Barlaam and Josaphat.

Another interesting post at the British Library’s Asian and African studies blog:

Europeans became increasingly interested in the cultures and religions of the Middle East and Asia, or what they later called ‘the Orient’, as a result of trade relations throughout the first millennium CE. Images of Buddha with the Greek lettering ΒΟΔΔΟ (‘Boddo’ for Buddha) were found on gold coins from the Kushan empire dating back to the second century CE. Buddha was mentioned in a Greek source, ‘Stromateis’, by Clement of Alexandria as early as around 200 CE, and another reference to Buddha is found in St Jerome’s ‘Adversus Jovinianum’ written in 393 CE. A religious legend inspired by the narrative of the ‘Life of Buddha’ was well known in the Judaeo-Persian tradition and early versions in Persian, Arabic, Hebrew, Armenian and Georgian have been discovered. The story became commonly known as ‘Barlaam and Josaphat’ in medieval Europe. The name Josaphat, in Persian and Arabic spelled variously Budasf, Budasaf, Yudasaf or Iosaph, is a corruption of the title Bodhisattva which stands for ‘Buddha-to-be’, referring to Prince Siddhartha who became Gotama Buddha with his enlightenment.

Fragments of early versions of the legend seem to have been preserved in Manichean texts in Uighur and Persian from Turfan, and it is thought that Manicheans may have transmitted the Buddha narrative to the West. From there the story was translated into Arabic, and into Judeo-Persian and Syriac. An early Greek version is attributed to St John of Damascus (c. 675-749 CE) in most medieval sources, although recent researches reject this attribution as it is more probable that the Georgian monastic Euthymios carried out the translation from Georgian into Greek in the 10th century CE. It became particularly popular throughout the Christian world after it was translated into many different languages in the Middle Ages, including Latin, French, Provençal, Italian, Spanish, English, Irish, German, Czech, Serbian, Dutch, Norwegian and Swedish.

The spread of the legend of Barlaam and Josaphat in medieval Europe was a cultural phenomenon second to none at the time. Poetic and dramatized versions of the legend became what today would be called ‘bestsellers’. In Christian Europe these two names were commonly known and the Buddha as St Josaphat became a Saint with his own feast day in the Christian calendar: 27 November. […] The legend became particularly popular in Germany through the Austrian poet Rudolf von Ems’ poetic German version that was composed on the basis of a Latin version around 1230 CE. In Scandinavia a translation into Old Norse was ordered by King Haakon Haakonsøn in the 13th century, which was the basis of later translations into Norwegian and Swedish. From a Syriac version translations into Old Slavonic and then Russian and Serbian were produced. […]

Europe was not the final destination of the Buddha narrative in form of the legend of Barlaam and Josaphat. The existence of the story was also known in Ethiopia, perhaps well before the 16th century. It was documented by Abha Bahrey, a 16th-century Ethiopian historian who mentioned the book, possibly a translation into Ge’ez (Ethiopic) from Greek, in his ‘Psalter of Christ’ dated 1528 CE. After the official adoption of Christianity in 330 CE, Ethiopian Christians began to translate the sacred texts: the Bible, the New Testament and the Pentateuch into the Ge’ez language. Many writings that were first compiled in Aramaic or Greek have been fully preserved only in Ge’ez as the sacred books of the Ethiopian Church. There is a vast corpus of scriptures that have survived exclusively only in Ge’ez.

That story got around! There are more details, and splendid illustrations, at the link; I myself have David Marshall Lang’s Englishing of the Georgian version, The Balavariani, and I now discover it’s available at, as are the Ems German version, a 13th-century French one, and perhaps still others. Thanks, Trevor!


  1. >>The name Josaphat, in Persian and Arabic spelled variously Budasf, Budasaf, Yudasaf or Iosaph, is a corruption of the title Bodhisattva which stands for ‘Buddha-to-be’, referring to Prince Siddhartha who became Gotama Buddha with his enlightenment.

    Until I read this, I was convinced that Josaphat was a Slavic version of Jehoshaphat from the Bible, and Barlaam is also Slavicized version of Greek/Church Slavonic Βαρλαάμ (maybe from Chaldean Bar Lechem according to Wikipedia)

  2. Yes, Josaphat and Jehoshaphat are tempting faux amis.

  3. David Eddyshaw says:

    Sadly, this is only available if you have a JSTOR account (though I imagine most of us do.)

    It’s about a Buddhist Jātaka story which is now found pretty much everywhere in the Old World, including multiple African versions and Chaucer’s Pardoner’s Tale.

    The Bodhisattva himself had sadly gone missing from the story by the time the story got to the Kusaasi, but the radix malorum est cupiditas moral is alive and well, and seems to be a favourite everywhere.

