiæ
    Syllabification: ke‧si‧ä



    1. (intransitive) (of outermost layer of skin) to peel off

    palanut iho kesii

    burned skin peels off.

  26. тәңкә



    Cognate with Kazakh теңге (teñge, “piece of money; Kazakhstan’s national currency”), Uzbek tanga (“currency in the historical Central Asian states; fish scale”), Uyghur تەڭگە‎ (tengge, “silver coin; ruble; fish scale”), Turkmen teňňe (“monetary unit”). Russian деньга (denʹga, “denga, an old coin”).

    For Bashkir this is clearly a cultural borrowing from Central Asia: tanka was the name of the currency in the Timurid Empire, as well as in the medieval Central Asian states of Bukhara, Khwarazm and Kokand. However, a currency named so was first introduced earlier, in the 12th century, in the state of Ghōr by Alā-ad-Dīn Husayn of the Ghurid dynasty.

    According to The American Heritage Dictionary of the English Language, the original currency name came from the Sanskrit word तन्कह् (tankah) (see the reference at History of the taka#Etymology). Indeed, the currency called Sultani tanka was introduced in North India in 1329 by the monetary reforms of Muhammad bin Tughluq, the emperor of the Delhi Sultanate. However, this etymology requires elucidation in relation to the original currency name, as the Ghurid dynasty is known to have spoken a Turkic language as their native and to have promoted Persian as the language of state.

    The relation to Turkish denge (“balance”) requires research.

    IPA(key): [tæŋˈkæ]
    Hyphenation: тәң‧кә


    тәңкә • (täñkä)

    1. (colloquial) ruble

    Synonym: һум (hum)

    2. (dated, historical) silver coin
    3. fish scale
    4. (traditional costume) metal pieces of decoration (originally silver coins) sewn on women’s traditional clothes or interwoven into hair braids




    From Proto-Turkic *kāpuk (“bark; skin; shell”).

    Cognate with Kazakh қабық (qabyq), Kyrgyz кабык (qabıq), Southern Altai кабык (kabïk), Uzbek qobiq (“skin, shell”), Turkish kabuk (“shell”), Tuvan хавык (xavık, “husk”), Chuvash хупӑ (hup̬ă, “bark”), etc.

    IPA(key): [qɑˈβɯ̞q]
    Hyphenation: ҡа‧быҡ


    ҡабыҡ • (qabïq)

    1. (fruit, vegetable) skin, rind, peel
    2. (nut, egg, etc.) shell
    3. (tree) bark
    4. (reptiles, inspects) skin, scale
    5. (turtle etc.) shell

    Ташбаҡа ҡабығы.

    Tašbaqa qabïğï.
    Turtle shell.

    6. (anatomy) cortex
    7. (geology) crust
    Ер ҡабығы.

    Yer qabïğï.
    Earth’s crust.

  27. кавык/qawıq



    1. dandruff

    Looks way too similar to кабык.

  28. ҡауаҡ


    Compare to Tatar кавык (qawıq, “dandruff”).

    IPA(key): [qɑˈwɑq]
    Hyphenation: ҡа‧уаҡ


    ҡауаҡ • (qawaq)

    1. dandruff

    Ҡауаҡҡа ҡаршы шампунь.

    Qawaqqa qaršï šampun’.
    Shampoo against dandruff.



    A derivation from Proto-Turkic *kạb- (“bark; shell; husk; bran”). Doublet of qovaq.

    Cognate with Turkish kepek (“dandruff”), Uzbek kepak, Uyghur كېپەك‎ (këpek), Kyrgyz кебек (kebek, “bran; sawdust, filings”), Chuvash кипек (kip̬ek, “hull; chaff; flaky skin”) (← Tatar кибәк (kibäk, “chuff”)), dialectal Tatar кибәк (kibäk, “dandruff”).

    Compare Persian کپک‎ (kapak, “bran”), and Mongolian хэвэг (kheveg), both Turkic borrowings.[1]

    IPA(key): [kæˈpæk], [kæˈpæj]
    IPA(key): [t͡ʃæˈpæt͡ʃ]


    kəpək (definite accusative kəpəyi, plural kəpəklər)

    1. bran
    2. sawdust, filings
    3. dandruff

    Synonyms: qovaq, (dialectal) dongra

  29. I have no thoughts on dandruff, but it happens I’ve just finished reading The Magic Mountain for the second time, this time in the Woods translation, and was duly impressed. A vast improvement, it seems to me, over the H.T. Lowe-Porter translation that was the only extant one for so many years. (I note, however, that I have no German with which to gauge its relative accuracy.)

  30. PlasticPaddy says

    “The new version appears to be free of ‘howlers’, he noted, but ‘the new American translator does misunderstand the German on quite a few occasions’ (1997: 658) . Buck illustrated his argument by listing some 17 mismatches between the German original and the new English translation…”
    I think you may have to buy the book to see more detail.

  31. David Eddyshaw says

    A vast improvement, it seems to me, over the H.T. Lowe-Porter translation

    Amen to that.

