In an earlier thread (under heading CALIBER), I showed that you can’t trust today’s English dictionaries to reliably summarize what’s known about a word’s origin, with the worst aspect being a failure to convey uncertainty. The word HAZARD provides an example where some of the dictionaries do in fact convey uncertainty, but they mislead the reader another way:

Merriam-Webster @ M-W.com: Middle English, from Anglo-French hasard, from Old Spanish azar, from Arabic al-zahr the die. First Known Use in English: 14th century.
Concise OED @ OxfordDictionaries.com: Middle English: from Old French hasard, from Spanish azar, from Arabic az-zahr ‘chance, luck’, from Persian zār or Turkish zar ‘dice’.
Collins English @ Dictionary.com: C13: from Old French hasard, from Arabic az-zahr the die
Random House @ Dictionary.com: 1250–1300; ME hasard < OF, perh. < Ar al-zahr the die
Webster’s New World @ YourDictionary.com: ME < OFr hasard, game of dice, adventure < ? Ar az-zahr, for Egypt colloq. Ar al-zahr, dice
American Heritage @ YourDictionary.com: Middle English hasard, dice game, from Old French, possibly from Old Spanish azar, possibly from Arabic az-zahr, the gaming die : al-, the + zahr, gaming die.

The compilation book “Word Origins” by John Ayto (2005) adopts the standard story above and puts it on synthetic steroids: “The word hazard was introduced to English as the name for a game played with dice. It was borrowed from Old French hasard, which came via Spanish azar from Arabic azzahr, earlier al-zahr ‘luck, chance’.”

The French word spelled ‘’hasart’’ is attested circa 1150 meaning a game of dice—ref: CNRTL.fr. An attestation in French 1200, also spelled “hasart”, meant « un certain coup au jeu [= a game] de hasard » (same ref). The Spanish word ‘’azar’’, which had the same meaning as the French, is not attested until more than a century later, in 1283 (same ref). There is no attestation in French for a wordform for ‘’hasard’’ that displays an ancestry in the Spanish ‘’azar’’, or at least no early attestation. There’s a later wordform in Latin ‘’azardum’’, which looks to me like a blend of the Spanish and French—ref: DuCange. Therefore, the Spanish origin proposition for French hasard must be highly uncertain and hypothetical. In records in England, the proper name “Hugo Hasard” occurs in 1167, and a “Walteri Hassard” occurs in 1197—ref: UMich MED. Those are Norman names I believe. They are unconnected to the game of dice but they show that a word of the form Hasard/Hasart could readily be created in (Anglo-Norman) French without Spanish fatherhood. Another example is in William of Tyre writing in Latin in the 1180s. He writes about a castle or fortified town he variously calls “Hasard”, “Hasart” and “Hasarth” --ref (hyperlink at upperleft)—this castle was controlled by people who spoke French as their vernacular languge, as did William of Tyre himself.

The Arabic ‘’az-zār’’ or ‘’az-zahr’’ is unattested in Arabic until a report in the early 19th century by the French linguist Ellious Bocthor, who found it in Egyptian oral dialect meaning “the dice”. (The word is absent in the great Richardson’s Arabic-English Dictionary of 1852). Bocthor’s attestation is more than 700 years too late. Nearly 800 years too late. It’s a piece of flotsam picked up along the wide wide shores of Arabic, dialectical or standard, for the purpose of retrofitting it to Spanish ‘’azar’’. It comes with zero support from history beyond the fact that Spanish was taking in Arabic words in the 12th century.

In the centuries leading up to the 19th, Egyptian Arabic borrowed many words from Europe especially from Italian. Italian speakers dominated the seaborne commerce of the Turkish Ottoman empire in and around the Renaissance centuries; Italian was the “lingua franca” of the Mediterranean rim, and a great many Turkish and Arabic words are from Italian as a consequence. Ernest Weekley’s 1921 etymology dictionary in its entry for hazard says the 19th century Egyptian Arabic dialectical ‘’az-zahr’’ “is a word of doubtful authority which may have been borrowed from Spanish ‘’azar’’ or from Italian ‘’zara’’, “a game at dice called hazard"."