  4. Not even faux amis. I think the early translators purposefully ‘translated’ the unfamiliar name into a similar-sounding but familiar one. Something like Zuckerman’s “camouflaged borrowing”, a.k.a. “phono-semantic matching”.

  5. ə de vivre says:

    As an aside, “the title Bodhisattva which stands for ‘Buddha-to-be’, referring to Prince Siddhartha who became Gotama Buddha with his enlightenment” seems like a weird way to define the word “Bodhisattva.” The important part of Bodhisattvahood is the delaying of becoming a Buddha, not the eventually becoming one. And identifying the historical Buddha as a Bodhisattva is a similar level of “technically true but pragmatically odd.” He’s essentially never referred to that way. Etymologically, Bodhisattva is a being of perfect knowledge, but as we all know etymology doesn’t tell you what a word means in use…

  6. Stu Clayton says:

    Dragging your feet as a virtue. This implies that there are a lot of Bodhisattva.

  7. Pro crasti nare om

  8. David Marjanović says:

    Haakon Haakonsøn


    Dragging your feet as a virtue.

    The point is to refuse nirvana until everyone else gets it, too.

  9. Haakon Haakonsøn

    Should be Old Norse Hákon Hákonarson or Norwegian Håkon Håkonsson. Somebody should alert the British Library.

  10. nobody ever asks about poor old Barlaam…

  11. There are Yiddish versions going back to the fifteenth century (translations of Ibn Hasdai’s Hebrew version). One Yiddish version, published in Żółkiew/Zhovkva in 1771, is particularly important because it is one of the earliest texts with clear elements of modern (Eastern) Yiddish, as detailed in Kerler’s Origins of Modern Literary Yiddish. Until now I had no idea of the origin of this text; I just thought it was a piece of medieval Hebrew literature.

  12. J.W. Brewer says:

    There was a Russian archbishop Barlaam (or Varlaam, if you prefer) in the early 19th century who got defrocked for allegedly publicly praying for Napoleon instead of the Czar during the period in 1812 when his diocese was under French occupation, as well as some early-medieval Russian saints of the same name. You’d think the name might have gotten rather skunked in Russian religious circles after the 14th century (when Barlaam of Calabria was the heretical-arch-villain opponent of St. Gregory Palamas in the so-called Hesychast controversy), but apparently not enough to completely eliminate the name from the available part of the lexicon.

  13. Seems to be pretty popular as an ecclesiastical name. (There are a couple of Greek and Bulgarian ringers in there, but the vast majority are Russian, and some quite recent.)

  14. Speaking of faux amis, does Barlaam ride an ass?

  15. Lars (the original one) says:

    Does Barlaam correspond to any specific one of the teachers that Gautama Buddha is said to have studied with? The most obvious, after a quick read of Wikipedia, would be Ārāḍa Kālāma or Udraka Rāmaputra, but it’s hard to see an obvious connection. What I can find online only traces the Boddhisattva > Josaphat development starting with 6th century Persian forms, and I don’t know if the Barlaam figure was present already then.

    Update: Well, it was. Belawhar o Būdāsaf according to Encyclopædia Iranica. And something about Manichean versions which I don’t know how to root out. So Barlaam is basically corrupted except for the starting consonant, but the Persian form is no closer to those Indian names.

  16. There are also Polish given names, male and female, Józefat / Józefata, <= Josaphat

  17. John Cowan says:

    Sadly, this is only available if you have a JSTOR account (though I imagine most of us do.)

    Far from it. We all, however, have an “account” on, should we choose to use it.

  18. Trond Engen says:

    Haakon Haakonsøn

    When did Danish historians stop using this convention for Norwegian kings? Or did they?

  19. Lars (the original one) says:

    The local prinsessenkrant had this in 2016, about the Danish kromprins participating in Birkebeineren:

    Ud over de fysiske strabadser er der det særlige ved løbet, at alle deltagere skal bære en rygsæk på 3,5 kilo, der symboliserer den nyfødte kongesøn Haakon Haakonsøn, der efter sin fødsel i 1204 måtte bringes i sikkerhed under den daværende borgerkrig.

    (So many commas… I’m agin that system, also the consecutive der relativizers, one of which could be som for better flow).

    But they seem to have it directly from the Norwegian(!) home page of the race. I also found a single reference with that spelling in the Danish Biographical Encyclopædia (entry from 1934), but the main entry in Den Store Danske Encyklopædi spells him Håkon 4. Håkonsson.

  20. Trond Engen says:

    Weird. Well, not that weird. Birkebeinerrennet was founded in the 1930’ies, when the Riksmål tradition still clinged to danified forms. Some still do.