    The L-P translations of Mann are astonishingly bad, so much so that I don’t think you even need to know German to tell. (For example: early on in Der Zauberberg, a group of characters are found lying around complaining about how “relaxed” they feel. You need only a basic feeling for English to know that something has gone badly wrong with the translation there. Much more follows in the same tin-eared vein.)

    I even speculated once that Lowe-Porter couldn’t be an English L1 speaker, but she was.
    Apparently she had some sort of armlock on the translation rights, and Mann thought she was wonderful.

    And she was responsible for Boris Johnson. There can be no excuses at the end of the day.

  32. January First-of-May says

    currency in the historical Central Asian states; fish scale

    Central Asian (and Russian) coins, esp. in the 14th-17th century, did look very much like fish scales (чешуйки in Russian), being small, thin, silvery, and roundish but not entirely round; the coin/fish scale analogy had been reinvented several times elsewhere, e.g. for US half dimes.

    In other words, I’m not sure we can rule out the “fish scale” word being an extension of “coin”. I think I’ve seen a few examples in that direction too, but can’t think of any offhand.
    (It helps that coins were frequently worn in large amounts in decorations vaguely resembling scale mail; the Slavic term for such decorations is monisto, which I always assumed was somehow related to “money” but TIL that it’s actually an inherited Slavic word that originally just meant “necklace”.)

    [EDIT: this is sense 4 in juha’s post: “(traditional costume) metal pieces of decoration (originally silver coins) sewn on women’s traditional clothes or interwoven into hair braids”. If the “coin” > “fish scale” shift wasn’t direct, it could easily have gone through something like that…]

  33. In other words, I’m not sure we can rule out the “fish scale” word being an extension of “coin”. I think I’ve seen a few examples in that direction too, but can’t think of any offhand.

    On words for ‘fish scale’ from ‘coin’, there is also Persian پول pūl ‘piece of a copper coin, money’ but also ‘fish scale’, along with Ottoman پول pul ‘small copper coin’ (Ottoman, from Persian), now Republican Turkish pul ‘fish scale, flake (as of red pepper); postage stamp, etc.’. (In present-day Persian, the fish scale is more often the diminutive پولک pūlak.) The Persian word is often said to descend from a Middle Iranian borrowing of Greek ὀβολός ‘obol’. Scroll down to the section entitled Units of currency and measure here. The Persian word has of course been widely borrowed in central Eurasia as a word for ‘money’.

    Arabic has فلس fals ‘small coin’ and ‘fish scale’, said to be (via Aramaic/Syriac) from Greek φόλλις ‘a small coin’, from Latin follis. (But for ‘scale’, cf. Greek φολίς, gen. φολίδος, ‘scale (usually of reptiles)’, so a more complicated situation? Or just a coincidence?)

  34. David Eddyshaw says

    The Laudator post links to

    which aligns with my own longstanding feeling (indeed, certainty, in the cases where I know enough of the relevant languages) that Pound is very often winging it in his translations.

    On the other hand, I doubt whether Pound would have been at all troubled by the point. I expect he’d have reponded along the lines that if it’s a translation you want, well, go look for a translation.

  35. Exactly.

  36. John Cowan says

    early on in Der Zauberberg, a group of characters are found lying around complaining about how “relaxed” they feel.

    No no, quite right. Relaxed used to mean ‘unstrung, enervated, lethargic’; the Victorians used to complain of the relaxing (i.e. too-warm) climate of India. Quoth Boswell: “It was a very wet morning. I woke relaxed and melancholy as in the country, and walked about an hour under cover, in the middle of the town.”

  37. Mann thought she was wonderful

    According to the story I have heard, Mann freely admitted that Lowe-Porter didn’t know any German at all. As long as they sold, he simply didn’t care about the quality of the English translations; his target audience were German readers.

    As for publishers correcting obvious mistakes, apparently the German translator of Bely’s Peterburg rendered the name of the Ukrainian philosopher Skovoroda (on the last page of the novel) as “Skoworod”. Reviewers quickly pointed out this obvious and embarassing mistake, but the translation was reprinted, without corrections, for decades (it may be still in print, for all I know).

    And then there is the story of James Joyce simply despairing about the ignorance of his German translator, Georg Goyert. The publisher claimed (wrongly) that Joyce had authorized Goyert’s translation, even when in the mid 1950s Arno Schmidt analysed the translation and came to the conclusion that it completely misrepresented Joyce’s novel. Still, German readers had to wait another twenty years for a new, improved translation (by Schmidt’s friend Hans Wollschläger).

  38. January First-of-May says

    The Persian word is often said to descend from a Middle Iranian borrowing of Greek ὀβολός ‘obol’.

    …I always thought – and I see that Russian numismatic dictionaries agree with me here – that both pul and fals ultimately derive from follis. This is the first time I’m hearing of the “obol” derivation, though it sounds intriguing.
    The 6th century follis (introduced by Anastasius) is, as it happens, about as unlike a fish-scale as a coin can be – it’s a large piece of copper, of a size unprecedented in Roman coinage since the demise of the sestertius.