The American Heritage Dictionary quoted above says the European word is “possibly from Arabic az-zahr, the gaming die : al-, the + zahr, gaming die.” That statement permits an ordinary reader such as myself with no further knowledge of the situation to mistakenly infer that ‘’az-zahr’’ was a real attested word in Arabic in the 12th century. The Chambers Dictionary @ ChambersHarrap.co.uk, distinct from the other dictionaries, says English hazard is from French hasard and doesn’t attempt to summarize the word’s history back any farther than that. That’s the best policy, I argue.

Before you go off on such rants about etymology, you really need to look at the OED. This is that dictionary’s etymological note on hazard (from the 1989 2nd edition):

The origin of the French word is uncertain, but its source was prob. Arabic. According to William of Tyre, the game took its name from a castle called Hasart or Asart in Palestine, during the siege of which it was invented: see Littré s.v. The true Arab name of this castle appears to have been ‘Ain Zarba (Prof. Margoliouth). Mahn proposes vulgar Arab. az-zahr or az-zar ‘die’ (Bocthor); but early evidence for this sense is wanting.

[from OED]According to William of Tyre, the game took its name from a castle called Hasart or Asart in Palestine, during the siege of which it was invented: see Littré

In my post above, I linked to the original Latin text of William of Tyre’s book, written in the 1180s. That link again is here. William of Tyre only speaks of Hasard|Hasart|Hasarth as a fortified town. He never speaks of hasard as a game of dice. Later on, his book was translated into French. The translator (or somebody) inserted the remark that the name of the game “hasard” took its name from the Hasert castle. I haven’t seen a definite date for when that insertion happened. According to an incidental remark in this article dated 1971, the insertion happened in the “thirteenth century”. But in any case it’s clear that it was an insertion made well after the word hasard was well established as a word in French, and well after William of Tyre was long dead. Therefore the story is totally apocraphal (sp?) and neglectable in any report about the origin of Hasard, and hence I expect that Littré and the author of the OED “hazard” article were unaware that the story was apocraphal. By the way, William of Tyre’s book also contains nearly a dozen references to dice playing, all of which you can readily get from the concordances lists in the above super-hyper-texted version of his Latin book. The Latin for dice is “alea__” and the Latin for game is “lud___”.

[from OED]The origin of the French word is uncertain, but its source was prob. Arabic.

What evidence can the OED present in support of the notion that the source was probably Arabic, other than the bogus item in William of Tyre?

Dave Wilton - 20 November 2010 03:48 PM

Before you go off on such rants about etymology....

My intention has been to make the point that the summary etymologies in all the English dictionaries are unreliable and untrustworthy. I don’t count the OED with the rest because the OED has time and space to go into details. I’m reporting the facts about the shitty etymology stories in the dictionaries. I do not forget that I made a bad mistake earlier concerning what cannons are made of. But I don’t think you’ve got a basis for slurring me with “ranting”, even though it’s true that I don’t have the benefit of whatever info is in the OED.

Furthermore my point above was that the “az-zahr’’ origin story for hazard is meritless. Since David Wilton hasn’t raised any item that would give merit to the ‘’az-zahr’ story, he’s got no basis in context for saying it’s “such rants”.

I find these notes interesting. Note that an alternative Arabic etymon is also presented in the above CNRTL entry. Perhaps some of these proposed old etyma should be marked with an asterisk or something when they are not attested from the correct time or otherwise dubious or conjectural.

What evidence can the OED present in support of the notion that the source was probably Arabic, other than the bogus item in William of Tyre?

Again, my advice is that you ask the OED, not us.  I know from personal experience that they are scrupulous in their requirements and store either hard copies or facsimiles of evidence in support of all their claims.  They don’t just invent theories.  They work from hard evidence whereas the constraints of the internet prevent you from physically showing us any hard evidence. Besides, we can’t rectify mistakes (even if there are any) by dictionaries.  But we can email them, as many of us have done, with new evidence.  If you have any evidence, I suggest you email it to them.