    I was actually a little surprised discovering that Danish conventions render the names of Norwegian kings in Norwegian form. Norwegian conventions name the Danish kings as e.g. Svein Tjugeskjegg, Harald Sveinsson and Svein Estridsson. For some reason, though, we use Erik for Danish and Swedish kings, even though we use Svein.

  21. David Marjanović says:

    Wait – -søn is real?

  22. Trond Engen says:

    Yes. And further reduced it became the -sen of modern Danish and Norwegian surnames.

    It’s also analogical from the umlauted plural.
    Nynorsk ein son – sonen – søner – sønene
    Bokmål en sønn – sønnen – sønner – sønnene
    Swedish en son – sonen – söner – sönene
    Danish en søn – sønnen – sønner – sønnerne

    Note that the inherited u-stem plural also yields the suffix vowel e in Nynorsk and Swedish. The plurals in Bokmål and (I think) Danish are regularized, though I have very little grasp of what rules the plurals -ene and -erne in Danish.

  23. PlasticPaddy says:

    @lars 11 July
    Almuth degener “barlaam the priest”, Zeitschrift der Deutschen Morgenländischen Gesellschaft vol. 164 November. 2 (2014) seems relevant. You could ask her to upload it to Academia…

  24. John Cowan says:

    Prescriptive German (at least the version taught to foreigners in the 1970s) says that if there would otherwise be three forms of der in a row, you must change the middle one to welcher. The example I remember is Die, die die Stimme hören, … ‘Those who heard the voice …’, which has to be written Die, welche die Stimme hören …, though otherwise welcher as a relative pronoun is considered to be Old High Pretension and is removed by copy editors. I don’t know if any of this is still true. (As an interrogative determiner/pronoun it is still current, of course.)

    Lojban has no word for interrogative which. One of the ways of expressing it is xomoi, combining the wh-word specific to numbers, roughly ‘How many?’, with the ordinal suffix. Thus ‘Which horse do you want?’ can be expressed as do djica le xomoi xirma lit. ‘you desire the how-many?-th horse’. (Lojban does not have wh-fronting.)

    An answer might be le semoi ‘The seventh’, but of course the answer need not meet the terms of the question directly: le blabi ‘The white one’ is equally cromulent.

  25. Trond Engen says:

    Me: Note that the inherited u-stem plural also yields the suffix vowel e in Nynorsk and Swedish.

    At least I hope that’s it.

    John C. Old High Pretension

    I’ll remember that, use it, and imagine I came up with it myself.

  26. PlasticPaddy says:

    On the web you can find, “die die das Sagen haben, werden nicht gewählt”. But it seems to be a misquote from the illustrious H. Seehofer.

  27. Stu Clayton says:

    though otherwise welcher as a relative pronoun is considered to be Old High Pretension and is removed by copy editors. I don’t know if any of this is still true.

    The use of welcher as a non-interrogative relative pronoun was not primarily OHP in the first place, as far as I know from my reading (18-19C). It was simply a practice which has fallen into desuetude today among people who know and care deeply, passionately, and peevishly about these things. The three-in-a-row rule I’d not heard of, but that’s OK because non-interrogative welcher is now Papierdeutsch anyhoo, as Duden remarks.

    There’s only NLH (New Low Highbrow) now, which ordinary people sometimes use in writing because they were taught in school that it’s the proper register for writing – Papierdeutsch eben. Maybe it should be called Schreibflächendeutsch these days.

    Of course this all may be different in Austria …

  28. The use of welcher as a non-interrogative relative pronoun was not primarily OHP in the first place, as far as I know from my reading (18-19C).

    Well, of course it wasn’t OHP in the 18-19C; the whole point of OHP is that it tries to keep old forms artificially alive when all the non-pretentious people have given them up as a bad job.

  29. Stu Clayton says:

    We have a three-phase model ! P1: unremarkable, P2: pretentious, P3: gradeschool retro.

    Your remarks seem to imply that old-timey folks can’t be pretentious, because they started it all. The way I see it, though, pretentiousness is not necessarily backward-looking. It’s always upward-looking.

  30. Your remarks seem to imply that old-timey folks can’t be pretentious, because they started it all.

    No, of course there were just as many pretentious bastards back then as now. I’m just saying that you can’t prove anything one way or the other about OHP by referring to usage a couple of centuries back.

  31. Stu Clayton says:

    Correction: pretentiousness is always upward- and downward-looking.

  32. Lars (the original one) says:

    Zeitschrift der Deutschen Morgenländischen Gesellschaft vol. 164 — volume 163 is up on their website, they are only 5 years behind. I didn’t find any DOI so no luck on sci-hub.