    Of course, as far as small thin roundish things are concerned, we might as well consider folium “leaf, petal”…

  39. λεπτός

    Ancient Greek

    From λέπω (lépō, “I peel”) +‎ –τός (-tós).

    λεπτός • (leptós) m (feminine λεπτή, neuter λεπτόν); first/second declension

    1. (rare, literally) peeled, husked (of grain)
    2. fine-grained (ashes, soil, etc.)
    3. thin, lean (people or animals)
    4. straight, narrow (spaces)
    5. small, weak, impotent
    6. light, slight (breezes)
    7. thin (liquids)
    8. (figuratively) refined, delicate, subtle



    Etymology 1

    From Ancient Greek λεπτόν (leptón), neuter form of λεπτός (leptós, “small”).

    lepton (plural lepta or leptons)

    1. A coin used since ancient times in Greece, serving in modern times as one hundredth of a phoenix, a drachma, and a euro (as the Greek form of the Eurocent).
    2. A small, bronze Judean coin from the 1st century B.C.E., considered by some to be the widow’s mite.


    Ancient Greek

    Perhaps from λέπω (lépō, “to peel”) +‎ –ῐς (-is, feminine suffix).

    IPA(key): /le.pís/ → /leˈpis/ → /leˈpis/


    λεπῐ́ς • (lepís) f (genitive λεπῐ́δος); third declension

    1. scale, flake, shell, husk
    2. epithelial debris
    3. plate of metal
    4. blade of a saw

  40. epithelial debris reminds me of Icarus I in Sunshine.

  41. فلس fals ‘

    I thought paisa was related, but it’s not:


    Inherited from Sanskrit पद (pada, “quarter”) +‎ अंश (aṃśa, “part, piece”).

    (Delhi Hindi) IPA(key): /pɛː.sɑː/, [pɛː.s̪äː]


    पैसा • (paisā) m (Urdu spelling پیسہ‎)

    1. paisa (a subdivision of currency, equal to 1⁄100 of a rupee)

    एक रुपये में सौ पैसे होते हैं‍।

    ek rupye mẽ sau paise hote ha͠i‍.
    There are a hundred paisas in a rupee.

    2. (historical) a quarter anna (1⁄64 of a rupee)
    money, wealth, riches

    उनसे इस व्यापार में बहुत पैसा कमाया।

    unse is vyāpār mẽ bahut paisā kamāyā.
    He earned a lot of money in this business.

  42. David Marjanović says

    the widow’s mite

    Mite? Like a tick?

    Oh, the rent, like Miete in German?

  43. The Widow’s mite. Scherflein in Luther’s translation.

  44. I thought paisa was related, but it’s not

    Paisa at LH.

  45. Joyce’s novel

    Do you mean Ulysses? He wrote two others.

  46. David Marjanović says

    Oh, so like a tick after all, just a mite metaphorical.

  47. tanga (“currency in the historical Central Asian states; fish scale”…
    In other words, I’m not sure we can rule out the “fish scale” word being an extension of “coin”…

    I was trying to learn a little Guaraní today, and I saw that the Guaraní word for ‘coin, money, bill of paper currency’ is pirapire, a compound of pira ‘fish’ and pire ‘skin, hide, peel, bark, scale of a fish’. I wonder if this is an instance of ‘fish scale’ > ‘piece of silver, coin’ directly and spontaneously?

    I checked CLICS for this colexification of ‘scale’ and ‘coin’/‘money’, and it was here.

    The Komi item there, сьöм ‘(fish) scale, money’, is evidently from Uralic ‘scale’ and thus illustrates ‘fish scale’ > ‘money’. The Uralonet site indicates here that this change is found in several other branches of Uralic too (Erzya and Moksha śav ‘Bart an der Ähre; Geld’; Eastern Khanty (Tromyogan dialect) såm; Mansi (Lower Losva River dialect) sē̮m.) I wonder, could the idea of ‘coin’ = ‘fish scale’ (as in Persian پول pūl) have diffused from central Eurasian khanates into these languages? Perhaps along with coins themselves? (Along the lines of Russian деньги ‘money’ from a Turkic source, for instance.)

    There is also this note on that Uralonet page:

    Zum Bedeutungsverhältnis ‘Fischschuppe’ ~ ‘Geld’ vgl. mord. (s-Laute 36) M kaloń-kapeka ‘Fischschuppe’ (kapeka < russ. копeŭка ‘kleine Kupfermünze’), ung. pénz ‘Geld’ ~ dial. halpénz ‘Fischschuppe’ (hal ‘Fisch’); selk. (Erd. 76,99) Ta. qomtä ‘Münze, Geld’ ~ komtä ‘Schuppe’.

    I gather that the earlier Tatar word for ‘ruble’ was сум (cf. Bashkir һум, Uzbek сўм (soʻm), Kazakh and Kyrghyz сом, etc.), from earlier Turkic som ‘in a single mass, with no admixture or intrusions, pure’ (as for example gold). I suppose the agreement in consonantism between the Turkic and Uralic is just a coincidence and doesn’t have anything to do with the Uralic developments.

Speak Your Mind