ᴚǝǝƶɐʍɐɈ - 20 November 2010 01:56 PM

There is no attestation in French for a wordform for ‘’hasard’’ that displays an ancestry in the Spanish ‘’azar’’, or at least no early attestation. There’s a later wordform in Latin ‘’azardum’’, which looks to me like a blend of the Spanish and French

I presume you must be referring to one or more of the following “discrepancies”:  (1) initial “h”; (2) “s” in place of “z”; and (3) final “d"/"t" in French hasard/hasart. But these are easily explained by the following features of French: 

(1) As noted in the CNRTL entry that you cited, (an unpronounced “h") was in the Middle Ages frequently added to words beginning with a vowel, particularly ones of foreign origin.

(2) “s” between vowels had for centuries been pronounced [z], so that there would have been no difference in pronunciation between hasard and hazard. The “normal” spelling for French [z] is in fact “s” not “z”.

(3) ard/art was a frequent ending (initially for words of Germanic origin) and so it is easily understandable that a final d/t was added to hasar.  Other examples include boulevard, first attested as bolewers (with unpronounced “s"), boyard (cf. English boyar), hussard (cf. English hussar). This process was facilitated by the fact that by the 12th century word-final d/t had ceased to be pronounced.

Thus medieval Spanish azar would have been pronounced indistinguishably from French hasard/hasart!

My “rant” comment is based on statements like “you can’t trust today’s English dictionaries to reliably summarize what’s known about a word’s origin,” “summary etymologies in all the English dictionaries are unreliable and untrustworthy,” and “shitty etymology stories in the dictionaries.” You make good points about the uncertainties surrounding the etymologies of the particular words you address, but you can’t extrapolate a general opinion about tens of thousands of etymologies based on a handful of cherry-picked examples. Hence, “rant.” Sticking to the cases at hand would serve your argument better.

I suspect the OED editors knew exactly what they were writing and the precise nature of the reference to the game in the manuscript of William of Tyre’s work. They are pretty scrupulous about their references. Although in this case, it appears as if the summation (even the OED has to summarize severely) limits how they can communicate the nature of the reference. Of course William, writing in Latin, did not make use of the French word. Furthermore, he’s writing about events in the past (I believe the First Crusade in this instance, but I could be wrong about that), which happened before the 1150 appearance of the word in French. I suspect his thirteenth-century translator knew this too when glossing the work. The fact that the etymological gloss comes later is not a reason to dismiss it outright as “apocryphal” and “neglectable.” The medieval etymology may or may not be accurate, but it probably reflects a general belief at the time about where the word came from and such beliefs are not always far off base.

For the most part, I am enjoying your contributions to the board. I just wish you didn’t generalize so much from specific examples.

For what it may be worth, the Spanish azar is derived from Arabic, unless one believes that the Royal Academy of the Language is
etymologically inept.

azar.

(Del ár. hisp. *azzahr, y este del ár. zahr, dado1, literalmente ‘flores’).

Rough translation: Azar, from Hispanic Arabic azzahr, and this from the Arabic zahr, a die, literally “flowers”.

It shows up, with a definition quite similar to that in modern English, in the Diccionario de Autoridades, 1770, with no etymology given.

The first appearance of the etymology is in an early 20th century edition. 

http://buscon.rae.es/ntlle/SrvltGUIMenuNtlle?cmd=Lema&sec=1.0.0.0.0.

The original poster would, I can assert with some confidence, say that Royal Academy of the Language is etymologically inept.

May I hog a ringside seat? I vow not to take sides. but to cheer on.

hazard
mid-12c., from O.Fr. hasard “game of chance played with dice,” possibly from Sp. azar “an unfortunate card or throw at dice,” which is said to be from Arabic az-zahr (for al-zahr) “the die.” But this is doubtful because of the absence of zahr in classical Arabic dictionaries. Klein suggests Arabic yasara “he played at dice;” Arabic -s- regularly becomes Sp. -z-. The -d was added in French in confusion with the native suffix -ard. Sense of “chance of loss or harm, risk,” first recorded 1540s; the verb sense of “put something at stake in a game of chance” is from 1520s.