    Wait – -søn is real?

    Very much so, you can actually have a p/matronymic surname of that form since a few years back. (A lot of -sen names are protected, but Larssøn is not the same as Larsen).

    But for historical names, Danish tradition uses -sen forms even though it’s a bit anachronistic. I sometimes see Svend Estridsøn, maybe because it’s a matronymic. (Should that really be metronymic? And is matriarchy unclassical as well, then?)

    -ene vs -erne: This one is easy, if the PL.INDEF. ends in -er, the PL.DEF. ends in -erne, otherwise -ene. In principle this extends to foreign plurals, though most people are a bit uncomfortable with the resulting form (but have no alternative):

    et virus
    nogle vira
    alle viraene

    English plurals in s feel even weirder, but what can you do. Alle dirtbikesene

  33. Should that really be metronymic?

    There exists such a word, but it’s rare; OED (December 2001):

    Etymology: Alteration of matronymic n., either after rare Hellenistic Greek μητρωνυμικόν, use as noun of neuter singular of (otherwise unattested) adjective μητρωνυμικός, or after its etymons ancient Greek μητρ-, μήτηρ mother n.1 and -ωνυμικός -onymic comb. form. Compare French métronymique (1898). […]

    1868 J. B. Lightfoot Epist. Philippians (1873) 55 In not a few instances a metronymic takes the place of the usual patronymic.
    1904 J. A. Nairn Herodas 9 It is noticeable that Gryllos has a metronymic, not a patronymic.
    1978 Norfolk Archaeol. 27 67 A metronymic derived from the Greco-Latin Ismena.

    Matronymic is “A borrowing from Latin, combined with an English element.”

  34. John Cowan says:

    I may have overgeneralized. Perhaps it is only three identical forms of the demonstrative/relative in a row that demand welcher, in which case die, die das is fine.

  35. David Marjanović says:

    It’s also analogical from the umlauted plural.

    Oh. That’s weird. I guess it makes some sense for an u-stem, though.


    I don’t think I was taught an explicit rule, but the result of the usual aversion to repeating any words at all was probably the same on average.

    the how-many?-th

    das wievielte

    No joke. Perfectly cromulent. 234,000 ghits, the first page nothing but dictionary entries. One of them gives an example: der wievielte ist heute – “which day of the month is today”. It never occurred to me how bizarre this is typologically; Bathrobe, can you say dì jǐ or dì duōshao in Mandarin?

    On the web you can find, “die die das Sagen haben, werden nicht gewählt”. But it seems to be a misquote from the illustrious H. Seehofer.

    Definitely needs an extra comma!

    (Also, “those who have actual power aren’t elected” is a strange thing to say for Seehofer.)

    Of course this all may be different in Austria …

    No. Reportedly, though, welch- remains common in writing in Switzerland, where Standard German is already OHP to begin with.

    Should that really be metronymic? And is matriarchy unclassical as well, then?

    Let’s just pretend they’re Doric. 🙂

  36. Lars: the article is available on library genesis:

    The interface is better there also. You can search by title or author, in the scientific articles subsection. It gives you the DOI, if you still need it after grabbing the article.

  37. earthtopus says:

    das wievielte

    To join the pile, Czech has the adjective “kolikátý”; “der wievielte ist heute?” has a Czech equivalent in “Kolikátého je dnes?” = what’s today’s date?

    In my head the “English” equivalent is always “whichth.”

  38. John Cowan says:

    According to my investigation, matriarchy is a calque of Mutterrecht using Latinate morphemes appearing around the same time in both English and French.

  39. David Eddyshaw says:

    I would imagine that both matronymic and matriarch as English words are likely to owe their unetymological vowels to the analogy of patronymic and patriarch.

    Again, in these Latter Days of the Law even the supposedly literate see nothing amiss with barbarisms like television and bicycle; it is (alas) all too probable that such degenerate offspring of illustrious forebears might imagine that the initial components of these compounds were in fact (shudder) Latin.

  40. *makes apotropaic gesture*

  41. “Ceterum censeo televisionem esse delendam” – Cato the Elder, in De mediorum cultura.

  42. Lars (the original one) says:

    shudder — yes, that’s basically what Hat’s OED quote said. I’m buying it.

    Also as usually quoted in Denmark, it’s praeterea censeo … and supposedly a backtranslation from Plutarch’s original Greek. Δοκεῖ δέ μοι καὶ Τελεόρᾱσῐν μὴ εἶναι — you be the judge.

    (Well, Καρχηδόνα. And should it be Τελόρᾱσῐν classically or is it only a linking -o- that gets elided before vowels?)

  43. David Marjanović says:

    might imagine that the initial components of these compounds were in fact (shudder) Latin.

    Each component of a new scientific name nowadays has a 50-50 chance of being declared Latin or Greek as a matter of uncommented fact when such a name is coined.

    “In Ancient Grome, where Classical was spoken…”

  44. Lars (the original one) says:

    @Gary, I can’t find the article on that site — lots of other articles by Almuth Degener, but none from ZDMG. Do you have a direct link?

  45. Rodger C says:

    According to my investigation, matriarchy is a calque of Mutterrecht using Latinate morphemes appearing around the same time in both English and French.

    Now you’re reminding me of all those already-“classic” books I read in the 60s about the culture-cycle of the mother-right by the nature-folks. (Okay, the last bit is a stretcher. But not the rest.)

  46. Even the Danish Wikipedia lists the variant with ceterum first; the German Wikipedia doesn’t know the version with praeterea at all. But as you say, it’s moot, as the original account about Cato’s phrase is in Greek.

  47. Lars (the original one) says:

    It’s an even bet that editors of the Danish WP start by uncritically copying English WP; the traditional Danish form was not added until 2014. Neither of the articles notes that it was transmitted through Greek. (Also English WP in the article “Punics” has the præterea form, so it’s not unknown outside Denmark).

    Den Store Danske Encyklopædi has “ceterum censeo, (lat.), for øvrigt mener jeg; for øvrigt stemmer jeg for; [see] præterea censeo” and then a much fuller explanation there (including the bits about Cato and Carthage and backtranslation). Also that’s what I learnt in school. Even my father who never went to high school could quote it.

    I am going to make a note of this, a rare case where Danish scholarly tradition does not slavishly follow the German! I haven’t found the ultimate source for the form with præterea, but there must have been more than one Latin translation of Plutarch; or maybe somebody made a Danish translation directly from the Greek and though they should put a Latin quote in Cato’s mouth.

    I just found The Authenticity and Form of Cato’s Saying “Carthago Delenda Est”, Charles E. Little, The Classical Journal Vol. 29, No. 6 (Mar., 1934), pp. 429-435. He quotes Livy: Catone suadente bellum et ut tolleretur delereturque Carthaginem, Pliny: Cato […] cum clamaret omni senatu Carthaginem delendam, and in a later source Cato inexpiabili odio delendam esse Carthago, et cum de alio consuleretur, pronuntiabat. His conclusion: Cato probably spoke the words Carthago delenda est, but there is no reported speech in the sources and Plutarch was dramatizing. (And there is no use of censeo in any source, that only appears when backtranslating Plutarch).

  48. Stu Clayton says:

    das wievielte

    Another German expression with puzzlement potential, depending on how it’s used: die wenigsten = “the fewest”.

    Die wenigsten wissen das = “Hardly anyone knows that”.

    The seasoned peever thinks: but that’s the superlative of few. “The fewest” should mean “none”.

    But that’s not how it works. Die wenigsten in the sample sentence above means “very few”. It’s not an absolute superlative, but rather an alternative to sehr wenige.

    Warning: die wenigsten is used in this way primarily by the masses, and by elites only in unguarded moments.

  49. Lars (the original one) says:

    Well, the same logic would apply to die meisten or most people, that should mean everybody. But it doesn’t. Language is nice that way, how else would we fill a whole blog?

  50. @Lars: Thanks for digging up some more sources!

  51. Stu Clayton says:

    “The hostess with the mostest.”

  52. Trond Engen says:

    Stu: Die wenigsten wissen das = “Hardly anyone knows that”

    No. De færreste veit at….

    Lars may have found a rare exception, but here we are right back in the German calquery.

  53. Etienne says:

    On “das wievielte”: French has an exact equivalent, “le/la combientième”, which matches the German form in meaning and in morphological make-up. A minor mystery (well, to me at least) involves the use of /t/ separating the suffix -“ième” from “combien”: one would expect simple denasalization instead, yielding “combiennième” : I have never heard the latter, but according to this source (which gives example sentences in French, plus equivalents of the word in other languages: regarding your question to Bathrobe, David, it does give a form “dì jǐ ” for Mandarin)-

    -it exists, albeit marginally. If I had to guess, I would propose that the /t/ is due to analogy with “vingt”: if “vingt” yields “vingtième”, then it is unsurprising that “combien”, ending as it does with the same phoneme, should in French speakers’ minds also require a linking /t/ before the suffix -“ième”.

  54. PlasticPaddy says:

    Or a calque from an Alsacien (ne)

  55. David Marjanović says:

    the nature-folks

    Ah yes, Naturvölker “peoples composed entirely of noble savages”.

    But as you say, it’s moot, as the original account about Cato’s phrase is in Greek.

    So the whole alliteration in [k]eterum [k]enseo Karthaginem is just an inference? Bummer.

    Warning: die wenigsten is used in this way primarily by the masses, and by elites only in unguarded moments.

    Maybe without a following noun; but with one it’s quite unremarkable. I’ve never encountered any peevery about it.

    (The peevers strike immediately, of course, when das einzige “the only” is doubled up to das einzigste.)

  56. John Cowan says:

    Or a calque from an Alsacien(ne)

    Long ago the calques actually ran the other way. The Latin large-ordinal ending -ēsimus as in vīcēsimus, trēsimus, fell together with the superlative -issimus in Pre-French as -iesme, and the Germans apparently got the idea that the highfalutin thing to say was twentiest, thirtiest. Note the perfect parallelism of fleißig:fleißigste::dreißig:dreißigste.

    (The Latin ending is really a pseudo-ending < older Latin *vīcēnssos + -imus < decimus.)

  57. PlasticPaddy says:

    What I meant was wieviel+t+e => combien+t+ième

  58. David Marjanović says:

    Vingtième strikes me as a much more likely source. Centième may have helped.

    the highfalutin thing to say

    Depending on how far west the Pre-French pattern reached, it may well be a substrate phenomenon.

    Anyway, there’s another ordinal-number phenomenon that French and German have in common: the completely regular ways to say “2nd” – zweit-, deuxième. But if that is due to contact, the contact must have happened much later: second still exists in French ( < Latin secundus, originally “following”), and “other”, the usual Germanic way to say “2nd”, was still in that use in German in the 16th century, even though zweit- is now universal (and productive as a noun prefix, too).

  59. Stu Clayton says:

    Since combien is an interrogative, maybe the “t” in combientième arrived by analogy with the “t” in the inverted subject-verb form of a question: combien y a-t-il de personnes là ? “how many people are there ?”

  60. Etienne says:

    David: your statement that “second still exists in French ( < Latin secundus, originally “following”)" is misleading: "second", in French, is most certainly NOT an inherited Latin word: it is a (rather old: first attested in the twelfth century) loanword, either from Latin itself or from Provençal (the fact that in French the word is pronounced /səgɔ̃/, and not */səkɔ̃/, despite the spelling, has been claimed to be due to its being borrowed from, or at least influenced by, (Old) Provençal "segond"). The Provençal word itself indeed is an inherited form deriving from Latin "secundum", with the regular voicing of Latin intervocalic /k/ to /g/ and the loss of the final vowel, which before its elimination was probably /o/ (final /m/ was already gone in Latin itself): this raising of the short /u/ to /o/ is an older, Italo-Western Romance innovation.

    Note that in French itself, intervocalic Latin /k/, when followed by a back vowel, is simply dropped: take for example an inherited French word such as "sûr', from Latin 'securum".

    Your suggestion of a substratum is to my mind intriguing. Consider the following map:

    which seems to indicate that Europe as a whole, in sharp contrast to the world outside Europe, is a hotbed for ordinal systems where "first" and "second" are suppletive, and where ordinals are regularly formed from 'three' up: I do know of one Romance variety which uses reflexes of ALTER, ALTERA as its ordinal for 'second", through the influence of a neighboring non-Romance variety.

    On the other hand, this same map indicates that, outside Europe, ordinal systems with "first" as the only suppletive member are the most common ones, and thus it is certainly possible that French and German each created "deuxième/zweite' independently of one another.

  61. What does the dialectal and documentary evidence say about the history of the spread of secundus throughout France?

  62. Etienne says:

    Y: As I wrote above, “second” is attested in the twelfth century in French: I consulted a few dictionaries, and both Gascon (South-Eastern France) and Franco-Provençal (Alpine area) have ‘second’-like and ‘deuxième’-like words: there thus does not appear, at first glance, to be any relic area(s) in France where neither word is attested. This says nothing about how old either word is, of course.

    Hmm. I once did some research on ordinals in Romance, and I think there is one area in the Southern Auvergne which lacks both “second”- AND “deuxième”-like forms. Will look it up and make sure my memory is not playing tricks on me…

  63. David Marjanović says:

    Note that in French itself, intervocalic Latin /k/, when followed by a back vowel, is simply dropped: take for example an inherited French word such as “sûr’, from Latin ‘securum”.

    …Oh. Oops!

    I had actually read just recently that sûr is regular.

    Your suggestion of a substratum is to my mind intriguing.

    I just meant a Romance substratum in southern German. Do you know when deuxième-type forms are first attested in French?

  64. Etienne says:

    David: In answer to your question: I do not know when “deuxième”-type forms are first attested, but the suffix -“ième” (whose exact origin and spread remain very obscure, by the way) only became general in French in the thirteenth century at the earliest. I have also found out two interesting things: 1-Old French typically used “autre” with the meaning “second”, and 2-A phonologically regular form ‘seont’, from Latin “secundum”, did exist in Old French, but seems not to have left any surviving reflex today.

    And I stand by the claim I made upthread: since ordinal systems with “first” as their only suppletive member are, outside Europe, the most common subtype, it is definitely possible that German and French each created their regular word for “second” (“deuxième”/”zweiter”) with neither influencing the other. Something similar happened in the history of the Gaelic languages: Old Irish had a suppletive form, “aile”, for “other, second”, but modern Gaelic languages use a form derived from “da” (two) as their ordinal today, with reflexes of ‘aile’ being confined to the meaning “other”. Do note that Brythonic languages preserve a suppletive ordinal for ‘second” which derives from a Proto-Brythonic cognate of Old Irish “aile” (cf. Modern Welsh “eil”), so this loss of an inherited suppletive form for ‘”second” is not a pan-Western European tendency.

  65. Savalonôs says:

    How do the jātakas refer to the future Gautama Buddha if not as the bodhisattva? I’ve never gotten the impression that temporizing is one of the main saliencies of the term bodhisattva, although it’s entirely possible that my impression is idiosyncratic.

  66. ə de vivre says:

    I mean, depending on the flavour of Buddhism the temporal location of bodhisattvas may vary. Theologically, being a bodhisattva means accumulating merit without the final extinction of nirvana. Culturally, Mahayana traditions use the bodhisattva model to make a distinction between themselves and Abhidharmika traditions that have the final extinction of nirvana as the “goal” of the practice.

    Tathagata is how the historical Buddha usually seems to be referred to in Mahayana scriptures (which came to dominate Central Asia during the relevant time frame), and Shakyamuni is pretty common ecumenically. The significance of “bodhisattva” in Buddhist thought and practice is that it’s not just a synonym of Buddha, even if the two happen to share a referent in the historical Buddha.

  67. Stu Clayton says:

    @ə: Interesting to learn that, ill-equipped as I am to appreciate it fully.

  68. David Eddyshaw says:

    Modern Welsh “second” should be ail.

    The WALS database may not be quite comparing like for like in this area.
    Kusaal and its kindred have a suppletive ordinal form for “first”, but only in the sense that deeng ” first” is the only ordinal as such: from “two” onwards you use periphrases like ayi’ daan “owner of two” or lini paas ayi’ “which adds up to two” etc. And the cardinal “one” itself is kinda internally suppletive, with two more or less synonymous unrelated stems yeong and arakon’. The Mooré cardinal ayimbre “one” is instead parallel in structure to the Kusaal adjective yimmir “single.”

    That wouldn’t detract from the basic point that Europe is peculiar, though.

  69. David Marjanović says:

    The WALS map does have 12 orange languages with “first, two, three” in a remarkably equatorial distribution, except Nivkh.

  70. John Cowan says:

    The nice thing about first is that (like its antonym last and the analytic superlative-maker most) it’s a superlative in form. The same is true of primus, which predates the just-pre-Classical standardization on -issimus as the only superlative ending. Other survivals are ultimus, intimus, infimus, imus, summus, supremus. The already-suppletive optimus, pessimus, maximus, minimus, plurimus also fit in here.

    Extrēmus ‘outermost’ is yet another example, but owes its r to the comparative exterior, as if ‘out-er-est’. Cf. English innermost, outermost, uttermost (the last two being a doublet).

    It’s also characteristic of suppletives that when they are replaced, the replacement is often itself a suppletive. Thus inherited other was replaced (as an ordinal) by the borrowed suppletive second, which itself replaced suppletive alter in Latin. Another unrelated example is OE eode being replaced by ME-ModE went.

    It so happens that I am writing Python code for $EMPLOYER to convert German numbers-in-words appearing in legal documents into numerical values, so ordinals are much on my mind right now. Some of these documents are quite old, so I have to deal with spaces and hyphens and other traditional baggage. Hopefully I’ll be able to open source this code.

    I have decided, however, to limit my recognition of common fractions to those with denominators of 100 or less, as I can’t get agreement on whether 1/103 is eins hundertdrittel or eins hundertdreitel — and who gives a damn anyway? I am handling anderthalb ‘1.5’ correctly as a special case, though.

  71. Stu Clayton says:

    I can’t get agreement on whether 1/103 is eins hundertdrittel or eins hundertdreitel

    Weder noch. The systematics are like so:

    1/100 ein Einhundertstel
    1/101 ein Einhunderteinstel (oder Einhundertundeinstel)
    1/102 ein Einhundertzweitel (oder Einhundertundzweitel)
    1/103 ein Einhundertdrittel (oder Eimhundertunddrittel)

    Where did you get that eins with “s” in connection with counting (fractions or not) ? You don’t say eins Drittel, but ein Drittel.

    Anderthalb is so cute, no ?

  72. Interesting. Russian (+ Ukrainian, Polish) also have a word for one and a half, полтора from пол (half) втора (Old Russian form for second). Wiktionary lists numbers in halves based to this model up to half-sixth. I would not be surprised if John Cowan found a wiki list of languages with special cases for 1.5. There is also a word for 150 – полторастра, but no word for 15 – полторадесят, though полтора itself works as a multiplier just fine, полтора десятка (1.5 x 10) works.

  73. English has a special term for one and a half, but it has different semantics than a numeral. If the first zucchini is 8 cm long and the other zucchini is 12 cm long, the second is “half again” the length of the first. Or—translating Archimedes—given a sphere and the smallest cylinder containing it, the cylinder is half again the sphere in both surface area and volume.

    I use “half again” pretty frequently in everyday speech (certainly more often than “one and a half”).

  74. John Cowan says:

    eins: just a thinko. Thanks for the alternative forms, though. Right now I only handle und between the ones and tens word.

  75. Trond Engen says:

    Scandinavian halvannen “1½” and obsolete halvtredje “2½”, halvfjerde “3½” etc. I’m sure I’ve told before of tremenning “second cousin”, halvtredjemenning “second cousin once removed”. The form is also underlying the Danish counting system halvtredjesindstyve > halvtreds “50”.

  76. David Marjanović says:


    I just spontaneously read that as ein Hundertdrittel. (I don’t do einhundert.)

    Anderthalb is so cute, no ?

    Likewise absent from my active vocabulary in favor of the systematic eineinhalb.

    All this, and -und-, is regional, but I don’t know any details.


    sestertius < semi-tertius “coin worth 2½ asses“. Allegedly abbreviated IIS, whence $.

  77. January First-of-May says:

    Allegedly abbreviated IIS, whence $.

    There’s about a dozen alleged origins for $, many of them utterly fanciful. (In particular, IIRC, historians still aren’t quite sure whether it originally had one vertical line or two, and apparently some believe that those two variants are of independent origin.)

    For what it’s worth, I learned the “Pillars of Hercules” explanation – not sure if it’s still accepted.

  78. PlasticPaddy says:

    If anything, in Irish, dara (second) can be used in place of eile (other) in ways I do not think are common in other languages. To say there was no one else there you could say “Ní raibh an dara duine ann” as well as “Ní raibh duine eile ann”. In fact I think to say he had no other choice you have to say “Ní raibh an dara rogha aige” (the version with eile sounds incorrect to my ears). The only similar English usage I can think of is “no two ways about it”.

  79. PlasticPaddy says:

    There is also aon “one, any”, which can be used with eile, so “Ní raibh aon rogha eile aige” would be OK.

  80. ensi, toinen — but on the other hand also the predictable regular yhdes- and kahdes- occur in yhdestoista ‘eleventh’, kahdestoista ‘twelfth’, kahdeskymmenes ’20th’, kahdessadas ‘200th’…

    [in] Europe … ordinals are regularly formed from ‘three’ up

    They might not be outright suppletive, but the likes of third, fifth, twelfth are not quite regular either. This also again goes on outside of Indo-European too: Finnish kolme but kolmas, ditto e.g. Northern Sami golbma : goalmmát; Komi куим /kujim/ : коймӧд /kojmɤd/…

  81. I was surprised to learn that the -s in kolmas, etc goes back to -nci, eerily reminiscent of (completely regular) (bir)inchi, etc in Turkic.

  82. And going back before that to *-mte, giving rise to Hungarian -d! A happy hunting ground of phonological developments…

  83. You can predict about this much already knowing just Finnish, given that the oblique stem is -nte- / -nne- (kolmantena, kolmannen). And yes, the Turkic parallel is one of the long-known morphological points alluded to as “Ural-Altaic”, though it probably would be neater still if the PU form had a *-č- rather than *-t-.

    PIE however seems to have had a remarkably messy ordinal system with not just suppletive ‘first’, but also a large gaggle of different suffixes for numerals above that (*-to-, *-o-, *-mo-, *-wo-)…

  84. The systems in the IE languages can be partially traced back to a state where *-o was added to the cardinal number. Later, the last consonants of the cardinal numbers were re-interpreted as part of an ordinal suffix, in different ways in different IE languages.

Speak Your Mind